The CDCDP Manual
The CDCDP Manual
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Some believe these figures and usage could be on the conservative side, and actually
higher than projected; either way, the demand for faster processing capabilities has
________________________________
seen increased power demands within the Data Centre community.
________________________________
Note: Statistical figures collated from numerous industry surveys during 2014-15
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The amount of data now being kept in digital format is expected to continue to rise,
and the amount of energy required to keep these data will consequently continue to
________________________________
rise, as will the associated level of carbon emissions.
At the regulatory level, this massive growth in energy demand and carbon dioxide
________________________________
emissions associated with running Data Centres is likely to draw scrutiny by regulators
looking for ways to reduce carbon dioxide emissions in line with Government and
________________________________
European Union targets. Data Centres that are able to achieve significant energy (and
carbon dioxide) savings have the potential to gain a competitive advantage through
________________________________
lower costs-per-byte.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The traditional culture does not promote change, but forward thinking and a clear
understanding of the impacts of increasing energy demands is now changing both
________________________________
business and design views.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Manny Data Centres are power-constrained, but we can no longer rely on additional
power from the grid (or being able to buy more). In many countries, these issues can
________________________________
be due to national limitations or local conditions.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The traditional culture is now out of step with the world and is changing.
► Do you know how much the energy used in your Data Centre costs?
►________________________________
How much CO is produced per kWh?
2
________________________________
The published Carbon Trust conversion table gives a figure of 0.5452 tonnes CO per
kWh of grid electricity.
2
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It should be noted figures will vary across the many global surveys, but they all
demonstrate there is still at present a high dependence on fossil fuels. However, as the
________________________________
global population continues to grow, new and sustainable energy sources will have to
be introduced.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A tremendous amount of activity occurs on the Internet every 60 seconds, with the
above being just selection from the top recorded activities. Although it provides many
________________________________
media and communication options, it also has a marked effect on bandwidth and
processing requirements, which in turn result in a continual demand for more energy
________________________________
to provide these expanding online services.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The pressure on operators to become more energy efficient has become a major
operational factor, due to the need to reduce operating costs whilst also having to
________________________________
meet regulatory demands and reduce CO2 emission rates. Meanwhile, they must
continue to remain effective (and even increase productivity) to meet business
________________________________
demands. Therefore, it is no surprise energy efficiency has become a major and
significant factor across the entire Data Centre industry.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Green energy generated from renewable sources is not readily available everywhere.
Some operators are making green energy a high priority when choosing a location.
________________________________
Washington State and Iceland, for instance, are susceptible to earthquake and volcanic
issues, but they have renewable sources of energy.
________________________________
Renewable energy has been mooted by the IT press as a possible source of additional
power capacity for the last number of years. In reality, the energy required by even the
________________________________
most efficient Data Centre makes solar, wind and even tidal energy unsuitable at
present as a prime source to support critical IT loads; but technology is continuing to
________________________________
improve, and Data Centre designs are now becoming a lot more innovative, with a
________________________________
drive to reduce energy consumption and implement greener options.
.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
An Environmental TCO certification badge means that a product has had to pass at
least 50 tests relating to emission levels, energy consumption, ergonomic design and
________________________________
ecological soundness. Environmental TCO is a European equipment energy efficiency
certification programme, developed in Sweden.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The picture can be confusing and negotiating the minefield is equally tricky.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
What challenges will the Data Centre community face when the Cloud becomes fully
mature and utilised? The arrival of ‘Big Data’ applications has introduced further
________________________________
challenges.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Since other industries may be involved in ensuring compliance of the Data Centre operation, the
CDCDP should also be knowledgeable about the related Standards, as well as local and national
________________________________
regulations appropriate to the region.
Worldwide
________________________________
► IEC 60364 defines common rules for electrical installations; IEC 61936 sets out rules for installations
>1kV. 90% of the electrical installations in the world are built to IEC 60364 and IEC 61936, or to the
US National Electrical Code. The general aims of electrical Standards are the same, and there has
________________________________
been a considerable effort over many years to bring the different Standards closer together.
► American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) Standards are
________________________________
also commonly quoted in Europe as a ‘standard’ for Data Centres.
► Building and utility codes, planning rules and Telco regulations can also apply to Data Centres, e.g.,
________________________________
ITU, Telcordia GRE, ETSI, etc.
Europe
________________________________
The European Standard EN 50110 ‘Operation of Electrical Installations’ consists of two parts:
► EN 50110-1 contains minimum requirements valid for all CENELEC countries and some additional,
________________________________
informative annexes dealing with safe working.
► EN 50110-2 is a set of normative annexes (one per country) which specify either the present safety
________________________________
requirements, or give the national supplements to these minimum requirements.
► BS 7671 is the country variant for the UK which ensures compliance with the IEC and EN
________________________________
USA
Standards.
________________________________
National Electrical Code (NEC), National Fire Protection Association (NFPA) Codes and Local Codes
apply. The CDCDP needs to keep abreast of the latest discussions regarding codes and Standards, as
they can often be wrongly (and expensively) applied. For example, NEC 645 and NFPA 75 do not apply
________________________________
to all IT facilities, and may or may not apply to a new Data Centre, so the designer must determine
applicable codes at the outset.
________________________________
The Electricity at Work Regulations (1989) - Enforces application of safe
practices into all electrical works in the UK. Relevant to all work activities and
________________________________
premises, except certain offshore installations and particular ships. Intended to assist
duty-holders in meeting regulatory requirements, it will be of interest and practical
________________________________
help primarily to engineers (including those involved in design, construction,
operation, or maintenance of electrical systems and equipment), technicians and
________________________________
their managers.
Electricity at Work: Safe Working Practices - The guidance covers key
________________________________
elements to consider when devising safe working practices, and is for people who
________________________________
carry out work on or near electrical equipment. It includes advice for managers and
supervisors who control or influence the design, specification, selection, installation,
________________________________
commissioning, maintenance, or operation of electrical equipment. This third edition
updates the guidance and provides sources of further information.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The Electricity at Work Act enforces application of safe practices into all electrical
works. BS 7671 is the recognised UK Standard for the safety of electrical installations.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The National Electrical Code (NEC), or NFPA 70, is a United States Standard for
the safe installation of electrical wiring and equipment.
________________________________
The NEC codifies the requirements for safe electrical installations into a single,
standardised source. It is part of the National Fire Codes series published by the
________________________________
National Fire Protection Association (NFPA), and while not itself a US law, NEC use is
commonly mandated by state or local law.
________________________________
The “authority having jurisdiction“ (or AHJ) inspects for compliance with these
________________________________
minimum Standards. ‘National Electrical Code’ and ‘NEC’ are registered trademarks of
the NFPA.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There are many political challenges facing everybody right now which come under the
‘green’ banner. Some of these are more obvious than others, but all have one thing in
________________________________
common for a Data Centre operator – they are going to cost you money.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Many countries have signed up to reduce CO2 emissions over time. These targets are
steep and penalties are being introduced to force businesses to comply.
________________________________
The EU has also taken a leading role in global environmental negotiations, especially the
signing of the Kyoto Protocol. At the 1997 UN Conference on Climate Change in
________________________________
Kyoto, Japan, the EU committed its members to reducing greenhouse gas emissions 8%
by 2012, compared to levels in 1990. The 2008 EU Climate Change package established
________________________________
the '20-20-20 Targets' for 20% of energy to come from renewable sources, and
committed to reduce greenhouse gas emissions to 20% by 2020.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The current CCA scheme started in April 2013 and will run until 31 March, 2023. An
operator that has a CCA will measure and report energy use and carbon emissions
________________________________
against agreed targets over four, two-year target periods.
If an operator has more than one eligible facility in the same sector, it could hold
________________________________
individual CCAs for each facility, or choose to group them together under one CCA.
The target could then be shared across the grouped facilities.
________________________________
Once a facility, or group of facilities, holds a CCA, it is referred to as a target unit. If the
________________________________
operator meets its targets at the end of each reporting period, it continues in the
scheme and is eligible for the discount on the CCL. Details of businesses that hold
________________________________
CCAs are published every month as reduced rate certificates for each sector.
Climate change agreements are voluntarily made by UK industry and the Environment
________________________________
Agency to reduce energy use and carbon dioxide (CO ) emissions. In return, operators
2
receive a discount on the Climate Change Levy (CCL), a tax added to electric and fuel
________________________________
bills. The Environment Agency administers the CCA scheme on behalf of the whole
UK. For operators who hold a CCA, the CCL will be reduced by:
________________________________
90% on electricity bills
________________________________
65% on other fuels
CCAs are available for a wide range of industry sectors.
________________________________
________________________________
________________________________
The US DoE has partnered with industry to develop a Data Center Energy
Practitioner (DCEP) programme and qualification as a means of training Data Centre
________________________________
professionals in the benefits, goals and techniques of Energy Efficiency in Data Centres.
This is a dynamic and fast-developing industry: No sooner has the ultimate, efficient
________________________________
Data Centre been designed and built, than new technology and thinking appears which
renders the site suddenly less efficient.
________________________________
Creating a structured framework around Data Centre Energy professionals and the
________________________________
industry encourages the generation of new ideas, innovation and the sharing of Best
Practices.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Developed by the European Union’s Joint Research Centre (JRC), the Code of
Conduct for Data Centres is an initial response to the accelerating power
________________________________
consumption of Data Centres across the EU, mirroring the United States’ “Save Energy
Now” campaign, which led to the DCEP programme.
________________________________
Data Centre power consumption is expected to double in the EU by 2020 (in 2007,
Western Europe alone reported use of 56 TWh, and the 2020 projection is 104 TWh).
Its________________________________
generation and distribution systems are already under strain, and international
evidence of power outages (through excessive demand requirements) are not
________________________________
uncommon.
Add________________________________
to this the concerns for the environmental impact of carbon-based energy
consumption, and it becomes clear a lead had to be taken. At this moment in time, the
________________________________
EU’s strategy is to provide clear and concise direction aligned with common sense,
without resorting to legislation. To enable the programme to be driven forward, the EU
has________________________________
developed a set of best practice guidelines – “Best Practices for the EU Code of
Conduct on Data Centres”. The guide draws on the experience and expertise of Data
________________________________
Centre professionals, operators, vendors, consultants, academics and industry bodies.
Although currently the Code is voluntary, not adopting it is likely to have future
________________________________
business implications, as winning and maintaining UK and EU government contracts will
________________________________
likely depend upon businesses demonstrating commitment to the Code of Conduct.
________________________________
________________________________
“BREEAM Data Centres” is an assessment method and certification scheme that can
be used at the design, construction, and refurbishment phases of the Data Centre’s
________________________________
building's lifecycle. The scheme is used to assess the environmental impact of
unoccupied or occupied Data Centre buildings.
________________________________
Data Centres are used to house computer systems and associated components, such
as telecommunications and storage systems. Data Centre’s tend to be heavy
________________________________
consumers of energy and, as they have grown in number, interest in applying BREEAM
to assess their environmental impacts has also grown.
________________________________
Launched in 2010, the BREEAM Data Centres’ scheme enables the assessment of this
________________________________
specialised type of building, using the standard BREEAM methodology. The scheme
recognises Data Centres’ energy-intensive nature, and seeks to encourage and reward,
________________________________
through the scheme, the use of “best practice” in designing, building or modifying Data
Centres. BREAAM’s Data Centre scheme includes innovation credits, which provide
________________________________
additional recognition for a building that innovates in the field of sustainable
performance, above and beyond the level that is currently recognised and rewarded
________________________________
within standard BREEAM assessments. Innovation credits therefore enable clients and
design teams to boost their building’s BREEAM performance and, in addition, help
________________________________
support the market for new innovative technologies and practices.
________________________________
BREEAM Data Centres assessments can be carried out on :
New Data Centre buildings
________________________________
Major refurbishment of existing Data Centre buildings
Fit-outs of existing Data Centre buildings
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The EU have made some recommendations. These are not regulations (yet!). It is
similar in the USA, and initiatives such as DoE DCEP are appearing in the Data Centre
________________________________
space.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
This module will not attempt to turn you into an electrical engineer, but merely
introduces you to the world of power, and allows you to hold a meaningful
________________________________
conversation with an electrical engineer when needed. Additionally, you will appreciate
the impact these electrical parameters may have upon the Data Centre distribution
________________________________
system.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Ohm’s Law has many derivatives and is actually quite complex; however, the simple
version shown above will suffice for our needs.
________________________________
This equation is used to calculate the current that will flow in a circuit when a voltage
is applied, and also the voltage drop along the route as a current flows through
________________________________
different circuit components, e.g., in cables.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
This equation is used to calculate the power dissipation of a load, and the conductor
size to meet the current-carrying requirements in a circuit.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Resistors in series: Add their values together to get total resistance:
Rt = R1 + R2
________________________________
Additional resistors in series - simply continue adding to the total circuit resistance.
________________________________
Rt = R1 + R2 + R3 + R4 + etc.
________________________________
Resistors in parallel: Add the reciprocal of their values together to get the
________________________________
reciprocal of the total resistance.
1/Rt = 1/R1 + 1/R2
________________________________
Additional resistors in parallel - simply continue to add their reciprocals.
________________________________
1/Rt = 1/R1 + 1/R2 + 1/R3 + 1/R4 + etc.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The resistances can be connected in series and parallel. The current flowing, as well as
the voltages across each component, depend on resistance values and how they are
________________________________
connected.
In order to analyse what happens, we need some additional rules.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Kirchhoff's Current Law:
What goes in must come out, at any junction.
In________________________________
the circuit above:
________________________________
►The sum of the currents I2 + I3 – I1 – I4 = 0
________________________________
Kirchhoff's Voltage Law:
________________________________
There is no such thing as perpetual motion.
In the circuit above:
________________________________
►V4 – V1 – V2 – V3= 0
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Of course, resistance is only part of the story - there are other parameters to
consider. The fundamental parameters of resistance, inductance, capacitance and
________________________________
conductance all present an impedance to current flow. These impedances affect the
relationship between voltage and current in different ways.
________________________________
________________________________
Resistance – Ohm (Ω): A measure of the degree to which an object opposes an
electric current through it
________________________________
Inductance – Henry (L): Is the property of an electrical current causing voltage to
be generated, proportional to the rate of change in current in a circuit (property also
________________________________
called self-inductance to discriminate it from mutual inductance).
________________________________
Conductance – Siemens (S): Electrical resistance is measured by its opposition to
the passage of an electrical current; the inverse quantity is electrical conductance -
________________________________
measuring how easily electricity flows along a certain path. Hence, one Siemens = The
reciprocal of one Ohm.Also referred to as the ‘Mho’
________________________________
Capacitance – Farad (F): Is the ability of a body / circuit to hold an electrical circuit
________________________________
________________________________
________________________________
________________________________
The fundamental parameters need to be converted into common units in order to be
useful in circuit analysis. The common currency is Ohms. The resistance (now called
________________________________
impedance) of inductors and capacitors varies with the frequency of the current.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Everything so far has used DC.
As useful and as easy to understand as DC is, it is not the only ‘kind’ of electricity in
________________________________
use. Some sources of electricity (most notably rotary generators) produce voltages
alternating in polarity, reversing positive and negative over time.
________________________________
Think about this as a voltage-switching polarity, or as a current switching direction
________________________________
back and forth; this ‘kind’ of electricity is known as AC.
With AC, it is possible to build electric generators and power distribution systems far
________________________________
more efficient than DC, and so we find AC is used predominantly across the world in
high-power applications.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
If a machine is built to rotate a magnetic field around a set of stationary wire coils with
the turning of a shaft, AC voltage will be produced across the wire coils as that shaft is
________________________________
rotated, in accordance with Faraday’s Law of Electromagnetic Induction. This is the
basic operating principle of an AC generator: As the speed of rotation changes, the
________________________________
frequency of the current and voltage reversal changes. This is called the frequency of
AC, and it carries the unit Hertz (Hz).
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The reason a generator produces sine-wave AC is the physics of its operation - The
voltage produced by the motion of the rotating magnet is proportional to the rate at
________________________________
which the magnetic flux is changing, perpendicular to the coils (Faraday’s Law of
Electromagnetic Induction).
________________________________
This is greatest when the magnetic poles are closest to the coils, and least when the
________________________________
magnetic poles are furthest from the coils.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Transformers work due to a phenomenon called mutual inductance. Two coils, the
primary and secondary windings, are wound onto a ferromagnetic core, usually steel. A
________________________________
current flowing in the primary winding induces a current to flow in the secondary
winding. This current is proportional to the turns ratio - therefore the voltage across
________________________________
the secondary winding can be higher (step-up) or lower (step-down) than the primary
voltage.
In________________________________
an ideal transformer, most of the power is transferred.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Other waveforms occur in electrical systems due to a number of factors. Not all
equipment generates smooth sine waves - for example, switch mode power supplies
________________________________
and uninterruptible power supplies (UPS). Unexpected distortions can be caused by
non-linear loads, harmonics and other imbalances.
In________________________________
addition to root mean square (RMS), average, peak (crest) and peak-to-peak
measures of an AC waveform, there are ratios expressing the proportionality between
________________________________
some of these fundamental measurements.
________________________________
The crest factor of an AC waveform is the ratio of its peak (crest) value divided by its
RMS value.
________________________________
The form factor of an AC waveform is the ratio of its RMS value divided by its average
value.
________________________________
Square-shaped waveforms always have crest and form factors equal to 1, since the peak
________________________________
is the same as the RMS and average values.
Sinusoidal waveforms have an RMS value of 0.707 (the reciprocal of the square root of
2)________________________________
and a form factor of 1.11 (0.707/0.636).
Triangle and sawtooth-shaped waveforms have RMS values of 0.577 (the reciprocal of
________________________________
the square root of 3) and form factors of 1.15 (0.577/0.5).
________________________________
The RMS and average values of distorted wave-shapes are not related by the same
ratios.
________________________________
________________________________
When faced with analysing an AC circuit, the first step is to convert all resistor,
inductor and capacitor component values to impedances (Z), based on the frequency
________________________________
of the power source. After that, proceed with the same steps and strategies learned for
analysing DC circuits, using the ‘new’ form of Ohm’s Law:
________________________________
V=IZ
________________________________
The power dissipated in a resistor is always positive, but note in the case of inductors
and capacitors, it cycles positive and negative.
________________________________
Negative power means the inductor is releasing power back to the circuit, while
positive power means it is absorbing power from the circuit. Since the positive and
________________________________
negative power cycles are equal in magnitude and duration over time, the inductor
________________________________
releases just as much power back to the circuit as it absorbs over the span of a
complete cycle. What this means in a practical sense, is the reactance of an inductor
________________________________
dissipates a net energy of zero. This is quite unlike the resistor, which dissipates energy
in the form of heat.
A________________________________
capacitor also results in a power wave that alternates equally between positive and
negative. This means a capacitor also does not dissipate power, it merely absorbs and
________________________________
releases power.
________________________________
________________________________
________________________________
Poor power factor makes for an inefficient power delivery system.
Any power factor less than one means the circuit’s wiring has to carry more current
________________________________
than would be necessary with zero reactance in the circuit to deliver the same amount
of (true) power to the load. This can require larger cables.
________________________________
Also bear in mind that useful electrical power is measured in kW, but the incoming
________________________________
supply is measured in kVA; the difference is due to the phase angle of the load. The
supply company will charge for kVA, so the Data Centre needs to control the phase
________________________________
angle it presents to the supply so kW and kVA are as close as possible.
________________________________
Power factor (PF) = True power kW
________________________________
Apparent power kVA
________________________________
Poor power factor can be corrected, paradoxically, by adding another load drawing an
________________________________
equal and opposite amount of reactive power to the circuit.
________________________________
Inductive reactance can be cancelled by capacitive reactance. These two opposing
reactances bring the circuit’s total impedance equal to its total resistance, which makes
________________________________
the impedance phase angle equal (or at least closer) to zero.
________________________________
________________________________
The current flowing in the neutral is in the opposite direction to that in the line
conductor.
________________________________
What goes down the live conductor must come back along the neutral.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
How do we get three AC voltage sources whose phase angles are exactly 120o apart?
The best way is to generate it at the source, so we construct the AC generator in such
________________________________
a way that the rotating magnetic field passes by three sets of windings, each set spaced
120o apart around the circumference of the machine.
________________________________
These windings can be connected together in a wye (Y) or as a delta (D). There are
pros and cons to each, but these are beyond our scope here.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The phase angle is the difference between voltage and current, as shown above.
In this example, the phases are balanced so the residual current is zero.
________________________________
If all phases are balanced loads, i.e., a three phase motor, then no neutral current flows.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
If the loads on each phase are different, either in current or phase angle, the residual
current becomes non-zero.
________________________________
The conductor should be sized appropriately.
In________________________________
a Data Centre, IT loads are typically unbalanced, therefore there is neutral current.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Harmonics superimpose on the fundamental frequency and cause distortion in the
current and voltage waveforms. In a power system, this is never helpful.
________________________________
Harmonics turn the sinusoidal wave into a square wave which, in turn, change the RMS
and average voltage values.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
This system is fully duplicated in the power path to the critical load. It has:
► Diverse electrical supplies from the utility.
►________________________________
Duplicated distribution networks to split the critical load from the non-critical load.
________________________________
The common belief here is the critical load is the IT load, and everything else is
non-critical.
►________________________________
Duplicated UPS of some description in each network to ensure the supply to the
critical load is maintained.
________________________________
► Duplicated standby systems (usually a generator) in each network to replace the
________________________________
utility supply in case of a utility failure.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
About the Power Grid
The US transmission system (“the grid” ) is operated by a number of interconnected
________________________________
Independent System Operators (ISOs). ISOs pool and control energy over large
sections of the grid, and manage energy markets within their regions. Electric power
________________________________
transmission is the bulk transfer of electrical energy from generating power plants
to electrical substations, located near demand centres.
________________________________
This is distinct from the local wiring between high-voltage substations and
customers, which is typically referred to as electric power distribution. Transmission
________________________________
lines, when interconnected with each other, become transmission networks. The
combined transmission and distribution network is known as the "power grid" in
________________________________
the United States, or just "the grid". In the United Kingdom, the network is known
as the "National Grid".
________________________________
A wide area synchronous grid, also known as an "interconnection" in North
America, directly connects a large number of generators, delivering AC power with
________________________________
the same relative phase to a large number of consumers. For example, there are
four major interconnections in North America (the Western Interconnection, the
________________________________
Eastern Interconnection, the Quebec Interconnection and the Electric Reliability
________________________________
Council of Texas / ERCOT grid), and one large grid for most of continental Europe.
To eliminate some of the risks associated with multiple independent power feeders,
________________________________
power is delivered synchronised (by regulation), so if one feed is lost, the switchover
is nearly invisible to machinery (like air conditioning system chillers, air handlers and
________________________________
other large mechanical systems).
________________________________
Much of the power system is outside of the Data Centre’s control. Only the final
connection point with the utility company is open to discussion. However, this can have
________________________________
an impact on the rest of the power system design.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A common misconception is that the Data Centre should be located close to the
generating station. Whilst there may be arguments (of varying but questionable validity)
________________________________
about whether this improves the resilience of the supply or not, there is no real basis
for it on efficiency grounds. In any case, such sites are not generally available.
________________________________
Once the electricity leaves the power station, very small losses in the transmission
infrastructure mean there is little to be gained from colocating.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The Data Centre Electrical Distribution System
________________________________
Most of Europe, Africa, Asia, Australia, New Zealand and most of South America use a
supply that is within 6% of 230V. In the UK and Australia, the nominal supply voltage is
________________________________
230V +10% / −6% to accommodate the fact that most supplies are in fact still 240V.
Japan, Taiwan, North America and some parts of northern South America use a voltage
________________________________
between 100 and 127V. The 230V standard has become widespread so that 230V
equipment can be used in most parts of the world.
________________________________
Many areas such as the US, which uses (nominally) 120V, make use of three-wire,
________________________________
single-phase 240V systems to supply large appliances. In this system, a 240V supply has
a centre-tapped neutral to give two 120V supplies, which can also supply 240V to loads
________________________________
connected between the two line wires.
Three-phase systems can be connected to give various combinations of voltage,
________________________________
suitable for use by different classes of equipment. Where both single-phase and three-
phase loads are served by an electrical system, the system may be labelled with both
________________________________
voltages such as 120/208 or 230/400V, to show the line-to-neutral voltage and the line-
to-line voltage. Large loads are connected for the higher voltage. Other three-phase
________________________________
voltages (up to 830 Volts) are occasionally used for special-purpose systems, such as oil
well pumps.
________________________________
________________________________
________________________________
A utility network is very complex. A simplified example is shown here with typical UK
values.
________________________________
There is interconnection at each level, but this can vary between locations.
________________________________
The single point of failure is typically at the lower-voltage part of the network.
The power leaving the generation station is transformed to a higher voltage for
________________________________
transmission. The voltage is stepped down throughout various stages of the
distribution network.
________________________________
A typical 11kV substation is shown in the dotted red box.
________________________________
Many Data Centres will connect at 11kV, but larger Data Centres (typically Tier III and
IV) could connect further back in the network, at 33kV or 132kV.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The supply to the Data Centre could be at 11kV through a single unit substation, like
the one shown, or an interconnected series of them. In other countries, the voltage
________________________________
could be between 10kV and 15kV, but the principle is the same.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A very large Data Centre can take a supply at a higher voltage. This could provide
higher levels of resilience and better tariffs at the expense of providing a privately
________________________________
owned electrical network.
Once we look inside a larger Data Centre, we are once again confronted with an
________________________________
interconnected network, which distributes the power.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
However, a large Data Centre network can be divided into a number of clusters which
perform the same functions as the utility network. The highlighted area contains 11kV
________________________________
substations, which are very similar to the ones a utility would provide.
The network comprises switches, transformers, protective devices and interconnecting
________________________________
cables or busbars, similar to the utility network.
________________________________
The connection into the computer room in this Data Centre is at 400V, exactly the
same as if it was connected to the utility network. The difference is the Data Centre
________________________________
provides the distribution network rather than the utility company.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A transformer is the most convenient device for transfer of power from one voltage
to another at the same frequency. It works on the principle of electromagnetic
________________________________
induction.
Transformers are of two types:
________________________________
Step-Up Transformer: Steps up the voltage at secondary side
________________________________
Step-Down Transformer: Steps down the voltage at secondary side
________________________________
When multiple feeders from diverse sources are present, the fault-tolerance demands
on the Data Centre, in practice, tend to reduce significantly.
________________________________
Some small Data Centres are supplied from utility pad-mounted transformers at low
voltage. Some Data Centres get dual feeds from a single substation or a single feed –
________________________________
this dramatically reduces the reliability of utility power.
________________________________
For the highest reliability, modern Data Centres supporting mission-critical data and
applications have the utility Main Power delivered by at least two diverse feeds,
________________________________
originating from separate utility substations. This represents the first step in power
reliability and protection.
________________________________
________________________________
________________________________
________________________________
The TVSS device absorbs or diverts all energy present in the surge, clamping the "let-
through" overvoltage down to a safe level for exposed circuitry.
To________________________________
protect against these surges, the installed TVSS devices connect to all points of
potential voltage threat, and to a level below the equipment "withstand" voltage.
________________________________
TVSS protection is typically applied at several points throughout a facility:
________________________________
Service entrance point
________________________________
Distribution panels
Branch panels
________________________________
Individual circuits
________________________________
The TVSS device is important to a mission-critical electrical system, and its benefits are
great.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The major design goal is to divert as much of the transient energy away from the load
as possible by providing a lower impedance path, then to address the effects of the
________________________________
surge with components of the SPD. Various types of surge suppression components are
used to provide this low-impedance pathway, depending on the specific application of
________________________________
the device.
Everyone who uses electricity (grid-supplied or locally-generated), telephone lines
________________________________
(voice only or fax / modem / data), or computer data lines (data collection, data
transmission, computer communication) needs some type of TVSS. In short, anyone
________________________________
who uses wire to accomplish any of these functions is in need of effective and reliable
________________________________
surge suppression.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The electrical power is presented into the Data Centre through a main switchboard
similar to the one shown, this for many Data Centres is the demarcation point
________________________________
between the utility service and the Data Centre distribution system. The outgoing
cables to the Data Centre environments are sized to carry the current and maintain
________________________________
voltage levels; these are protected by an appropriate fuse or main circuit breaker
(MCB), according to the national electrical regulations.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The electrical power is presented into the computer room through a switchboard
similar to the one shown, exactly the same as if the supply had been from a utility. The
________________________________
outgoing cables to the equipment are sized to carry the current and maintain the
voltage levels; they are protected by an appropriate fuse or MCB, according to the
________________________________
national electrical regulations.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Transmission connection charges apply to generators, and the generation company
pays the transmission company for the right to connect. This is not normally an issue,
________________________________
but some Data Centres may have their own generators, and wish to export any spare
electricity to the public network. There are various charging policies in place which can
________________________________
have very different impacts. However, even in today’s political climate, there is no great
evidence to suggest that ‘green’ generators are given any special incentive by the
________________________________
transmission companies.
Use of System (UoS) charges apply over the transmission and distribution networks,
________________________________
i.e., from the power station to the Data Centre. These can vary dramatically from
________________________________
country-to-country. In some countries (England and Wales for instance), the UoS
charges are split between the generator and the end user; in other countries, the end
________________________________
user bears the entire cost.These charges consist of some or all of:
► Fixed charges (£ / month)
►________________________________
Capacity charges (£ / kVA agreed Maximum Demand)
________________________________
► Demand charges (£ / kVA above agreed MD), i.e., a penalty
► Usage charges (£ / kWh)
________________________________
Modifications are sometimes implemented, such as:
________________________________
► Time of year or time of day
► Type of load / power factor
________________________________
Tariffs are complex, and the correct (OR INCORRECT) choice can have a major
impact.
________________________________
In today’s world, climate change is on the agenda. Supply companies offer ‘eco-tariffs’ to
customers to reflect a contribution to reducing carbon footprint.
________________________________
Many organisations are choosing electricity supplied from more sustainable generation
sources to reduce their environmental impact. Most UK suppliers offer levy-exempt
________________________________
energy contracts when the electricity comes from accredited generation technologies
under the Climate Change Levy (CCL) exemption scheme.
________________________________
Typically two types:
►________________________________
Green energy from renewable generation technologies such as wind, small-scale
hydro and biomass.
►________________________________
Low carbon energy from good quality CHP. According to Ofgen, every 1MW of
________________________________
CHP operating in the UK helps reduce carbon emissions by 700-900 tonnes per
year. Electricity generated from CHP is more readily available than electricity
________________________________
generated from renewable sources.
Of course it is not possible for all the electricity to come from renewable sources.
________________________________
Renewable energy only accounts for 6% of the UK’s electricity supply, with targets to
reach 20% by 2020.
A________________________________
practical possibility for the Data Centre to get a better deal is to agree to be ‘cut
off’ when the utility grid becomes fully-loaded. This is not as dramatic as it sounds, as
________________________________
the utility will check before they do it to give the Data Centre time to run up the
________________________________
generators and carry out a smooth changeover. If there is a problem, then the power
will not be cut off. Reasonable tariff advantages can be achieved by initiatives like this.
________________________________
Bloom Energy Server™ is a new class of distributed power generator, producing
clean, reliable, affordable electricity at the customer site.
________________________________
Fuel cells are devices that convert fuel into electricity through a clean electro-chemical
process, rather than dirty combustion. They are like batteries, except they always run.
________________________________
The particular type of fuel cell technology is different than legacy ‘hydrogen’ fuel cells
in three main ways:
________________________________
Low-cost materials: Cells use a common sand-like powder instead of precious
________________________________
metals like platinum, or corrosive materials like acids.
High electrical efficiency: They can convert fuel into electricity at nearly twice the
________________________________
rate of some legacy technologies.
________________________________
Fuel flexibility: The systems are capable of using either renewable or fossil fuels.
Each Bloom Energy Server™ provides 200kW of power, enough to meet the baseload
________________________________
needs of 160 average homes, or an office building (day and night) in roughly the
footprint of a standard parking space. For more power, simply add more energy
________________________________
servers.
The Bloom Energy Server™ system has been successfully deployed in support of
________________________________
numerous Data Centre facilities in California, and is now being reviewed by many more
Data Centre operators.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
At the heart of every Energy Server is Bloom’s patented solid oxide fuel cell
technology. Each Energy Server consists of thousands of Bloom's fuel cells. Each cell
________________________________
is a flat, solid ceramic square made from a common sand-like ‘powder.’
Each Bloom Energy fuel cell is capable of producing about 25W, enough to power a
________________________________
light bulb. For more power, the cells are sandwiched, along with metal interconnect
plates, into a fuel cell ‘stack’. A few stacks (together about the size of a loaf of bread)
are________________________________
enough to power an average home.
________________________________
In an Energy Server, multiple stacks are aggregated into a ‘power module’; multiple
power modules, along with a common fuel input and electrical output, are then
________________________________
assembled as a complete system.
For more power, multiple Energy Server systems can be deployed side-by-side.
In________________________________
addition to Bloom's unmatched performance, this modular architecture offers:
________________________________
Easy and fast deployment
Inherent redundancy for fault-tolerance
________________________________
High availability (one power module can be serviced while all others continue to
________________________________
operate)
Mobility
________________________________
________________________________
________________________________
Increased reliability – and lower total cost of ownership (TCO): Bloom Energy
delivers a continuous source of UPS-quality power. Modular architecture enables
________________________________
operators to specify the level of availability, and N+M configurations to suit multiple
needs (where N = number required and M = the number installed). Bloom draws fuel
________________________________
from the highly-reliable natural gas grid, and utilises the electric grid as backup. This
delivers an order-of-magnitude greater reliability than traditional approaches, at lower
________________________________
lifetime costs.
Avoid uninterruptible power systems (UPS) and diesel generation: A Data
________________________________
Centre powered by Bloom Energy eliminates the need for traditional backup
equipment like diesel generators, UPS, batteries, and complex switchgear. This suite of
________________________________
legacy components is still too expensive, unreliable, and cumbersome. Bloom enables
Data Centre supplies to rely on a simpler, cleaner, lower-cost solution.
________________________________
Energy-efficient, no water use: Bloom Energy Servers generate power at
unmatched electrical efficiencies. Further, by generating power onsite, they avoid 7
________________________________
to 15% losses from transmission across the grid, and also avoid similar additional losses
from duplicate UPS systems (themselves employing inefficient, dual-conversion
________________________________
architecture). Bloom cells can run on 100% renewable biogas, and consume no water
during normal operation.
________________________________
Future applications: This is just the beginning - Bloom Energy can be deployed
________________________________
across the globe, tailored to any need. From urban nodes for low-latency, to core Data
Centres at low TCO, Bloom Energy's solution will be a catalyst for reimagining Data
________________________________
Centre design and deployment worldwide.
Reduced CO emissions: Dramatically more ozone-friendly against traditional fossil
2
fuel-based applications.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Data Centres are complex, and the entire building infrastructure is critical to
continuous operation. The power distribution system is at the heart of any critical
________________________________
facility, and it's vital everyone working in and around critical sites knows at least the
basics of your power distribution system.
________________________________
One of the most important aspects of a Data Centre is its power and power quality. A
diagram of a typical Data Centre power flow is shown above. In the diagram, power
________________________________
flow is shown from main utility power, to the delivery, to the non-critical and critical IT
________________________________
load (racks / servers). Components along this route include:
Low-voltage switchgear / switchboard
________________________________
Transfer switches
Switchgear
________________________________
Power Factor Correction (PFC) units (if required)
UPS system with input / output switchboard and UPS distribution switchboard
________________________________
Panelboards and Main Circuit Breakers (MCBs)
Power Distribution Units (PDUs) and Remote Power Panels (RPPs)
________________________________
Cabinet power strips
________________________________
________________________________
________________________________
________________________________
Dumb (No instrumentation - not manageable)
Metered (Equipped with a display showing current load on each phase)
________________________________
Switched (Receptacles can be individually switched on or off remotely)
________________________________
One challenge in selecting PDUs is to balance the relatively high cost, great
functionality and low risk of a switched PDU, versus the relatively low cost (but higher
________________________________
risk and lack of manageability) of a dumb PDU.
With a dumb PDU, your Data Centre runs the risk of phase unbalancing; devices may
be________________________________
unexpectedly plugged in, possibly tripping a circuit breaker. This could result in a
hard shutdown of critical equipment, causing data loss or corruption, and costly
________________________________
hardware damage.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
What can the different switching devices do?
Circuit breakers: Make and break all currents within the scope of their ratings, from
________________________________
small inductive and capacitive load currents up to the full short-circuit current; all fault
conditions in the power supply system, such as earth faults, phase opposition, and so on.
________________________________
Switches: Switch currents up to their rated normal current and make on existing short-
circuits (up to their rated short-circuit making current).
________________________________
Disconnectors (Isolators): Used for no-load and opening operations. Their function is to
“isolate” downstream devices so they can be worked on.
Three-position disconnectors: Combine the functions of disconnecting and earthing in
________________________________
one device. Three-position disconnectors are typical for gas-insulated switchgear.
Switch-disconnects (Load-break switches): The combination of a switch and a
________________________________
disconnector, or a switch with isolating distance.
Contactors: Load-breaking devices with a limited short-circuit making or breaking
________________________________
capacity. They are used for high switching rates.
Earthing switches: To earth isolated circuits.
________________________________
Make-proof earthing switches (Earthing switches with making capacity): Used for
the safe earthing of circuits, even if voltage is present, that is; also used in the event the
________________________________
circuit to be earthed was accidentally not isolated.
Fuses: Consist of a fuse-base and a fuse-link. With the fuse-base, an isolating distance can
________________________________
be established when the fuse-link is pulled out in a de-energised condition (as in a
disconnector). The fuse-link is used for one single breaking of a short-circuit.
________________________________
Surge arresters: Discharge loads caused by lightning strikes (external overvoltages),
switching operations and earth faults (internal overvoltages). They protect the connected
________________________________
equipment against impermissibly high voltages.
________________________________
Medium Voltage Switchgear
Typically fed directly from the utility and usually marks the service entrance to the
________________________________
building.
Generally located in the electrical space of large-capacity Data Centres (i.e., greater
________________________________
than 1 MW IT load).
If a MV generator is present, it also feeds the MV switchgear.
________________________________
Other than simply distributing power, the MV switchgear is responsible for
disconnecting faults and controlling the MV power distribution system. For example,
________________________________
when isolating a redundant section for maintenance.
MV switchgear typically includes meters, breakers, contactors, fuses, surge arresters,
________________________________
earthing switches for IEC equipment, voltage / current transformers, control and
protection relays, and an overall control system.
A________________________________
medium voltage switchgear should also be capable of the following:
________________________________
Normal ON / OFF switching operation
Short-circuit electric current interruption
________________________________
Switching of capacitive currents
________________________________
Switching of inductive currents
Some special applications
________________________________
All the above-mentioned functions must be carried out with a high degree of safety
________________________________
and reliability.
________________________________
Power must be shut down to the load for any maintenance performed on the Data
Centre equipment.
________________________________
One of the more common LV configurations is the secondary selective, or MTM. For
maximum reliability, draw-out power circuit breakers with automatic transfer should
be________________________________
used. The automatic transfer feature switches to the alternate source in three
seconds, rather than the 20 minutes it may take to manually transfer.
________________________________
Unavailability is reduced more than 10 times compared to the single-ended radial value.
________________________________
The transformer and MV conductor have long repair times, and the MTM allows them
to be automatically bypassed. Still, loss of the bus, one of the bus breakers, or the
________________________________
feeder conductor to the load, requires the load to be down until repairs are
completed.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Electrical load like motors, transformers, cooling units, etc. require REACTIVE power
to set up the magnetic field, while the ACTIVE power produces the useful work. Total
________________________________
Power is the vector sum of the two and represents what you pay for, i.e., kVA service
capacity charge. Power Factor Correction (PFC) equipment provides the means of
________________________________
reducing the reactive power being supplied by the utility. Reducing the reactive power
supplied by the utility results in lower electrical bills, since the kVA demand is also
________________________________
reduced.
An automatic PFC unit consists of a number of capacitors, switched by means of
________________________________
contacts. These contactors are controlled by a regulator that measures power factor in
________________________________
an electrical network. Depending on the load and power factor of the network, the
power factor controller will switch the necessary blocks of capacitors in steps to make
________________________________
sure the power factor stays above a selected value.
Ensuring a good Power Factor has numerous energy efficiency benefits:
►________________________________
Environmental benefit, reduction of power consumption by improved energy
efficiency
►________________________________
Reduced power consumption means less greenhouse gas emissions and fossil fuel
depletion by power stations
►________________________________
Reduction of electricity bills
► Extra kVA available from the existing supply
►________________________________
Reduction of losses in transformers and distribution equipment
► Reduction of voltage drop in long cable runs
►________________________________
Extended equipment life
► Reduced electrical burden on cables and electrical components
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
As well as transferring the load to the backup generator, a transfer switch may
command the backup generator to start, based on the voltage monitored on the
________________________________
primary supply. The transfer switch isolates the backup generator from the electric
utility when the generator is on, providing temporary power. The control capability of a
________________________________
transfer switch may be manual-only, or a combination of automatic and manual. The
generator supplies power to the electric load, but is not connected to the electric
________________________________
utility lines. It is necessary to isolate the generator from the distribution system to
protect the generator from overload in powering loads beyond the house. (And for
________________________________
safety, as utility workers expect the lines to be dead.)
________________________________
When utility power returns for a minimum time, the switch will transfer the Data
Centre back to utility power and command the generator to turn off, after another
________________________________
specified amount of "cool down" time with no load on the generator.
A transfer switch can be set up to provide power only to critical circuits, or to entire
________________________________
electrical (sub) panels. Some transfer switches allow for load-shedding or prioritisation
of optional circuits, such as heating and cooling equipment. More complex emergency
________________________________
switchgear (used in large backup generator installations) permits soft loading, allowing
load to be smoothly transferred from the utility to the synchronised generators, and
________________________________
back; such installations are useful for reducing peak load demand from a utility.
________________________________
________________________________
________________________________
Static transfer switches shift power between two sources in less than a quarter
electrical cycle. These products are power transfer switches designed to move power
________________________________
between multiple sources without interrupting the critical load; they are designed for
mission-critical applications & Data Centre requirements.
A________________________________
static transfer switch uses power semiconductors, such as silicon-controlled
rectifiers (SCRs), to transfer a load between two sources. Because there are no
________________________________
mechanical moving parts, the transfer can be completed rapidly, perhaps within a
quarter-cycle of the power frequency. Static transfer switches can be used where a
________________________________
reliable and independent second source of power is available, and it is necessary to
________________________________
protect the load from even a few power frequency cycles of interruption time (or from
any surges / sags in the prime power source).
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Automatic Transfer Switches (ATS) provide a dependable and positive transfer to an
emergency power source in the event of normal power source failure. The switch
________________________________
returns the system to the normal source in an orderly fashion when power is restored.
In the event the emergency source fails, the switch will automatically return to the
________________________________
normal source when it becomes available.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The main power distribution subsystem generally includes transformation to a lower
voltage, feeders, and overcurrent protection. The main power distribution also includes
________________________________
the electrical supply to the Data Centre cooling equipment. The typical high-capacity
main distribution switchgear or switchboard will have draw-out, mounted circuit
________________________________
breakers. This improves flexibility, as each circuit breaker can be removed from service
without needing to shut the entire line-up off.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There are a number of ‘definitions’ of IT load that have been used, sometimes with no
real justification.
________________________________
Even so, a clear and consistent definition of what is included and excluded in IT load is
needed. For instance, should we include:
►________________________________
Networking equipment to the wide area network (WAN)
►________________________________
The operations centre equipment
► Security equipment in the Data Centre if it is IP-connected
►________________________________
Management systems
________________________________
Understanding the IT load is a key factor that needs to be fully understood to properly
evaluate the power distribution required to support the Data Centre (IT load and
________________________________
supporting environments).
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The NCPI is provided to meet the full IT design capacity, and then some! All of this
oversized NCPI equipment requires that power be provisioned for it. It is an additional
________________________________
burden, which results in oversizing, affecting efficiency. The cooling capacity must also
include IT hardware plus these additional loads which, again, increase power provision.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The historical practice (still widely used by many Data Centres) of using nameplate
values to estimate power resource requirements has, in many cases produced highly
________________________________
unreliable results, including substantial undersizing (miscalculating by derating too far)
or oversizing (using the label as-is without understanding hardware options and
________________________________
equipment utilisation) of electrical and cooling systems. In a typical example, a server
rated at 475W uses 142W at start-up and 133W under normal use, only 30% and 25%
________________________________
of its nameplate rating, respectively. Therefore, calculating power requirements on the
nameplate ratings significantly bloats the requirements.
________________________________
These miscalculations can, and do, increase the possibility of service loss, plus excessive
________________________________
cost of operations. The more accurately power consumption data can be calculated,
the more efficient and reliable the electrical distribution systems will become. The
________________________________
accuracy of the calculations will assist with improving service availability and
optimisation of operational costs. Data Centres are now challenging vendors to supply
________________________________
more accurate power consumption data; manufacturers are also working with
governments to improve the efficiency of equipment.
________________________________
Finally, do not lose sight of the fact these miscalculations have a considerable cascading
effect upon the entire physical infrastructure of the Data Centre. More power, more
________________________________
cooling, and more capital expenditure.
________________________________
“Time to move forward and become more accurate”
________________________________
________________________________
Due to nameplate rating factors and design calculators, many assumptions are built into
the design criteria of the power infrastructure. In many cases, this results in excessive
________________________________
levels of capacity, often exceeding the actual business demands. However, due to poor
design or infrastructure limitations, excessive power capacity cannot actually be
________________________________
distributed to expand the Data Centre capability, and subsequently becomes stranded
power.This has the knock-on effect of additional cost and reduced efficiency.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The power requirement calculated does not give any insight into the Data Centre
loadings. It is very much an average of averages, and is based on empirical assumptions
________________________________
(in the absence of anything better). Whilst it may give some indication of the power
required at some of the blocks in the diagram, at some future point in time, it does not
________________________________
help very much in the detailed electrical design.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Codes and regulations require that electrical systems are split into circuits, so
distribution of power can be carried out safely.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
In a circuit supplying a load, we are concerned with voltage and current.
1. The voltage presented to the load must be within code or regulatory limits.
2.________________________________
The circuit must be capable of continuously carrying the design current.
3.________________________________
The circuit shall be protected by some device to limit and disconnect the current
flowing in the event of a fault.This is usually a fuse or circuit breaker.
________________________________
________________________________
Ib = Circuit design current
In = Rating of protective device
Iz________________________________
= Is the current carrying capacity of the circuit
________________________________
The relationship among circuit design current (Ib), nominal rating of protective device
(In), and effective current-carrying capacity of conductor (Iz) for an electrical circuit
________________________________
can be expressed as:
Ib<= In< =Iz
________________________________
________________________________
________________________________
________________________________
________________________________
As well as the three phase conductors, an electrical system has a neutral and an earth.
This arrangement shows a separate earth and neutral back to the source. This is called
________________________________
Terra Neutral-Separate (TN-S). This is the preferred earthing system for Data Centres,
as defined by Standards such as BS 7671, EN50310 and IEC 60364 for in-building
________________________________
earthing systems. It requires a neutral conductor, a separate earth conductor and a live
conductor.
________________________________
NB: These conductors have a variety of different names in Regulations and Codes, but
________________________________
the terms “live”,“neutral” and “earth” are generally well-understood.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The diagram above shows the main component parts of a Data Centre in a schematic
layout.
________________________________
This sort of diagram is useful for planning the electrical network - working out power
routes, cable sizes, protective devices, etc.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The copper loss of every main circuit connecting a distribution transformer and the
main incoming circuit breaker of a LV switchboard should be minimised by either:
(a)________________________________
Locating the transformer room and the main switch room immediately adjacent to,
or on top of each other
(b)________________________________
Restricting its copper loss to not exceed 0.5% of the total active power
________________________________
transmitted along the circuit conductors at rated circuit current
The cross-sectional area of neutral conductors should not be less than that of the
________________________________
corresponding phase conductors.
In any electrical circuit, some electrical energy is lost as heat which, if not kept within
________________________________
safe limits, may impair the performance and safety of the system. This energy (copper)
loss, which also represents a financial loss over a period of time, is proportional to the
________________________________
effective resistance of the conductor, the square of the current flowing through it, and
________________________________
the duration of operational time. A low conductor resistance therefore means a low
energy loss.
________________________________
The length of the main distribution circuit conductors connecting the distribution
transformer and the main incoming circuit breaker (MICB) of the LV switchboard
________________________________
should be as short as possible, by means of locating the substation and the main LV
switch room adjacent to each other. A maximum conductor length of 20m (66ft) is
________________________________
recommended. Due to the possibility of large, triple-harmonic currents existing in the
neutral conductor for Data Centre loads (with a large proportion of non-linear
________________________________
equipment), it is not recommended to use neutral conductors with a cross-sectional
area less than that of phase conductors in the main circuit.
________________________________
Sub-mains circuits could be UPS to PDU, for example.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Guidelines exist identifying the total harmonic distortion (THD) limits for HVAC units
according to their size:
________________________________
Current Maximum THD
________________________________
I<40A
40A<I<400A 15.0%
20.0%
________________________________
400A<I<800A 12.0%
________________________________
800A<I<2000A
I>2000A
8.0%
5.0%
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Unless specified otherwise, feeder circuit calculations are normally based on the
assumption that the supply voltages and load currents are sinusoidal and balanced
________________________________
among the three phases in a 3-phase, 4-wire power distribution system.
Care must be taken if the 3-phase feeder circuit is connected to nonlinear loads such
as________________________________
UPS systems, variable-voltage, variable-frequency (VVVF) systems and Variable Speed
Drive (VSD) motor systems.
________________________________
The design current used for cable sizing must take harmonic currents into account.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Traditional PDUs, RPPs, and RPDUs are located in the IT space to distribute, control,
and monitor critical power from the upstream UPS system to IT racks.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The floor-mounted and rack-mounted PDUs in a Data Centre can be more
sophisticated, providing data for power usage effectiveness (PUE) calculations.
A________________________________
floor-mounted PDU, sometimes called a main distribution unit (MDU), provides an
important management bridge between a building's primary power and various
________________________________
equipment cabinets within a Data Centre, or network operations centre (NOC). Each
PDU can handle larger amounts of energy than an ordinary power strip (300 kVA and
________________________________
higher, depending on the manufacturer and model), and typically provides power to
multiple equipment cabinets.
________________________________
A cabinet-mountable PDU installs directly in an equipment cabinet so it can control
________________________________
and monitor power to specific servers, switches and other Data Centre devices,
assisting to balance power loads. Cabinet-mountable PDUs are known by several
________________________________
different names, including smart PDUs and intelligent PDUs.
Such PDUs include three-phase displays for devices sharing power, as well as remote
________________________________
management tools that use the Simple Network Management Protocol (SNMP) to
provide administrators with the ability to adjust and monitor power demands from
________________________________
offsite locations.
The most basic large power strip without surge protection is sometimes (incorrectly)
________________________________
called a PDU. It is designed to provide standard electrical outlets for Data Centre
________________________________
equipment, and has no monitoring or remote access capabilities.
________________________________
________________________________
The device is basically a transformerless PDU that meets a particular form factor
requirement. Generally equipped with standard panelboards, this cabinet
________________________________
accommodates branch circuit breakers. Like its transformer-based relative, it can be
configured with single or dual-source feeds to accommodate an A-B bus and dual-
________________________________
corded downstream equipment. RPPs may contain up to four panelboards and a
monitoring system.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
In panelboard distribution, the main Data Centre power is distributed to multiple wall-
mounted panelboards. In general, a single branch circuit panelboard has 42 pole
________________________________
positions, which can accommodate 42 single-phase branch circuit breakers.
Panelboards (typically rated from 1.5 kVA to 75 kVA) may come assembled by the
________________________________
vendor, or they may be assembled in the field by the electrical contractor. Individual
branch circuits or power cables are cut, terminated, and connected in the field by the
________________________________
contractor.
________________________________
Panelboard Advantages
Lowest first cost, primarily driven by lower-cost components
________________________________
Accommodates unusual physical room constraints
________________________________
Flexibility in breaker and cable combinations (not preconfigured assemblies)
________________________________
Parts can be obtained very fast, i.e., from a local electrical supply
Disadvantages
________________________________
Increased risk of human error, since installations are custom-engineered
________________________________
If the air plenum is used, cabling restricts air flow, impacting cooling
________________________________
Creates an environment that is not readily changeable
Tracing cables is difficult, due to power cables under the floor (or in trays)
________________________________
________________________________
Final circuits could be the cable run to the cabinet socket, or to the cabinet power
strip.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Modular Distribution Bustrack can also be used. This is not new, and has been used in
the industry for many years. It is a simple and versatile solution for supplying power
________________________________
and can be quickly tapped at any location, with a variety of plug-in units, to allow for
moves, adds and changes in a Data Centre.
________________________________
Bustrack generally has some form of polarising method to prevent mismatched
components from being inadvertently connected to each other, and to prevent cross-
________________________________
phasing during installation. Not all systems are compatible.
________________________________
The calculation of current capacity is straightforward, but the voltage drop is slightly
more complicated, as it is a distributed load, depending on the equipment point of
________________________________
connection.
Onsite placement involves installing sub-feed breakers, then securing the bustrack in
________________________________
place by attaching it with threaded rod and Unistrut to the ceiling; wire is run from
breakers to the bustrack feed units, then plug-in outlet boxes are inserted, routing
________________________________
them to the racks. The power bus is generally installed over IT equipment rows - MDB
is generally installed upfront for the maximum expected load.
________________________________
Bustrack usually includes:
________________________________
The feed unit connected with an upstream LV electrical switchboard
The power bus
________________________________
Plug-in units equipped with overcurrent protection devices
________________________________
Connection fittings and their accessories
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Cabinet Power Strips are packaged in a standard ICT equipment cabinet form factor,
so they can be located either in the same rack / cabinet, or in-row, close to the IT
________________________________
equipment racks. Cabinet Power Strips are typically cord-connected and available in a
variety of form factors. Some are cabinet-mounted and take up 1-4U of horizontal
________________________________
space; others mount outside of the U area within a cabinet (for example, mounted
vertically in the rear of the rack or cabinet).
________________________________
Cabinet Power Strips are selected (three-phase or single-phase) based on the
expected cabinet power density and / or system configuration. The diameter of the
________________________________
cable determines its "ampacity," or the number of amperes it can safely carry (and its
________________________________
cost). The voltage used will determine how much power can be delivered at different
voltages over the same conductor size.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A chart like this can be a handy reference.
The CDCDP can do something like this using the appropriate voltages and volt drop
________________________________
limits allowed in his / her local codes and regulations.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A chart like this can be a handy reference.
The CDCDP can do something like this using the appropriate voltages and volt drop
________________________________
limits allowed in his / her local codes and regulations.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There are some codes of practice that give guidance to electrical engineers on how to
design energy-efficient electrical networks.
________________________________
These may become applicable to Data Centres in the future.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Why Infrared Scanning? Increased resistance and heat are the primary reasons most
electrical components fail. Before electrical components fail, they heat up. Infrared
________________________________
scanning is used to inspect electrical equipment, because excess heat is usually the
first sign of trouble. Loose connections, imbalanced and / or overloaded circuits,
________________________________
defective breakers, damaged switches, faulty fuses, and material defects all lead to
equipment failure. Using infrared scanning as an early warning tool for potential
________________________________
electrical equipment / system failure is an important part of maintaining mission-
critical uptime.
________________________________
Infrared Scanning Responsibilities - It is good practice to scan an electrical panel
________________________________
whenever accessed. Anyone authorised to work in an electrical panel or high voltage
cabling needs to be trained on infrared scanning. If any hotspots are found, or any
________________________________
components are outside of normal operating temperatures, an electrician or
facilities supervisor needs to troubleshoot the problem.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Higher voltage, DC power systems may involve fewer components, which can result in
higher reliability and lower total cost of ownership when compared to AC power
________________________________
systems. However, there are some drawbacks with DC power systems that need to be
noted:
________________________________
Bigger cables
________________________________
Shorter distances
Working practices
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Consider these facts:
The greatest risks of data loss are preventable, or can be highly mitigated. Only 3% of
________________________________
data loss incidents are caused by a natural disaster, according to Price Waterhouse
research.The following issues account for the majority of destructive influences:
________________________________
Software error (14%)
________________________________
Human error (32%)
Hardware failure (44%), often triggered by power problems
________________________________
________________________________
After a power outage disrupts IT systems:
More than 33% of companies require more than one day to recover
________________________________
10% of companies take more than one week
________________________________
It can take up to 48 hours to reconfigure a network
It can take days or weeks to re-enter lost data
________________________________
________________________________
________________________________
________________________________
________________________________
An uninterruptible power supply, (UPS) or battery / flywheel backup, is an
electrical apparatus that provides emergency power to a load when the input power
________________________________
source (typically the utility mains) fails. A UPS will provide instantaneous or near-
instantaneous protection from input power interruptions by means of one or more
________________________________
attached batteries and associated electronic circuitry. The on-battery runtime of most
uninterruptible power sources is relatively short - 5 to 30 minutes being typical for
________________________________
smaller units, but sufficient to allow time to bring an auxiliary power source online, or
to properly shut down the protected equipment.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The following factors should be taken into account when specifying a UPS:
Ride-through capability is the ability of a power source to deliver usable power for a
________________________________
limited time during a power loss.
Hold up time is the length of time that the power supply used in an application can still
________________________________
provide a viable output without an input supply.
Transfer time is the time taken by the UPS to switch from mains to the battery backup in
________________________________
case of mains failure or from battery power to mains power when power is restored.
Power handling is the normal power consumption of the application which the UPS must
be able to supply during emergency periods.
________________________________
Inrush current is the momentary peak input current required by electrical equipment
when it is initially switched on, due to charging inductive and capacitive loads.
________________________________
Backup time is the time period for which the UPS is required to support the load in case
of mains failure. The required backup time depends on the user's system design.
________________________________
Power conditioning (or line conditioning) is the provision of a regulated voltage and
frequency power line with a pure sine wave, free from electrical noise and ripple.
________________________________
Isolation is the complete electrical separation of the UPS output from the input power
line.
________________________________
Orderly shutdown is the process of sequentially closing down hardware and / or
software processes in the application, so no corruption of data occurs and / or no safety
________________________________
hazards are created.
Load-shedding is the ability to selectively shut off power to less critical loads during an
________________________________
extended power failure, while maintaining power to the more critical loads, thus extending
the effective backup time of the UPS.
________________________________
Load sequencing is the ability to selectively turn on or off specific loads in a pre-set
pattern during start-up and / or shutdown.
Recharge time is the length of time a UPS takes to recharge batteries after a discharge.
Hot-swapping is changing equipment while power is being supplied and used.
________________________________
UPS Input Switchgear
The UPS input switchboard feeds the UPS from the power control center of the
________________________________
upstream LV switchgear / switchboard.
________________________________
The UPS output switchboard not only provides power from the UPS output to the
downstream circuits, it also consists of static bypass and maintenance bypass circuit
________________________________
breakers, to allow utility power to clear faults or isolate the UPS for maintenance. In
some cases, isolation transformers are installed in the input or output switchboard.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A power semiconductor is an electronic device, consisting of two layers of silicon
wafer, with different impurities forming a junction made by diffusion. The joining of
________________________________
these two wafers provides control of current flow.
The power semiconductor permits current to flow in one direction from the anode
________________________________
to the cathode, whenever the anode voltage is positive, relative to the cathode.
When the anode voltage is negative relative to the cathode, the power diode blocks
________________________________
the flow of current from the cathode to the anode.
________________________________
The power semiconductors may be either Silicon Controlled Rectifiers (SCR) or
transistors; these are bipolar transistors, field effect transistors (FET}, and insulated
________________________________
gate bipolar transistors (IGBT). The IGBTs are significantly more efficient and easier
to control than the other power semiconductors. The use of IGBTs has allowed for
________________________________
static UPS as large as 750 kVA without paralleling units.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The SCR operates in two modes:
ON-state (forward-conduction, low-impedance)
________________________________
OFF-state (open-circuit, almost infinite impedance)
________________________________
While turning on the SCR is very efficient, the SCR requires a commutation circuit to
turn it off. The turn-off time is slow in comparison to the transistors, which are non-
________________________________
latching devices. Other drawbacks to the commutation circuit include the introduction
of more equipment to the circuit, additional noise, and power consumed.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Control of voltage and current is achieved by adjusting the pulse width as required by
load changes. The effects of these load changes are sensed and fed to the control logic
________________________________
circuitry through feedback circuits, comprised of voltage and current sensing devices,
thus adding design complexity.
________________________________
The switching frequency of this type of inverter is between 30 KHz and 50 KHz, and
due to the fast switching speeds, turning off the device is crucial.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Best practice requires all UPS to comply with the latest recommendations.
Most modern UPS are capable of appropriate power factor correction, or harmonic
________________________________
filtering to improve the PF to a minimum of 0.85, restricting the total harmonic
distortion (THD) of input current.
________________________________
Battery-backed UPS systems will occasionally be required to charge a fully-discharged
________________________________
battery. It is typical to assume a 20% rating for this.
Dedicated feeder circuits should be provided for individual UPS plants, to facilitate
________________________________
separate metering and monitoring of energy consumption for future energy
management and auditing purposes.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Some UPS have automatic voltage stabilisation (AVS), often transformer-based,
between the mains input and the transfer switch. This keeps the output voltage within
________________________________
a tight tolerance band. This feature eliminates the inverter ‘kicking in’ and discharges
batteries at low mains voltage, increasing system reliability and battery lifetime.
________________________________
They are used in small computer room installations, but are not generally suitable for
critical IT equipment in a large Data Centre. However, they can be used for essential
________________________________
systems, such as cooling.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Online systems comprise two conversion stages. The first rectifies the incoming AC power to
DC, thus creating a DC bus, which keeps the battery charged in normal operation. The
________________________________
inverter then converts the DC power back to AC in order to feed the critical load. This
isolates the IT load from the incoming supply and removes the effects of voltage and frequency
________________________________
fluctuations.
There is no interruption during switchover. This makes it ideal for Data Centre loads.
________________________________
However, there are losses in both conversion stages. The bypass is provided to enable
maintenance on the UPS system, or can backup the UPS in case it malfunctions. The transfer
switch can automatically transfer the load to bypass in a UPS overload or failure.
________________________________
In normal operation, load energy comes from mains, via the rectifier and the inverter. Upon
mains failure, the battery supplies the energy, which the load requires (by the DC / AC
________________________________
Inverter). Should the mains voltage return before the battery is fully discharged, the rectifier
will feed the load through the inverter and start a battery recharge regime, to compensate for
________________________________
the lost charge. Otherwise, the UPS will turn off when the battery becomes fully discharged.
The UPS will automatically resume normal operation upon mains restoration. Additionally, the
________________________________
rectifier will feed the load via the inverter, recharging the battery. The output voltage of an
Online UPS is generally stabilised within 1% tolerance. Output frequency can be set by a free-
________________________________
running, crystal-controlled clock, so the online system can act as a frequency converter.
The merits of the double-conversion online system have made it the preferred choice, in spite
________________________________
of the capital and electricity costs.
► They protect against all types of mains disturbances.
________________________________
► They are available in a wide range of sizes.
► Batteries can be added to increase the protected time.
________________________________
► Units can be modular and connected in redundant configuration for resilience.
► They have other features - e.g., they can work from generators and support a range of
power factor loads.
________________________________
The delta converter serves a dual purpose:
► To control the input power characteristics - the active front-end draws sinusoidal
________________________________
power, minimising reflected harmonics; this ensures optimal utility and generator
system compatibility, reducing heating and system wear.
►________________________________
To control input current in order to regulate charging of the battery.
________________________________
The output characteristics are the same as the double-conversion UPS.
However, the input characteristics are often different.
►________________________________
Delta conversion designs provide dynamically-controlled, power factor-corrected
________________________________
input, without the inefficient use of filter banks associated with traditional solutions.
► During steady-state conditions, the delta converter allows the UPS to deliver power
________________________________
to the load with much greater efficiency than the double-conversion design.
► The input power control makes the UPS compatible with all generator sets and
________________________________
reduces the need for oversizing.
________________________________
________________________________
________________________________
________________________________
________________________________
Three modes of UPS operating as defined by IEC 602040-3:
► VFI:Voltage and frequency of the output is independent of the input voltage
________________________________
► Double-conversion UPS
________________________________
► Highest level of power conditioning
► Efficiencies up to 96%
________________________________
► VFD:Voltage and frequency of the output is dependent on the input voltage
________________________________
► Offline UPS
► No power correction (Volts in = Volts out)
________________________________
► Efficiencies up to 99%
►________________________________
VI:Voltage of the output is independent of the input voltage
► Line-interactive UPS
________________________________
► Corrects sags, harmonics, input PF and swells, but not frequency
________________________________
► Efficiencies of 96% to 98%
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Transformer-based UPS tend to be used for loads with ‘bad’ characteristics such as
motors. Modern IT equipment is ‘good’ and a Data Centre UPS would often (but not
________________________________
always) be transformerless. The UPS can either be zoned so that areas of the Data
Centre have separate UPS systems, can be run in parallel operation, or in a
________________________________
combination of both modes.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
UPS are often specified with an output power factor displacement graph like the one
shown in the slide.
________________________________
Remember that Real Power (W) = Apparent power (VA) x power factor (PF), so at a
PF of 1.0, 100VA produces 100W, and at a PF of 0.8, 100VA produces 80W.
________________________________
UPS have traditionally been designed to support loads with unity and lagging power
________________________________
factors. Provided the wattage and voltage ratings are not exceeded, the UPS operates
normally and will try to satisfy the output load. Beyond that, it will realise it is
________________________________
becoming overloaded and react.
Leading power factors are a relatively new challenge for UPS and these need to be
________________________________
treated carefully, as they can overload the UPS; the UPS may not react to it, because it
may not detect the leading PF, so a 0.9 leading PF could result in a permanent 11%
________________________________
overload without the UPS control circuitry being aware.
________________________________
One typical way of dealing with this is to size the UPS so that it can service a specified
load across a range of power factors without becoming unstable.
________________________________
This graph shows a 100kVA UPS suitable for a range of output PF loads without any
reduction in power (in this case 80kW) from +0.9 to -0.8. In this case, it is obviously
________________________________
less robust on the leading PF side than on the lagging PF side.
It would, of course, be perfectly possible to have a similar graph for a 100kVA UPS
________________________________
designed to supply 90kW between +0.9 and -0.9, trading kW for PF flexibility.
________________________________
________________________________
There are a number of desirable characteristics for any UPS.
The ability to accept a wide range of inputs is critical if they are to be used with
________________________________
generators.
________________________________
The ability to accept a range of output power factors is determined by the
characteristics of the connected load. A UPS supplying an IT load (which is fairly linear)
________________________________
at PF 0.9 has a different challenge to one supplying non-linear fans and pumps, with PF
0.7 and five times the inrush current on start-up.
________________________________
The ability to accept an unbalanced load is a factor in a Data Centre with mainly single-
phase loads, though every effort will be made to minimise this.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Resilience in a Data Centre is important, so the capability for parallel operation may be
a key selection factor.
________________________________
Dynamic stability is the faculty to ride through dramatic swings in the load. In a Data
Centre, this may not be a major factor – as an example, if fans were connected and
________________________________
they all started together, it should not cause the UPS to have a problem.
________________________________
Crest factor enables the UPS to handle non-sinusoidal loads. This is not a major factor
in modern IT equipment.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Harmonics caused by the IT load used to be a problem due to the power supplies
used. In IT equipment, modern switched mode power supplies are much better and the
________________________________
problem has diminished.
Harmonics generated by the rectifier stage of a static UPS are still present, but 6 and
________________________________
12-pole IGBT rectifiers and good filters have reduced this significantly.
________________________________
No-break static switches are important in critical applications, and a variety is generally
available.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Large Data Centres = >100 racks / >200kVA UPS - Typically centralised
Midsize Data Centres = 20 to 100 racks / 60 to 200kVA UPS - Typically zoned,
________________________________
allowing computer room growth
________________________________
Small Data Centres = 5 to 20 racks / 10 to 60kVA UPS - Typically rack-mounted,
provides ultimate scalability or high-density areas
________________________________
Rack power density:
►________________________________
Low density – 1 to 4kW per rack
► Medium density – 5 to 9kW per rack
________________________________
► High density – 10 to 15kW per rack
►________________________________
Ultra high density - >15kW per rack
________________________________
Sizing and resilience:
Example of an N=500kW Data Centre using 120kVA UPS modules, optimised between
PF________________________________
0.9 lead and 0.9 lag
________________________________
Resilience UPS kW UPS kVA
N = =
________________________________
N+1 = =
2N________________________________
= =
2N+1 = =
________________________________
Not all UPS are available in the same sizes.
Typical transformerless UPS are up to around 200kVA, transformer-based ones up to
________________________________
around 800kVA. However, they can be paralleled together, providing manufacturer
guidelines are strictly followed.
________________________________
Different manufacturers place different limits on the number of UPS units that can
________________________________
operate in parallel.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
System scalability - A key benefit of modular UPS, making them future-proof with
significant commercial benefits. Modularity gives the flexibility to right-size the
________________________________
system to the critical IT load.
Adaptability and flexibility - Means to meet changing business demands and
________________________________
provide the flexibility to meet future growth expectations.
________________________________
Hot and safe swap out - The ability of UPS systems to tolerate the live insertion
and removal of UPS modules, rather than having to bypass or shutdown, exposing
________________________________
the critical load to the raw AC mains supply.
Energy efficiency - Significantly less energy compared to a standalone unit; the
________________________________
reduced carbon footprint offered is tremendous, and helps move toward cleaner
and greener power.
________________________________
Reduced maintenance impact - Allows maintenance to be undertaken without
________________________________
directly impacting the operational capability of the UPS.
Closer alignment to the IT load - Reduction of poor load losses.
________________________________
Reduced MTTR - Mean time to repair (MTTR) is reduced significantly in a
modular UPS system; some modules can be easily swapped in less than half an hour,
________________________________
which makes its availability 99.9999%.
________________________________
ROI and reduced TCO - Offers multiple savings in terms of operational and
infrastructure costs, spares and maintenance.
________________________________
________________________________
Continually monitor battery energy
Inspect and test every battery monthly
________________________________
Many industry best practices recommend replacing batteries every 3-5 years
________________________________
Conditions That Affect Battery Life:
________________________________
Temperature, as well as the depth and frequency of battery discharges.
________________________________
The "shelf life" of batteries decreases with higher ambient temperature.
Particularly important if the UPS is to be stored without periodically recharging the
________________________________
batteries. For example:
________________________________
At a 40ºC (104ºF) ambient, battery life is reduced approximately 70% from
what would be expected at 25ºC (77ºF).
________________________________
Even when the UPS is in use and the batteries are fully charged, higher ambient
conditions will reduce "float" service life.
________________________________
________________________________
________________________________
________________________________
________________________________
High-rate batteries are designed to deliver a large amount of current over a
short amount of time (approximately 15 minutes); this is achieved by designing the
________________________________
batteries with thin plates.This design is most common for UPS applications.
Medium-rate batteries are designed for general use, and deliver medium current
________________________________
over a period of approximately 1 to 3 hours. The design consists of medium-width
plates; this design is most common with switchgear and control applications.
________________________________
Low-rate batteries are designed for delivery of power over a long amount of
________________________________
time (approximately 8 hours). The battery design consists of thicker plates, and is
most common for applications, such as emergency lighting and telecommunications.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A well-designed charger will act to replenish a discharged battery by maintaining the
correct balance between overcharging and undercharging, so as not to damage the
________________________________
battery; it must assure battery discharging is limited to the point where the cells
approach exhaustion, or where the voltage falls below a useful level (usually about 80 %
________________________________
of rated capacity).
________________________________
Overcharging results in increased water use and over-discharging tends to raise the
temperature, which may cause permanent damage if done frequently. Safety margin is
________________________________
provided on the charger to take care of any overload condition at the inverter output.
After an emergency condition, when the charger starts charging the battery bank, a
________________________________
heavy current will be drawn by the battery bank. This should be taken care of by the
current-limiting circuit of the charger.
In________________________________
static UPS systems, the battery is continually connected to the charger, float-
charging the load and battery. During float charging, the battery charger maintains a
________________________________
constant DC voltage that feeds enough current through the cells, while supplying the
________________________________
continuous load; this replenishes local losses, and replaces discharge losses taken by
load pulses exceeding the charger's current rating.
________________________________
Periodically, the charger voltage is set at a level 10 percent higher than the floating
voltage to restore equal state-of-charge at the individual (lead-acid) cells. This mode of
________________________________
charging is called "equalising charge".
________________________________
Following battery discharge, the charger is set at this higher voltage to drive a higher
charging current that more quickly replenishes the battery, restoring it to a fully-
charged state.
________________________________
UPS Static Bypass
Should the inverter fail, the static bypass ensures the load drops automatically onto
________________________________
the mains input feed.
________________________________
With an Online UPS, if an inverter fails, it will (in all probability) occur whilst main
power is present.
________________________________
The static bypass, depending on the UPS manufacturer, will allow about 90-95% of
preventative maintenance to be performed on the UPS module without interruption to
________________________________
the critical load.
________________________________
Most (but not all) repairs can be accomplished with an internal maintenance bypass.
Not all UPS manufacturers include this type of redundant bypass source in their UPS
________________________________
module designs.
A________________________________
static bypass is normally part of the UPS internal circuitry, but may be invoked
manually using an external switch.
In________________________________
larger UPS, it will synchronise UPS output with the mains cycle before switching, and
so may take a few seconds to engage.
________________________________
________________________________
________________________________
________________________________
On larger UPS, the internal static bypass should be engaged before throwing the
maintenance bypass. Some bypasses have interlocks to ensure this correct operation.
________________________________
After setting the UPS into bypass mode (to synchronise with the mains supply),
SW1 is operated to provide raw mains to the load.
________________________________
SW2 may be opened to allow safe working on the UPS, or may be left closed if the
________________________________
engineer wishes to make tests with the power on.
The Maintenance Bypass Cabinet permits transfer of connected loads to an alternate
________________________________
power path, allowing full isolation of the UPS. The UPS can then be turned “OFF” and
removed from service, with no interruption of power to connected loads.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The true definition of a rotary UPS is any UPS whose output sine wave is the result of
rotating generation. Therefore, the UPS in the figure above (although it utilises a
________________________________
flywheel as a rotating, temporary energy storage source in case the utility fails), is not,
by definition, a rotary UPS.
________________________________
The flywheel is coupled to the M-G shaft and supplies stored energy to drive the
generator upon momentary loss of the motor output. The flywheel inertia is selected
________________________________
such that the stored energy is sufficient to supply the generator, while operating at
________________________________
rated power for a duration not exceeding 0.5 second; while keeping the speed from
falling, it maintains the frequency drop to a maximum of 0.5 Hz.
In________________________________
addition, it acts to stabilise the generator frequency by maintaining the rotational
speed, following transient frequency variations at the motor power supply, or sudden
________________________________
load changes.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Technical Features:
Pure sinusoidal output waveform
________________________________
High short-circuit capacity
Unlimited crest factor and no derating at leading power factors
________________________________
Self-ventilating
No filters or power capacitors
________________________________
Commercial Benefits:
________________________________
Exceptionally high levels of reliability
20-year service life
________________________________
Efficiencies up to 97% (with energy store connected)
Low Total Cost Ownership (TCO)
________________________________
Reduced operational / reputational risk
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
How does it work?
In the heart of the unit is its kinetic energy accumulator, a clever but simple system
________________________________
to store and retrieve kinetic energy. The kinetic energy accumulator consists of two
rotating parts: The outer rotor runs mechanically-free around the inner rotor. The
________________________________
inner rotor, driven by the main shaft, rotates at 1,500 rpm (50 Hz) or 1,800 rpm (60
Hz). It contains two sets of windings - a three-phase,AC winding and a DC winding.
________________________________
________________________________
Energy storage
In conditioning mode, the AC winding is powered to generate a rotating magnetic
________________________________
field, which turns the outer rotor to the speed of maximum 3,000 RPM. However,
________________________________
the relative speed between the inner and outer ring of the bearings is only 1,500
RPM, since the shaft itself is rotating at 1,500 RPM. The outer rotor stores kinetic
________________________________
energy advantageously, since the amount increases quadratically with the distance
from the centre of rotation.
________________________________
Energy retrieval
________________________________
In independent mode, the DC winding is powered and the outer rotor is electrically
________________________________
coupled to the inner rotor by induction. Its kinetic energy is transferred to the
inner rotor, which drives it. This energy transfer is regulated by accurately
________________________________
controlling the current injected into the DC winding. The kinetic energy
accumulator is a totally brushless system.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
BS ISO 8528-1 defines various classifications for the application, rating and
performance of generating sets, consisting of a reciprocating internal combustion (RIC)
________________________________
engine, alternating current (AC) generator and any associated control gear, switchgear
and auxiliary equipment. There are three principal ratings, as defined in the Standard:
________________________________
Emergency standby, Prime and Continuous.
► Emergency standby: The maximum power for which an engine-generator is
________________________________
capable of delivering, for up to 200 hours per year. The allowable average power
output over a 24-hour run period is 70% of the standby rating, unless otherwise
________________________________
agreed by the manufacturer.
►________________________________
Prime power: The maximum power for which an engine-generator is capable of
delivering continuously with a variable load, for an unlimited number of hours. The
________________________________
allowable average power output over a 24 hour run period is 70% of the prime
rating, unless otherwise agreed by the manufacturer.
►________________________________
Continuous power: The maximum power for which an engine-generator is
capable of delivering continuously, for a constant load over an unlimited number of
________________________________
hours.
________________________________
________________________________
________________________________
________________________________
Most Data Centres have standby generator sets, usually diesel powered.
It is necessary to maintain the generator so it will start within a few seconds when
________________________________
required. This requires regular maintenance, and also means the engine needs to be
kept in a state of readiness.
As________________________________
with a diesel engine in a car or truck, if the engine gets cold, it can be difficult to
________________________________
start; if the fuel gets cold, it can start to freeze. Heaters maintain the temperature of
the generator systems.There are various recommendations and common practices:
________________________________
► Although generators may be rated to work in low ambient temperatures, it is
typical to maintain the environment above 5ºC
________________________________
► It is important to ensure the entire generator is kept above dewpoint
________________________________
temperature
► Water jackets are used to keep the generator warm when it is not running -
________________________________
these are typically maintained around 21ºC (70ºF)
► Block heaters are used to heat the engine block, allowing combustion to take
________________________________
place at a faster rate on initial start-up. These circulate coolant around the
cylinders and maintain the temperature between 40 ºC and 55ºC (100 ºF and
________________________________
120ºF)
________________________________
These heaters can require several kW of power.
________________________________
________________________________
Generators are an integral part of business continuity planning; but while companies
invest large amounts of capital expenditure (CAPEX) in this type of equipment, often
________________________________
generators fail to properly start when needed.
The best practice of high-quality Data Centres is to test the full load of the Data
________________________________
Centre itself, which tests all systems including transfer systems, and not an external,
artificial, load bank. For the highest level Data Centre operations, all generators should
be________________________________
tested weekly, and at least monthly with full load. It’s not enough to simply test your
________________________________
generators and keep them fueled.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
If a genset is located at high altitude (especially over 1,000ft), or exposed to higher-
than-rated ambient temperature, it must be derated. Specific derating factors must be
________________________________
obtained from the supplier.
When the genset must be located outdoors, you should consider all the elements that
________________________________
could cause problems - wind, ice, snow, flooding, lightning, fire, earthquakes and
vandalism. A variety of measures can be taken to minimise these problems. For
________________________________
example, where flooding is a possibility, install the genset on a raised platform or
concrete mounting base. In locations where temperatures fall below 50ºF, special
________________________________
accessories such as an electric water jacket heater or manifold heater will be needed
________________________________
to ensure dependable starting. Use of heaters is also recommended if the genset is
installed in a humid location, to prevent moisture collection in the generator windings.
________________________________
Where exposed to excessive sun, an enclosure painted white will help keep ambient
temperatures down.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A genset must be mounted on a substantial and level base which supports the weight
of the engine-generator set. It maintains correct alignment between engine and
________________________________
generator, and isolates any vibration produced.
The foundation may be a large concrete mass or simply a fabricated steel base,
________________________________
depending upon the size of the genset and the type of soil or supporting material.
Proper preparation includes excavation, setting of forms, reinforcing of structural steel
________________________________
(where the unit is installed on an upper floor), vibration isolation, anchor-bolt
positioning, and shimming and grouting.
________________________________
Smaller gensets have rubber vibration isolators located between the engine and
________________________________
generator assemblies, and the skid.
Steel spring vibration isolators are normally provided on engine-generator sets rated
________________________________
from about 150kW up. These can reduce up to 98% of the vibration produced. When
additional vibration reduction is required, heavy-duty isolators and / or specially-
________________________________
designed structural components can be provided, such as ‘floating’ floors or
foundations that rest on rubberised,‘hockey-puck’ isolators.
________________________________
________________________________
________________________________
________________________________
________________________________
Adequate ventilation and cooling must be provided. Ensure openings are provided and
located so cool clean air can be brought in, forced through the engine cooling system,
________________________________
and directed out without recirculating.
Adequate volume of air, based on the genset size, is required. For example, a 350kW
________________________________
unit may require as much as 15m per second (33,000cfm) of air when running.
3
________________________________
In most instances, an engine-mounted radiator and fan provide genset cooling. If fan
noise is excessive for a nearby area, a remote fan and radiator can be used. However,
________________________________
the further away a remote radiator and fan are located, the higher the installation and
maintenance costs.
________________________________
One metre (3ft) must be allowed around the genset for service (as required by
regulations) and to assure free flow of cooling air.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Exhaust fumes are poisonous, so great care is needed when installing the exhaust
system.
________________________________
Exhaust pipes should be wrought iron, cast iron or steel, and should be freestanding,
not supported by the engine or muffler. Where exhaust pipes attach to the engine, they
________________________________
must be connected with flexible connectors to minimise vibrations that can cause
damage to the exhaust system. The exhaust pipe should terminate outdoors, away from
________________________________
doors, windows, or other building openings.
________________________________
Never connect the genset exhaust system to an exhaust system serving other
equipment. Outdoor gensets housed inside weather-protective enclosures should have
________________________________
their exhaust directed away from buildings and their air intakes.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Local codes and regulations vary from location-to-location. In Europe, 2000/14/EC
provides a guide.A 10dB rise above normal levels can lead to complaints.
________________________________
Noise comes from four sources:The engine, generator, cooling fan and exhaust.
________________________________
Noise from the engine and generator can be reduced by providing baffles, or a
sealed enclosure lined with sound absorption material.
________________________________
The major source of noise is the engine exhaust. A variety of mufflers are available - in
most instances, a critical-grade muffler on the exhaust piping is most effective.
________________________________
However, it is also the most expensive.
The direction of the muffler discharge is an important factor in noise reduction.
________________________________
Pointing the exhaust pipe straight up is the most effective direction, though a rain cap
is then required to keep moisture out of the system. Another approach is to point the
________________________________
terminating exhaust pipes away from any nearby buildings or locations to be protected
________________________________
from noise.
When reduction of noise is vital, it is best to bring in an engineer or firm specialising in
________________________________
sound reduction.
________________________________
________________________________
________________________________
________________________________
The load acceptance is important for UPS compatibility. A generator and UPS system
can never be matched on a 1:1 ratio, because a UPS is not 100% efficient. A PPR rating
________________________________
is better suited to a UPS.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
These factors affect the generator size.
A PPR level two (80%) would require the generator to have a 125% UPS rating.
________________________________
Other loads and harmonics can double that.
________________________________
Generator performance reduces as the temperature rises, either because the
generator is in a plant room or a hotter climate. This can also cause damage - in such
________________________________
circumstances, it is normal to derate the generator.
Typically, however, it is the size of the maximum load that has to be connected, and the
________________________________
acceptable maximum voltage drop which determines the set size, not the ratings
themselves. If the set is required to start motors, then the set will have to be at least
________________________________
three times the largest motor, which is normally started first. This means it will be
________________________________
unlikely to operate at anywhere near the ratings of the chosen set.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Generators in parallel can be achieved fairly readily, and there are two common
methods in use.
________________________________
More than two generators can be synchronised, but this requires more complex
controls.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Wet stacking occurs when carbon or unburned fuel oil accumulates on the injectors,
on the exhaust valves, or in the exhaust system. Permanent damage due to cylinder
________________________________
scarring is possible, although unlikely.
Fortunately, the condition is usually corrected rather easily by running the engine for a
________________________________
few hours under sufficient load (typically 30% to 40% of rated load) to bring the engine
to nominal operating temperature and burn off the deposits. Load banks may be
________________________________
required to achieve this if normal building load is not sufficient.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Fuel storage is an issue to consider. Day tanks may be fitted to the generator giving
about eight hours running, or bulk tanks may be needed for longer periods. Health and
________________________________
safety regulations commonly require a double-bunded tank (tank within a tank) to
capture any fuel spillage.
________________________________
Fuel suppliers may have a tier system for priority service. Some countries or states
require a triple tank, which is tested every five years and also has double wall piping.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It is necessary to maintain the fuel in good condition, or it will deteriorate and the
generator may not start when required.
________________________________
Instead of washing the oil, some Data Centres carry out a full system test and switch
over to the generators every six months, using all their fuel. This may be viewed as
________________________________
‘brave’ or ‘sensible’, depending on your viewpoint.
________________________________
A well-designed and operated Data Centre may consider this option, as it proves the
standby power system will work when truly needed, but the carbon commitment may
________________________________
influence this.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Why use load banks?
The only way to verify a backup power system will perform during an outage is to
________________________________
periodically test it under load. Both generators and UPS may appear to run well when
lightly loaded, but may fail to deliver full load power if they are not regularly tested to
________________________________
ensure they are up to the task.
________________________________
However, unlike the “real” load, which is likely to be dispersed, unpredictable and
random in value, a load bank provides a contained, organised and fully controllable load.
________________________________
Where the “real” load is served by the power source and uses the energy output of
the source for some productive purpose, the load bank serves the power source, using
________________________________
its energy output to test, support or protect the power source.
Understanding load bank specification, installation, and operational issues can help
________________________________
ensure electrical and mechanical system reliability. Anywhere there are critical standby
or life safety electrical systems, a load bank may be desirable (or even mandated by
________________________________
code) to help ensure the reliability of those systems.
Although the concept of a load bank as an artificial electrical load is relatively
________________________________
straightforward, consideration must be given to the detailed needs of the site when
load banks are specified and installed.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
As the average power density of today’s modern IT equipment continues to climb, it
has become a major challenge for many Data Centres, as they struggle to match this
________________________________
increase with sufficient and precise cooling; this, in turn, results in additional costs and
potential risk to uptime.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A survey carried out on a number of Data Centres in the US revealed that when
cooling systems are inefficient, not only do they use more energy, but the environment
________________________________
in the computer room deteriorates. This seems at first to be counterintuitive, but is
covered in more detail in the Data Centre Efficiency module.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There are a number of major contributing factors, most of which can be seen to result
from incorrect design. Some are (of course) operational issues, but a good design and
________________________________
controlled handover process can alleviate many of these at the outset. Operators
should not have to live with bad design issues.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Data Centres must be able to effectively measure and monitor the cooling system to
understand its actual capability to meet both present and future demands. To put it
________________________________
simply:“If you cannot measure it, how can you manage it?”
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
With the continual advances in technology and increased business demands to remain
competitive, the four physical constraints (power, cooling, space and IT connectivity) of
________________________________
many Data Centres are coming under intense pressure. For many, cooling is becoming
the leading concern, also impacting some of the other physical constraints.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Note: The European Union’s (EU) current Fluorinated Gas Regulations (F-GAS) are
designed to minimise environmental damage from refrigerant gases.
________________________________
Fluorinated gases (F gases) are powerful greenhouse gases that contribute to global
warming if released into the atmosphere. Their effect can be much greater than carbon
________________________________
dioxide. F gases currently make up about 2% of the UK’s annual greenhouse gas
emissions.
________________________________
Hydrofluorocarbons (HFCs) are the most common type of F gases, and are mainly
________________________________
used as refrigerant in HVAC and commercial refrigeration systems.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
As with any element of managing a Data Centre, you need to assemble accurate
information from all the respective environments, and be able to analyse it before
________________________________
deciding what needs to change (this is especially relevant to the cooling system). This is
also true of calculating a Data Centre’s Carbon Footprint. There are many carbon
________________________________
footprint calculators available, but they can provide differing results, as few actually use
the baseline assumptions on lbs/Kg, CO2/kW or CO2 emissions per energy distance
________________________________
travelled.
Knowing your carbon footprint calculation is merely the starting point. Priorities must
be________________________________
baselined against operational expectations and actual delivery capabilities before a
________________________________
roadmap can be formulated to implement actions to successfully reduce carbon
emissions.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There are many challenges facing the CDCDP when planning cooling systems for a
given Data Centre. It needs to be flexible and scalable, with appropriate redundant
________________________________
cooling elements to ensure a consistent performance. Adaptability is also a major
challenge for the projected lifecycle of the cooling system, especially supporting
________________________________
constant IT advances.
Maintenance and servicing programmes also need to be addressed during the design
________________________________
process to ensure they can be undertaken without affecting the operational balance of
the cooling system.
________________________________
These factors drive the choices, which will need to be compared with an impact
________________________________
analysis of each option on the total cost of ownership (TCO), inevitably a major
business driver.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There are various elements that need to be considered; these depend on the nature of
the design work being undertaken:
________________________________
Existing extension:
________________________________
Understand the limits of the existing Data Centre
Implement all best practices
________________________________
Consider enhanced airflow segregation
________________________________
Add liquid / local cooling
Major retrofit, or existing building conversion:
________________________________
Understand the baseline capabilities of the building
________________________________
Perform detailed engineering analysis and TCO evaluation of various options, air and
liquid
________________________________
Greenfield site:
Understand rack power values - peak and average
________________________________
Plan for high-density if possible
________________________________
Consider air-side or water-side economisers
________________________________
Implement air cooling where possible for lowest cost to build and operate
________________________________
________________________________
Rack Cooling documentation and analysis:
Identify airflow obstructions within, above, and below the rack environment.
________________________________
Identify racks where air mixing is likely to occur, and where blanking panels should
________________________________
be installed.
Examine rack enclosures for airflow suitability.
________________________________
Determine the airflow pattern (i.e., front-to-back, side-to-side) of racked and non-
racked equipment.
________________________________
Detect and document areas where rack inlet temperatures exceed industry
________________________________
Standards and guidelines, using infrared thermography.
Analyse rack arrangements for air distribution effectiveness.
________________________________
Measure the airflow and temperature of raised-floor air distribution system and
________________________________
overhead grilles, if accessible.
Locate significant obstructions under the raised floor and in the drop ceiling that
________________________________
affect airflow. This will be collected onsite or by using customer-provided mechanical
drawings.
________________________________
Determine areas of significant air leakage in the Data Centre.
________________________________
________________________________
________________________________
Redundant equipment levels are essential to compensate for equipment failure, and
provide an opportunity for maintenance of the normal load.
________________________________
Actual redundancy design levels are usually based upon anecdotal evidence, as well as
experience of end-users, IT Managers, Facilities Management and engineers; OR, by the
________________________________
basic assumption that for every kW drawn by the IT load, another kW of power will
be required to remove the heat generated.
________________________________
That said, there is a fine balance between underspecifying the cooling requirement, or
________________________________
drastically over calculating it. It is essential to review the following:
► The projected IT load (Day 1 and Day 2 onwards)
►________________________________
Consider the impact of future IT advances
►________________________________
Review cooling system options
► Design flexibility and adaptability into the planned cooling system
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There are a number of areas where the CDCDP can influence performance and
efficiency of the Data Centre cooling systems. The list above is not exhaustive, but all
________________________________
of the items MUST be incorporated at the design stage.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It is essential that management strategies do not overlook the fact that good working
practices and maintenance schedules can also bolster the integrity of the cooling
________________________________
systems on a daily basis. Poor installation practices, coupled with a lack of maintenance
can (and will) lead to significant and unnecessary losses of pressure and cooling
________________________________
capabilities throughout the Data Centre. The following are basic examples that highlight
this fact:
►________________________________
Failure to decommission and remove redundant servers
________________________________
► Neglecting the hot and cold aisle principle (classic example - air grille tiles placed in
the hot aisle)
►________________________________
Abandonment of redundant underfloor cabling, reducing the effectiveness of the
plenum
►________________________________
Blanking panels not being fitted in cabinets
►________________________________
Cabinet floor entries not being sealed properly (brushes and grommets)
► Open cable cut-outs in the hot aisle
►________________________________
Cable routes blocking airflow to equipment within cabinets
►________________________________
Forgetting to firestop and seal wall or floor penetrations
► Poorly maintained raised floors, causing unacceptable air gaps
________________________________
Individually, they may have a limited impact, but combined, they will have a significant
________________________________
bearing upon Data Centre cooling efficiency, with unnecessary waste increasing cost.
________________________________
Ideal opportunity to assemble all key stakeholders to assess cooling capability against
the present and foreseeable operational demands.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
NEBS Standards were introduced in the USA by Telcordia and have become the de
facto standard for Telcos. ETSI Standards are used in Europe by Telcos.
________________________________
These Standards are end-user centric, i.e., they consider typical environmental
conditions in equipment rooms, and specify IT equipment which must ensure reliable
________________________________
operation.
________________________________
It is then down to IT manufacturers to design equipment that meets specifications.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
ASHRAE produced the document that is commonly used as a reference Standard for
Data Centres - the Thermal Guidelines for Data Processing Environments handbook.
________________________________
This adopts many of the concepts used in NEBS GR-63 and GR-2038, and aims to:
________________________________
Establish standard operating environments for Data Centres
Define a common interface for the equipment and its surroundings
________________________________
Provide guidance on how to test the operational health of the Data Centre
________________________________
Provide a methodology for reporting
However, it is written from an equipment manufacturer’s perspective, whereby
________________________________
equipment manufacturers state their environmental specifications for ensuring reliable
________________________________
operation. It is now down to the end‐user to design environmental systems to meet
those specifications.
________________________________
CDCDPs generally accept the recommendations of ASHRAE and use them as a basis
for their designs. The operators generally accept them and build them into their
________________________________
operating processes.
________________________________
________________________________
________________________________
________________________________
There are two ranges defined in ASHRAE:
► The recommended range is where equipment is designed to operate reliably, i.e., it
________________________________
will maintain its meantime before failure (MTBF).
________________________________
► The allowable range is where the manufacturers have verified IT equipment will
continue to work, but it may shorten the MTBF if extended periods are spent within
________________________________
it.
Outside of the allowable range, equipment may no longer continue to work.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
We are concerned with the operating environment of the IT equipment.
The conditions within the IT equipment are managed by the manufacturer, who defines
________________________________
the environment required at the air intake. We should measure as closely as possible to
this point, to minimise the effects of any local recirculation within the cabinet.
________________________________
The temperature in the middle of the aisle, or the return temperature to the
________________________________
computer room air conditioner (CRAC) has only a very loose relationship to rack
cooling effectiveness. The CDCDP should be alert though - measuring in these
________________________________
positions is still recommended in the current version of several Standards. In historical
environments with poor (or no) separation of hot and cold air, measuring the ambient
________________________________
temperature made some sense, insofar as there was little else to measure; however, it
makes no sense in a well-designed Data Centre.
________________________________
________________________________
Definitions:
CRAH: Computer room air handling unit – includes only fans and a cooling coil, often
________________________________
using chilled water to remove heat from the Data Centre.
CRAC: Computer room air conditioner – includes an internal compressor using
________________________________
direct expansion of refrigerant to remove heat from the Data Centre
________________________________
________________________________
________________________________
There are many different types of equipment found in a large Data Centre, and the
CDCDP must be aware of all of them. Their environmental requirements may be
________________________________
significantly different, so the CDCDP must account for this in the design.
The recommendations from ASHRAE, NEBS and ETSI are (to some degree) a
________________________________
consensus taking into account the weakest links in the equipment chain, such as the
most sensitive equipment.
________________________________
PDH/SDH: Plesiochronous digital hierarchy / Synchronous digital hierarchy
________________________________
PBX: Private branch exchange
RAID: Redundant array of independent disks
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
ASHRAE Standards have, up until very recently (2008), recommended very tight
parameters within the computer room. Some of these have been extended to align
________________________________
with NEBS / ETSI as better knowledge of Data Centres has been gained.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The allowable ranges are very different though, the NEBS/ETSI range being well beyond
the ASHRAE range.
________________________________
The question of how well the IT equipment measures up to these ranges can be seen
on the following psychrometric chart.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Recommended: 18-27°C = 64-81°F
A1 Class: 15-32°C = 59-90°F
A2________________________________
Class: 10-35°C = 50-95°F
A3________________________________
Class: 5-40°C = 41-104°F
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Consideration points:
Most Data Centres still continue to operate within the ASHRAE Recommended
________________________________
Class envelope. Does anyone question why this is the case?
________________________________
ASHRAE has introduced the allowable envelope (A1-A4) that has wider tolerances.
Could this be applied to Data Centres?
________________________________
What about the IT equipment - what does it require? After all, this is the only
reason Data Centre cooling is used.
________________________________
Most manufacturers already make equipment that will operate across a wider
tolerance band.The example shows a standard ‘off the shelf’ Dell server.
________________________________
Other manufacturers will have similar capabilities. They also make ‘hardened’
________________________________
equipment that can extend the envelope even further, at a price.
The telecoms industry already operates to very wide tolerances, primarily due to
________________________________
the type of equipment being operated.
________________________________
________________________________
________________________________
________________________________
________________________________
As a note: Reducing cooling energy (raising intake parameters) could actually increase
the IT energy usage demand. For example, raising inlet temperature set-points allows a
________________________________
reduced burden on the cooling plant efficiency, but could potentially increase server
fan power draw with variable speed fans.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Rate of temperature change was quite limited in ASHRAE pre-2008 (5ºC per hour).
This has been relaxed in the 2008 revision as shown.
In________________________________
contrast, the NEBS / ETSI allowable rate is much greater; this specification is stricter
and results in more robust equipment.
________________________________
Note that these are rates of change, not absolute temperatures. For instance, 96°C
________________________________
per hour means the temperature can rise 16°C in 10 minutes - it does not mean the
temperature can rise by 96°C.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Using dew point as a humidity control parameter will allow separate control of latent
and sensible component control. The dew point should not exceed the temperature of
________________________________
the coldest surface in the Data Centre space. Another useful feature of controlling on
dew point is that it's easier than control based on relative humidity. Changes in dry
________________________________
bulb temperature across a space means relative humidity (RH) varies widely
throughout the building, which makes the system “hunt” to achieve control within a
________________________________
defined range.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Humidity is the measurement of moisture content within the air. A Data Centre that is
too humid will result in condensation building on computer components, causing
________________________________
possible short-outs. Also, high humidity can cause condensation to form on the coils of
a cooling unit, causing increased workload to rid itself of the condensation; this in turn
________________________________
can lead to wasted cooling (also called latent cooling) which costs money. Subsequently,
many Data Centres struggle to find the optimal humidity balance - in many cases, it is
________________________________
actually an afterthought.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There are a number of industry solutions available to combat humidification:
Steam humidifiers - In central air conditioning systems, steam humidifiers are
________________________________
often used in the AHU after the cooling coil, to reintroduce moisture lost through
condensation during cooling. This type of humidification offers close control and
________________________________
rapid response to a humidity requirement. However, due to the significant capacities
required in large Data Centres and continual 24x7 operation, energy consumption
________________________________
can be high. In hard water areas, there can also be a high cost associated with
humidifier maintenance and spare parts.
________________________________
Low energy - A low energy alternative is to use cold water evaporative or spray
________________________________
humidifiers. In order to evaporate the moisture into the airstream, the air needs to
be above a certain temperature. This often necessitates pre-heating, which negates
________________________________
some energy savings of cold water humidifiers. However, by using a heat recovery
system on heat generated in the data hall, this required pre-heating can be achieved.
________________________________
This results in a humidification system with 150 times less energy consumption of
steam humidifiers, and massive energy savings for the overall HVAC system.
________________________________
Evaporative cooling - As well as economic humidification, cold water humidifiers
can provide up to 12°C of evaporative cooling to the air being humidified. This
________________________________
adiabatic cooling can be applied directly to the incoming fresh air, or alternatively to
________________________________
the outgoing extract air. Outgoing air can have its humidity increased up to 100%
RH with the resulting drop in temperature. This chilled thermal energy is then
________________________________
transferred to the incoming supply air with a heat recovery system. Combining
these two techniques can reduce the reliance on mechanical cooling by up to 95%.
________________________________
There are now a wide range of sensor applications available, providing appropriate
monitoring and measuring capability across the elements of the Data Centre physical
________________________________
infrastructure.They are available in both fixed and wireless applications.
As the power density of IT hardware continues to increase with technology advances,
________________________________
deploying sensors is becoming more of a necessity; it allows accurate monitoring of
environmental conditions, and can identify areas of energy inefficiency.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Floor obstructions such as network and power cables will obstruct airflow, and have a
negative effect on the cooling supply to the racks, also affecting subfloor static
________________________________
pressure.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The data points can be presented graphically and summarised into two numbers.
However, the data points can be stored and analysed to show each occasion when
________________________________
excursions outside the compliant limits took place; these can be tracked back to
individual cabinets or sensors, to identify potential problems before a failure takes
________________________________
place.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
RTI can be presented as a single number. However, the data points can be stored and
analysed to show each occasion when excursions outside of the ‘desirable’ limits took
________________________________
place, and also the magnitude of the excursion. This can be useful if correlated with
other activity known to have taken place at the same time (i.e., installation or
________________________________
maintenance work).
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Air conditioning, as a basic principle, is simple:
► Heat is generated by equipment inside a room
________________________________
► The air conditioning system removes the heat and dumps it outside
________________________________
The cooling system should be able to remove all heat generated by the IT load, but the
efficiency decreases at partial loads. There may be just one, or there may be several
________________________________
stages in the cooling system.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The most important thing to understand in cooling theory is that cold is the absence
of heat. Heat can be removed in a cycle called the refrigeration cycle. The refrigeration
________________________________
cycle is present in all air conditioning systems, and is made up of four main
components:
________________________________
Refrigerant: Refrigerants are engineered chemicals developed to have very specific
boiling and condensation temperatures. They come in many different flavours, with
________________________________
cryptic names like 410a, R22, and water.
________________________________
Compressor: Compresses refrigerant, turning it from a warm liquid to a hot gas.
This compression assists in the movement of heat and refrigerant within the system.
________________________________
Evaporator: Evaporators are heat exchangers (devices built for efficient heat
transfer from one medium to another), so an evaporator passes heat from the air to
________________________________
the refrigerant.
________________________________
Condenser: Condensers are also heat exchangers. The condenser releases trapped
heat in refrigerant outside the space being cooled.
________________________________
This is a very simplified explanation of the refrigeration cycle components; only these
four components or steps have been mentioned, because they are common to all
________________________________
HVAC systems regardless of size or type, and you can apply them to any Data Centre
cooling system.
________________________________
________________________________
________________________________
Basic Refrigeration Cycle
The compressor pumps this gas from the evaporator, increases its pressure, and
________________________________
discharges the high-pressure, high-temperature gas to the condenser.
________________________________
In the condenser, heat is removed from the gas, which then condenses and
becomes a high-pressure, high-temperature liquid.
________________________________
Liquid in the condenser is cooled and reduced in pressure by passing it through an
expansion valve.This provides the input to the evaporator.
________________________________
In the evaporator, the low-pressure liquid expands, absorbs heat, and evaporates,
changing to a low-pressure, low temperature gas at the evaporator outlet, and the
________________________________
cycle repeats.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Air cooling in the traditional sense involves the use of computer room air conditioners
(CRACs) to convert warm air to cool air by removing heat to the outside. CRACs can
________________________________
be used in a number of basic configurations that focus on cooling the entire room, just
a row, or just a rack.
________________________________
A CRAC unit is exactly like the air conditioner at your house. It has a direct expansion
(DX) refrigeration cycle built into the unit. This means the compressors required to
________________________________
power the refrigeration cycle are also located within the unit. Cooling is accomplished
by blowing air over a cooling coil filled with refrigerant.
________________________________
A CRAC is typically constant-volume – therefore, it can only modulate on and off.
Recently, some manufacturers have developed CRAC units that can vary the airflow
________________________________
using multistage compressors, but most existing CRAC units have on / off control only.
Whole-room air conditioning situates CRACs such that a certain temperature is
________________________________
maintained fairly evenly throughout the room (in some sense, the same way you might
want to cool a room of your house).
________________________________
Owing to the inefficiency of mixing warm and cool air, whole-room cooling designs
have been refined to isolate warm air from cool air.
________________________________
________________________________
________________________________
________________________________
________________________________
Cooling is accomplished by blowing air over a cooling coil filled with chilled water.
Typically, chilled water is supplied to the CRAHs by a chilled water plant. Chilled water
________________________________
is pumped in pipes from the chiller to the CRAH units located in the IT environment.
CRAHs are similar to CRAC’ in appearance, but work differently: CRAH units cool the
air________________________________
(remove heat) by drawing warm air from the computer room through chilled water
coils filled with circulating chilled water.
________________________________
CRAHs can have VFDs that modulate fan speed to maintain a set static pressure, either
________________________________
underfloor or in overhead ducts.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
In the hot aisle - cold aisle model:
► The hot sides of adjacent rows face each other to create the hot aisle and
________________________________
similarly, the cold sides create the cold aisle.
________________________________
► Air conditioning units pump cold air into the floor void or into ceiling ducts.
► Cold air is delivered by vents in the floor (or dropped through diffusers) to the
________________________________
front of two lines of equipment cabinets facing each other – the ‘cold aisle’.
► Hot air escapes through the back of the cabinet into the ‘hot aisle’, where it
________________________________
returns to the air conditioning unit.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Using power density can lead to confusion as shown in the slide above. APC calculated
five scenarios for the same Data Centre which gave vastly different answers. The
________________________________
CDCDP must understand which parameters to discuss with the HVAC engineer.
Note: New standards now refer to W per m³ (volume)
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The cooling capacity of any system depends on the ability to transfer heat. As shown in
the equation above, the heat transfer depends upon the refrigerant / fluid used, the
________________________________
mass flow and the allowed temperature differential.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There are many manufacturer calculators and modelling tools available on the market
that can be utilised to evaluate energy requirements. They are useful and allow you to
________________________________
formulate a realistic design baseline.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Once the airflow has been determined, the amount of resistance to it must be
considered. This resistance to flow is referred to as system impedance, and is
________________________________
expressed in static pressure as a function of flow rate.
Obstructions in airflow will change the static pressure at different points throughout
________________________________
the system, which can significantly affect airflow at those points. Larger underfloor or
ceiling voids are better than tight ones.
________________________________
These factors can have a dramatic effect on the effectiveness and efficiency of the
________________________________
cooling system, and the CDCDP must ensure they are carefully considered and dealt
with.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Cabinets with greatly different loading make the problem even worse.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The CDCDP must understand the limitations of the traditional hot aisle - cold aisle,
room-cooled system.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The traditional approach has been to install cabinets, preferably in a hot aisle - cold
aisle arrangement, calculate average power density and cool the room. This gives a
________________________________
typical answer of 1.5kW per rack (2008 figures). However, this requires the cooling
requirement is spread fairly evenly throughout the computer room. Once certain
________________________________
pieces of IT equipment deviate too far from the average, then hot aisle - cold aisle and
room cooling starts to have problems.
In________________________________
order to deal with a range of requirements, the CDCDP needs to consider close-
coupling the heat extraction and the heat generation. The closest that could be
________________________________
achieved is to have individual cooling for each piece of IT equipment, but this is difficult
________________________________
and costly to achieve in practice. Falling back from that, we can have coolers dealing
with individual racks; stepping back again, we can group the racks in such a way that we
________________________________
can cool several of them with a row / zone cooler.
Going still further, we can install airflow management, in the form of a containment
________________________________
system - the CDCDP should consider this as the starting point for a new Data Centre
design for reasons of performance and efficiency.
________________________________
________________________________
________________________________
________________________________
________________________________
Bypass air: Conditioned air that does not reach computer equipment, escaping
through cable cut-outs, holes under cabinets, misplaced perforated tiles, or holes in
________________________________
the computer room perimeter walls - limits the precise delivery of cold air at the
server intake.
________________________________
Hot air recirculation: Where waste heat enters the cold aisle, requiring the
cooling infrastructure to supply colder air at the equipment to offset this mixing.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Air-side economisers are simple, effective systems that significantly reduce cooling
energy usage and improve indoor air quality. They increase HVAC system energy
________________________________
efficiency such that most energy codes require them in many commercial applications.
The air-side economiser is integrated into a central air handling system, with ducting
________________________________
for both intake and exhaust; its filters reduce the amount of particulate matter (or
contaminants) brought into the Data Centre.
________________________________
With Data Centres needing to be cooled 24 x 7, 365 days per year, air-side
economisers may even make sense in hot climates, where they can take advantage of
________________________________
cooler evening or winter air temperatures.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Air-side economisers have numerous benefits, as detailed in the slide. One often
overlooked is they can act as an emergency backup to save a Data Centre, if there is a
________________________________
loss of mechanical cooling.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
When assessing the implementation of air-side economisers, how many hours you
would be able to run on free-cooling will be crucial in the design. The Green Grid is
________________________________
offering tools on its website that help you get an indicative answer on available hours
of cold air, depending on the geographic region your Data Centre is located.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It is possible to use air-side economisers which introduce external cold air directly into
the cooling distribution system, effectively cutting out the rest of the cooling system.
In________________________________
the off mode, dampers are positioned so the hot air from the computer room is
directed through the terminal unit (CRAC / CRAH).
In________________________________
the on mode, dampers are positioned so cold air is brought in directly from outside
________________________________
through filters, and the hot air from the computer room is vented to the outside.
In reality, the dampers can be in an intermediate position, so it is not an “all or
________________________________
nothing” scenario.
A benefit that is often overlooked is that economisers can act as an emergency backup
if ________________________________
there is a loss of cooling plant.
________________________________
ASHRAE Standard 62 recommends the use of MERV 13 filters on units with outside air
economisers. Lawrence Berkeley National Laboratory has shown in field studies that
________________________________
units with MERV 7 filters have been able to keep dust at an acceptable level.
Bringing smoke into the Data Centre could cause the fire suppression system to
________________________________
trigger. Ultra-low leakage dampers and some form of early detection outside the
intakes is recommended.
________________________________
________________________________
MERV = Minimum Efficiency Reporting Value - It ranges from 1 to 16, and the higher
number, the finer the filter.
________________________________
________________________________
Historically, the subfloor cooling distribution with CRAC or CRAH units has been
used within Data Centre environments, whereas overhead cooling distribution with
________________________________
central air handlers have been used in telco environments. This perception is now
showing signs of closing, as new Data Centre designs are being built directly onto floor
________________________________
slab, alleviating the need to deploy a raised floor. The overhead delivery option has a
number of design considerations worth reviewing:
________________________________
►Provides for higher-density environments exceeding 5kW.
________________________________
►Direct supply to the cold aisle, reducing the level of bypass losses.
►Introduces a centralised supply system utilising a single air plenum, fed by fewer but
________________________________
larger air handler units.
________________________________
►Units have larger fans and motors which improve efficiency as they are scaled up.
►Air handlers can easily be located outside the compute space.
________________________________
►Supply ducts need to be sized significantly larger than a typical office delivery system.
________________________________
►Ducts need to be correctly positioned to maximise the potential of the return air
path.
________________________________
________________________________
________________________________
________________________________
It is possible to contain the cold air in a cold corridor. One such system is shown
above - it is becoming accepted that new Data Centre designs will incorporate some
________________________________
form of air containment.
Cold air containment can provide a number of environmental advantages:
►________________________________
Stops air mixing together - cold aisle temperatures will be consistent, allowing
________________________________
optimum aisle temperatures.
► Aisle will self-balance for temperature and airflow.
►________________________________
The hot aisle changes will give you improved return air temperatures to the CRAC
units and building chillers.
________________________________
► Minimises bypass airflow through racks and floor tiles.
►________________________________
Claims to reduce cost and downtime for maintenance and replacements.
________________________________
Cold air containment means the designer must consider issues such as health and
safety, as well as fire systems within the room.
________________________________
________________________________
________________________________
________________________________
________________________________
Rack and row cooling may not remove 100% of the heat, so there may be a
requirement for room cooling to remove the residual heat. They can however, remove
________________________________
the majority of the heat, so a high-power cluster looks like a low-power cluster as far
as the room cooling is concerned.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The HVAC engineer, like all engineers, has their own terminology. It is useful for the
CDCDP to understand some of this.
________________________________
DECS (Datacom Equipment Cooling System): An isolated loop within the rack that
transfers heat from the equipment to a heat exchanger within the rack.
________________________________
TCS (Technology Cooling System): A dedicated loop typically contained within the
________________________________
technology space to transfer heat to the chilled water cooling loop via CDU.
CDU (Coolant Distribution Unit): A heat exchanger that conditions the TCS loop
________________________________
by connecting to a central cooling plant.
________________________________
Liquid cooling refers to a number of different approaches that all share the same
characteristic of transferring heat to a fluid at or near the source, rather than
________________________________
transferring it to the room air.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Rack cooling can be used where needed. However, if the air exhausted at the rear of
the cabinet is cooled too much, it can be difficult to integrate into a standard hot aisle
________________________________
- cold aisle scenario.
Backdoor options can have integral fans, or can rely on equipment fans to make the air
________________________________
flow through the heat exchanger. They all introduce more fan energy into the facility,
possibly falsely increasing the measured IT load.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
In-row cooling attempts to recirculate hot air back into the cold aisle, as shown. It is
often used in conjunction with air containment. If they are to be deployed, the CDCDP
________________________________
must take into consideration that they will require floor space, and will vary in size
between suppliers. That said, there are a number of possible benefits created by
________________________________
utilising in-row coolers:
► No dependency upon a raised floor
________________________________
► Removes fighting / duelling between row and perimeter coolers
►________________________________
Simpler redundancy deployment
► Reduced capital expense
________________________________
► Lower energy costs from removing over-provisioned coolers
►________________________________
Provides additional capacity to cold / hot aisle containment
________________________________
► An increase in cooling capacity per unit that reduces the overall cooling footprint
► More effective capture of hot air enables a warmer supply temperature and
________________________________
reduces the humidification make-up requirement.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Water systems are not new, and can be very efficient. Traditionally, IT managers have
been very wary of introducing water into close proximity with the IT equipment.
________________________________
However, the volume of water required to remove a specified amount of heat is much
less than the equivalent volume of air, as the specific heats and densities are different.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Traditional refrigerant gases have been used, but these are being phased out, due to
environmental and health concerns. Global Warming Potential (GWP) of Ozone-
________________________________
Depleting Substances (ODS) is available from the United States Environmental
Protection Agency (EPA).
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
► Carbon dioxide is highly efficient
► Weight for weight, carbon dioxide absorbs 7.2 times more heat than water – if it
________________________________
changes phase during the process
________________________________
► 30-50kW cooling per cabinet is possible
► It needs additional care and management, which the CDCDP must evaluate if he /
________________________________
she is considering it
► High pressure exists in the system (300 – 400psi), so there are safety concerns
________________________________
► CO2 kept above 14ºC to prevent condensation
►________________________________
Intermediate coolers may be needed
► Additional pipework is required in the Data Centre
________________________________
► Specialised skills are needed to work on it
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Once the heat has been transferred from the equipment to the CRAC, CRAH or
CDU, the next step is to take it outside and ‘dump’ it.
We________________________________
all accept it would be better to reuse the heat rather than reject it to the
atmosphere; opportunities and innovative technologies will appear to allow this, but
________________________________
right now, the heat is purely a waste product in most Data Centres.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It is worth remembering that the cooling system is a series of loops, each transferring
heat to the next.This helps visualise what is happening at each stage.
________________________________
Also, remember not all loops need to be present, for example:
________________________________
► Introduction of free cooling removes the need for the middle three loops
► Introduction of a water-side economiser removes the need for the condenser loop
________________________________
What is the effect of raising the temperature to the IT equipment?
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The heat can be transported or removed in a maximum number of ways via the
________________________________
following steps:
►________________________________
Heat coming from the IT server components is transported into the air, which is
forced through the servers.
►________________________________
The heat transported by the air of the server equipment is transferred via heat
exchanger to a secondary water circuit, located in the IT room.
________________________________
► The heat removed by the secondary water circuit is transferred to a primary water
________________________________
circuit, which goes to the outdoor environment.
► If the outdoor temperature is too high, the primary water circuit needs to be
________________________________
cooled through an additional compressor circuit or via evaporative cooling.
________________________________
The upper cooling infrastructure in the above figure is the most extensive application
in________________________________
this comparison. Unfortunately in some Data Centres, the complexity goes even
further due to poor design considerations and deployment. The goal is to reduce the
________________________________
number of heat transfers, because each is accompanied by losses. This eventually
reduces the outdoor cooling capacity, requiring more work to be done by the
________________________________
compressors. A more ideal situation would be to directly utilise the outdoor
conditions, allowing outside air to enter the Data Centre directly when conditions are
________________________________
within range (see middle picture of the above figure). Ultimately, the most ideal
situation would contain no cooling at all (see lower left picture of above figure).
________________________________
However, this would require servers designed to withstand such conditions.
Selecting the best cooling infrastructure is key in saving money, providing a reliable and
________________________________
resilient environment, minimising the carbon footprint and providing good working
________________________________
conditions for people working in the Data Centre environment. When a Data Centre
is characterised, floor space in square meters (m²) is mentioned in relation to designed
________________________________
power consumption (kW); this relation is referred to as the heat density in kW/m².
Although the m² parameter is relevant in understanding site size and, therefore, the
________________________________
amount of cabinets and servers that can be housed, it is not considered the best
parameter for judging what type of cooling system should be applied. Another more
________________________________
valuable parameter in designing a cooling system is the available m³. If the space in
cubic meters is positioned against the required cooling capacity, a new dimension
________________________________
reveals the volumetric dimension. This metric is based on a physical law – the formula
of Bernoulli: It states the total amount of pressure remains the same on a streamline
________________________________
from one point to the other. However, we will not delve into the theory of Bernoulli’s
formula, as that is in the world of the HVAC engineer.
________________________________
________________________________
________________________________
________________________________
________________________________
Example of an available industry cooling calculator that could be used to establish the
________________________________
baseline requirement.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
This slide shows a dry cooler DX with a water and glycol loop. They are not very
efficient, as the dry cooler has low heat rejection. However, they can share a common
________________________________
cooler, and thus reduce the number of roof penetrations required compared to air-
cooled DX systems.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Definitions:
________________________________
The term ‘direct’ refers to the position of the evaporator with respect to the airside loop.
Whereas the term ‘expansion’ refers to the method used to introduce refrigerant into the
________________________________
cooling coil (the evaporator). This reduces the pressure and temperature of the refrigerant to
the point where it is cooler than the air passing through the coil.
DX________________________________
cooling systems also come as:
► Split DX system: The components of the refrigerant loop maybe split apart, allowing for
________________________________
increased flexibility in the system design.
________________________________
► Packaged DX system: The components of the DX unitary system refrigerant loop may
be packaged together, which provides for factory assembly and testing of all components,
________________________________
including the electrical wiring, the refrigerant piping and the controls.
________________________________
This slide shows a typical air-cooled CRAC unit. The compressor is in the evaporator (indoor)
unit and the condenser is usually located on the roof. There are a number of key issues that
________________________________
the CDCDP needs to be aware of:
►For each CRAC unit, there is a separate refrigerant loop to an associated condenser, leading
to________________________________
many roof penetrations for the pipe work and condenser unit mounts. Each penetration is a
potential leak.
►________________________________
The condensing units can take up a lot of the available roof space.
________________________________
► The number of condensers required leads to a high electrical load.
► A dark roof can increase the air entering the condensing units by as much as 5°C (9°F),
________________________________
which decreases the CRAC unit efficiency and capacity. Remember that capacity issues due
to cooling capacity are faced by many Data Centres.
► Access must be provided to the roof in order to service the units.
________________________________
Evaporative pre‐cooled DX is like air‐cooled DX, with either water sprayed over the
condensing unit coils, or wetted media placed on the air intakes of the condenser. The
________________________________
wetted media is preferred, as direct spray onto coils can cause fouling by plating of
dissolved solids, with airbourne dust sticking to the wetted coils. This reduces heat
________________________________
exchange, thereby decreasing efficiency and capacity. It can also cause the coils to
corrode away.
________________________________
Obviously it needs a water supply to each unit, so there is additional pipe work
required, and possibly a storage tank to be housed. It depends on water being able to
________________________________
evaporate into the ambient air, so it is not ideal in humid conditions. There is a risk of
________________________________
the coils freezing, so protection is required.
It is not a standard option from the manufacturers but there are a number of retrofit
________________________________
kits available; these can only be mounted on side inlets! It will probably become a
standard option in the near future.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The components of a water‐cooled CRAC or air conditioning unit system are shown
above. Many CRAC units can be associated with a single condenser line, so only a few
________________________________
penetrations are required; these can often be at low level rather than through the roof.
The condenser water lines in this system typically run at 16-32°C (60-90°F), so they
________________________________
don’t need to be insulated against freezing. However, if they are in a subfloor plenum
used for supply air distribution, it might be worth considering.
A________________________________
central heat exchanger at the cooling tower is generally recommended, as it is much
________________________________
easier to maintain the system.
In a medium to large Data Centre (>400KW IT load), this system can often be cheaper
________________________________
than multiple air-cooled CRACs, due to reduced need for electrical systems and
maintenance.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There are many advantages to chilled water systems (CRAHs or AHUs).
► Efficiency
►________________________________
Cost-effective on larger projects
►________________________________
Easier to get competitive pricing, as there are many custom unit manufacturers that
will compete with the standard CRAH manufacturers to produce units more
________________________________
efficient and usually less expensive.
► Easier to incorporate economisers with either air or water. With air-side
________________________________
economisers, there is no risk of freezing the coil as with DX. Water-side
economisers can be incorporated in the plant.
________________________________
► Require less maintenance and are quieter
►________________________________
Avoid the problem of roof penetrations
________________________________
There are some disadvantages:
►________________________________
Cost for smaller Data Centres
► Space for the chillers - although this can be less than air-cooled CRACs, outdoors
________________________________
for air and evaporative-cooled or indoors for water‐cooled
________________________________
► Room for insulated pipe to distribute chilled water to the CRAHs or AHUs
► Introduces water into the Data Centre floor. Leaks are rare and can be easily
________________________________
protected against, using sensors
► Added complexity to the controls
________________________________
Adiabatic humidification, designed to eliminate any risk of droplet carryover, is a
more efficient alternative. It is the process of injecting water into an incoming air
________________________________
stream, using natural convection or forced spray from nozzles. The exiting air has
both a lower dry-bulb temperature (due to evaporative cooling) and a greater
________________________________
moisture content.
Some characteristics of adiabatic cooling:
________________________________
► 6 bar water pressure at nozzles (fine mist)
________________________________
► Droplets 50-100µm to minimise risk of Legionella
► UV sterilisation treatment
________________________________
► Water kept moving, so stagnation is avoided
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Air at entering condition of 35°C dry-bulb (1), 21°C wet-bulb (2) at 28% relative
humidity (3), has moisture added by sprays (5). Evaporation of this moisture
________________________________
increases humidity and reduces dry-bulb temperature (4) to provide a lower
effective air onto the cooler (6)
________________________________
Adiabatic cooling refers to the humidification of air under adiabatic (isocaloric)
conditions, so that energy is neither added or removed. The heat necessary for
________________________________
evaporation is taken from the air, which subsequently cools down.
________________________________
Adiabatic cooling can be direct or indirect.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The cold water (CWS) and chilled water (CHW) plant consists of chillers, pumps,
cooling towers and economisers. In short, everything that provides chilled water to the
________________________________
cooling system. Simple?
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Not so simple! Typical equipment in a chilled water system is shown.
The variables that determine the performance and efficiency of each are identified.
________________________________
This is a very complex interaction, and optimising a chilled water plant is therefore
________________________________
very complicated – too complicated for us to go into here.
The key thing for the CDCDP is to make sure the cooling engineer has full knowledge
of________________________________
the likely operating conditions of the Data Centre, including part load operation, so
he / she can make the best selection to meet the requirements.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
How variable-speed drives work?
Many existing pump and fan systems are based on throttling arrangements; the motor
________________________________
is driven at full speed and the flow of liquid or gas is regulated by valves, vanes or
similar throttling mechanisms. Throttling the output in this way wastes energy. An AC
________________________________
drive can increase system efficiency by adjusting the motor speed to the correct
operation point, eliminating the need for throttling.
________________________________
A small reduction in speed can make a big difference in energy consumption. A
________________________________
centrifugal pump or fan running at 80% speed consumes only half as much energy as a
unit running at full-speed. This is because the power required to run a pump or fan
________________________________
changes with the cube of the speed.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
First, you should understand the terminology of chiller efficiency. Chiller efficiency is
________________________________
generally expressed in terms of kW per ton (kW / ton). A kW is a kilowatt of
________________________________
electrical input. A ton of cooling is equivalent to 12,000BTU of cooling per hour. More
efficient chillers will have lower kW / ton ratings, indicating they use less electricity to
________________________________
deliver the same amount of cooling.
You should also realise chillers may have a full-load (FL) rating and an Integrated Part
________________________________
Load Value (IPLV) rating. The IPLV is a weighted average of efficiency measurements at
various part-load conditions. It is a standardised way of comparing chillers at
________________________________
conditions more representative of field conditions. IPLV is preferred for situations with
variable loads.
________________________________
However, operators may wish to put more emphasis on full-load performance in
________________________________
installations with staged chillers. New chiller efficiency (IPLV) can vary widely, from as
low as 1.25kW / ton (standard efficiency, air-cooled screw compressor) to as high as
________________________________
0.38kW / ton (high-efficiency, water-cooled centrifugal compressor.) In other words,
energy operating cost can be more than 2½ times as high for the lower-efficiency
________________________________
chiller. There are three main types of chillers distinguished by their use of water or air
to reject heat:
________________________________
►Water-cooled chillers
________________________________
►Glycol-cooled chillers
► Air-cooled chillers
________________________________
Air-Cooled Chillers
________________________________
With air-cooled chillers, heat removed from the returning chilled water is rejected to a
________________________________
device called an air-cooled condenser that is typically integrated with the chiller.
This type of chiller is known as a packaged chiller, and can also be integrated into a cooling
________________________________
facility module. Air-cooled chillers are typically located outdoors.
________________________________
Glycol-Cooled Chillers
________________________________
Glycol-cooled chillers look identical to water-cooled chillers. With glycol-cooled chillers,
heat removed from the returning chilled water is rejected to a glycol loop for transport to
________________________________
the outside atmosphere.
The glycol flows via pipes to an outdoor device called a dry cooler, also known as a fluid
________________________________
cooler. Heat is rejected to the outside atmosphere, as fans force outdoor air through the
warm glycol-filled coil in the dry cooler.
________________________________
Glycol-cooled chillers are typically located indoors.
________________________________
Water-Cooled Chillers
________________________________
Heat removed from the returning chilled water is rejected to a condenser water loop for
transport to the outside atmosphere.
________________________________
The condenser water is then cooled using a cooling tower, the final step in rejecting the
heat to the outdoors.
________________________________
________________________________
► Chilled water pumping systems may involve primary and secondary loops, and can be quite
complex. Conventional primary / secondary pumping systems are starting to lose favour as
strategies to save energy. Ensure your system designer has evaluated the pros and cons of
________________________________
alternative chilled water pumping strategies.
________________________________
►Cooling range
►Approach to wet bulb temperature
________________________________
►Mass flow rate of water
________________________________
►Wet-bulb temperature
►Air velocity through tower or individual tower cell
________________________________
►Tower height
Other design characteristics to consider are fan horsepower, pump horsepower, make-up
________________________________
water source, fogging abatement, and drift eliminator.
________________________________
Operation considerations
► Water make-up: Water losses include evaporation, drift (water entrained in discharge
________________________________
vapour), and blow-down (water released to discard solids). Drift losses are estimated to be
between 0.1 and 0.2% of water supply.
►________________________________
Cold weather operation: Even during cold weather months, the plant engineer should
maintain the designed water flow rate and heat load in each cell of the cooling tower. If less
________________________________
water is needed due to temperature changes (i.e., the water is colder), one or more cells
should be turned off to maintain the design flow in the other cells. The water in the base of
________________________________
the tower should be maintained between 60-70°F by adjusting air volume if necessary.
Usual practice is to run the fans at half speed, or turn them off during colder months to
maintain this temperature range.
For Data Centres with water or air-cooled chilled water plants, a water-side
________________________________
economiser uses the evaporative cooling capacity of a cooling tower to produce
________________________________
chilled water, and can be used instead of the chiller during the winter months. Water-
side economisers offer cooling redundancy because they can provide chilled water if a
________________________________
chiller goes off-line. This can reduce the risk of Data Centre downtime. Water-side
economisers are best-suited in climates where the wet bulb temperature is lower than
________________________________
55°F for 3,000 hours or more per year.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A typical example of a water-side economiser schematic layout.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The efficiency of hot aisle containment (HAC) will be higher because it can maintain
higher temperatures, and removes recirculation; it can also work well in conjunction
________________________________
with increased operating parameters. The increased return air is efficiently captured
and routed back to the air-handling unit.
If ________________________________
your Data Centre is fitted with a ceiling plenum, the hot air return system might be
the best option to duct exhaust back to the air-handling units, and may actually fall into
a ________________________________
medium-term approach. However, if you have an open slab ceiling, the cost will
dramatically increase; a ducting system will need to be installed to channel the hot
________________________________
return air back to the air-handling units.
________________________________
There are a number of factors to be assessed before introducing HAC. The
environments will be considerably hotter and need tighter working considerations. The
________________________________
following factors have been identified by The Green Grid:
►Health and safety concerns for operating personnel
________________________________
►Airflow optimisation throughout the Data Centre to avoid hot-spots
________________________________
►Increased temperature in the hot aisle, which may extend beyond limits of cabling
►Increased temperature, humidity, and particulates effects
________________________________
►PUE anomalies, due to server fan power consumption at high temperatures
________________________________
________________________________
________________________________
It is possible to contain the hot air within a rack. One such system is shown above, and
the racks have a ‘chimney’ to take the hot air directly into the ceiling plenum.
In________________________________
this arrangement, the open room can be kept at comfortable temperatures for
humans. The racks shown here are deeper than normal in order to accommodate the
________________________________
chimney. There are, of course, other ways of achieving hot air containment, but the
principle is the same.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It is possible to use heat wheels, which introduce external cold air indirectly into the
HVAC distribution, effectively cutting out the rest of the cooling system.
________________________________
The Kyoto system has a number of advantages and is very efficient - it overcomes
many concerns about external contamination and humidity, as external air is not
________________________________
brought into the Data Centre.
________________________________
The CDCDP could consider this option for a new build, but must be aware of the
wheel size and the space required.
►________________________________
Wheel is 6m wide
► Weighs 11 tonnes
►________________________________
Moves at 3RPM
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
EcoBreeze from APC by Schneider Electric™ is an energy-efficient and innovative
cooling approach, employing a modular design for capacity, redundancy and service
flexibility.
________________________________
________________________________
EcoBreeze is installed where hot air from servers is controlled, through hot
aisle containment
________________________________
How it works:
► The hot air rises to the ceiling plenum and is fed into the EcoBreeze return ducting.
________________________________
The IT exhaust air is then cooled and supplied back to the Data Centre through a
raised floor plenum or cold aisle diffusers
►________________________________
The cool air is forced through the perforated tiles in front of the server racks and
is used to keep the servers at a safe operating temperature
________________________________
EcoBreeze installed where hot air from servers is controlled, through
________________________________
ducting connected to the racks
How it works:
►________________________________
The hot air rises to the ceiling plenum through ducting, and is fed into the hot air
return of the EcoBreeze
►________________________________
The IT exhaust air is then cooled and supplied back to the Data Centre through
cold aisle diffusers. The diffusers have a bowling alley-type discharge that sends the
________________________________
air down the cold aisle
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
When the Open Compute Project launched with its first Data Centre in Prineville,
Oregon, it selected a system that uses both outside air and the cooling power of water.
________________________________
Prineville is in a high-desert region, where rainfall averages just 10” per year and the
summer temperatures reach 80-90°F. The 147,000ft2 facility relies completely on
________________________________
outside air and a MeeFog™ cooling and humidification system to keep temperature
and humidity within the specified range.
A________________________________
MeeFog™ system was designed consisting of 56 x 7.5HP positive displacement fog
pump units with variable-frequency drives, each providing 7.62gpm @ 100 psi. Two
________________________________
pumps serve each of the Data Centre’s 28 x air handling units, one on active duty, the
________________________________
other on standby, with automatic switchover.
Designed for a PUE of 1.15, measured PUE of 1.06.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
In Iceotope’s approach, each server motherboard is completely immersed in a sealed
bath of liquid coolant; this passively transfers heat away from the electronics to an
________________________________
exchanger formed by the wall of the module, where water is continuously recirculated
by low-power pumps. The system is nearly silent and requires no cooling outside the
________________________________
cabinet, which (in theory) would allow Data Centre operators to eliminate expensive
room-level cooling schemes. Iceotope says its system uses just 80W of power to
________________________________
harvest the heat from up to 20kW of IT equipment.
________________________________
Source: Data Center Knowledge
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
transfers heat to water in the pump module, where it is made available for reuse. The
CarnotJet™ system can reliably create waste water at 50°C (122°F), a temperature
________________________________
suitable for reuse in almost every application.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Examples of Data Centres reusing heat:
An IBM Data Centre in Switzerland is being used to heat a nearby swimming pool.
________________________________
Hot air generated by the Uitikon Centre flows through heat exchangers to warm
water that will be pumped into the nearby community pool. IBM says the volume of
________________________________
heat thrown off by GIB-Services’ Data Centre is enough to warm 80 homes - or
one swimming pool, it would appear.
________________________________
Quebecor channels excess heat produced by servers at its Data Centre in
________________________________
Winnipeg, Canada to the nearby offices of a local newspaper. The company ran a
second duct out of the exhaust plenum to the intake duct of the editorial office
________________________________
upstairs. The process was controlled by pneumatic baffles that open and close,
depending on thermometer readings within the ducts.
________________________________
Excess heat from servers at the new Telehouse West Data Centre in the Docklands
section of London will soon be used in nearby houses and businesses. The waste
________________________________
heat from the $180 million Telehouse colocation facility will be used in a district
heat network, expected to produce up to nine megawatts of power for the local
________________________________
Dockland community
________________________________
Source: Data Center Knowledge
________________________________
________________________________
________________________________
► Commissioning must begin in the design phase, with the verification the scope
________________________________
documentation is in-line with the basis of the design.
► Commissioning verifies the installation meets design parameters.
________________________________
► Commissioning equipment and systems to assure they perform within
specifications and meet the design intent.
________________________________
► Commissioning guarantees the proper training of operational personnel.
________________________________
► Opportunity to identify any areas of weakness or future system improvements.
________________________________
The commissioning process can be broken down to a number of elements:
► Factory acceptance tests (Level 1) – Tests prior to leaving manufacturer.
________________________________
► Field component verification (Level 2) – Verification upon receipt of
products.
________________________________
► System construction verification (Level 3) – Field inspections to verify
________________________________
components are assembled and correctly integrated into the system.
► Site acceptance testing (Level 4) – Demonstrate system components
________________________________
function to the expectations of the design criteria.
► Integrated system tests (Level 5) – Testing of redundant and backup
________________________________
components to demonstrate they respond as predicted to expected and
unexpected circumstances.
________________________________
The definition of planned preventative maintenance (PM) is the systematic inspection
and identification of potential failures before they actually occur.
________________________________
________________________________
It involves establishing a comprehensive, planned preventative maintenance strategy
to ensure the cooling system remains operational, maximising efficiency within the
________________________________
design criteria. It should as a minimum cover:
► Identification of potential issues, undertaking immediate action to prevent a future
________________________________
failure.
► Implementation of a repair schedule to identified issues.
________________________________
► Ensure equipment is working within the designated manufacturer’s operating
________________________________
guidelines.
► Identification of defects, and system checks to prevent reoccurrences.
________________________________
► All equipment and sensors are fine-tuned and calibrated.
________________________________
► Performance reports are prepared and reviewed.
► Alignments undertaken to meet technological changes.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The evolution of the Data Centre has led to an ever-increasing demand for power in
today’s business environment, where 24/7 operation and complete resiliency are
________________________________
critical.
These demands have historically increased exponentially, and could continue to do so
________________________________
unless positive action is taken.
________________________________
However, secure power is still one of the key requirements for a Data Centre facility.
Rising energy costs, diminishing stocks of fossil fuels, carbon footprint and global
________________________________
warming are of great concern.
There should be a common motive for becoming more efficient in our energy usage.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Many Data Centres are power-constrained, but we can no longer rely on additional
power coming out of the grid (or the ability to buy more).
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Energy costs are significant - various studies show a wide range of costs.
A simpler approach is to work it out based on what we actually know:
________________________________
Facts:
________________________________
There are 24 x 365 = 8,760 hours in a year (Leap Years excepted)
Assumptions:
________________________________
We pay eight pence per unit (kWH)
________________________________
Calculation:
________________________________
Each kW used continuously costs £8,760 x 8/100 = £700 per annum
Conversely, each continuous kW saved will reduce the annual bill by £700.
________________________________
(You can substitute your own unit costs and currency here; the hour and year are de
________________________________
facto worldwide standards - the Jacobin attempt at a metric year has not been used
since 1806).
________________________________
________________________________
________________________________
________________________________
________________________________
Many Data Centre operators still have a battle to fully understand where all power
consumption is dissipated within their sites, and therefore struggle to fully implement
________________________________
effective energy efficiency measures.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There are many factors that impact the utilisation and distribution of power
infrastructure, such as:
►________________________________
Poor power factor
________________________________
► Cable transmission losses
► Switch and cutover systems
►________________________________
Inefficient equipment
________________________________
► Equipment conversion losses
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Data Centres are critical to the organisations that operate and use them. Therefore,
they are designed so services are available 24 x 7 x 365. Historically, this was achieved
________________________________
by building-in very high levels of resilience at the physical infrastructure level.
IT equipment has been, and often still is, running at low processor utilisation - so there
________________________________
can be more physical equipment (all using power) installed than is actually needed to
provide IT services.
As________________________________
a result, most Data Centres operate at very low utilisation levels, which is not
________________________________
efficient.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The Data Centre is a complex environment designed to house IT equipment. Utility
power entering the Data Centre has to pass through a number of stages of voltage
________________________________
transformation, distribution and cleaning before finally being delivered to the IT
equipment. Most power used within the facility is converted to heat, requiring
________________________________
significant cooling system capacity, which draws an additional load in a traditional,
recirculating air Data Centre. There are a number of ancillary support systems in the
________________________________
Data Centre such as lighting, generator pre‐heaters, fire suppression, as well as human-
occupied areas which also require power.
________________________________
The slide shows a simplified representation of the power delivery and loss path in a
________________________________
Data Centre.
Utility power enters the building on the left and passes through the delivery chain to
________________________________
the IT equipment on the right. Each stage in the delivery chain has inherent losses,
shown by red arrows, as well as the specific overheads shown as their own paths.
________________________________
The actual implementation of a Data Centre is considerably more complex than this
diagram, and details such as whether the computer room air conditioner (CRAC) units
________________________________
are fed from the UPS may vary.
________________________________
________________________________
________________________________
________________________________
Here the whole chain is assembled, and we can see that whilst our DCiE of 40%
looked bad, it is a small loss compared to the overall system with 98.6% inefficiency.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Data Centres have developed to meet demands of performance and resilience –
energy efficiency was just not on the radar. The resulting typical Data Centre consists
________________________________
of layer upon layer of energy inefficiency, by design!
Inefficiencies arise in a number of areas in the Data Centre, each with
________________________________
interdependencies on the others.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
For many Data Centres, the issue of failing to meet an appropriate level of the original
design is the fact that during the design phase, numerous assumptions had to be made,
________________________________
resulting in additional infrastructure. However, the design assumptions failed to come
to fruition and have resulted in dramatically underutilised components; this has
________________________________
resulted in significant knock-on effects:
► Components continually failing to meet desired load factors
________________________________
► Equipment being retired and replaced without performing to capabilities
►________________________________
Unnecessary procurement and maintenance costs
► Additional and unnecessary resource support
________________________________
► Dramatic levels of inefficiency
________________________________
This understanding of the actual present and future business requirements must be
known before undertaking the audit process. Then and only then can a full and
________________________________
accurate audit review be undertaken that provides a concise and honest evaluation of
the design capability to meet business needs.
________________________________
________________________________
________________________________
________________________________
________________________________
So far, the resilience in the Data Centre has been provided at the physical level –
duplicate this, double up on that, etc.
________________________________
Physical resilience comes at a price, actually at several prices. There is the first cost
capital expenditure (CAPEX) of provision, but there is also the operating expenditure
________________________________
(OPEX) of feeding the resultant inefficiencies.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Data Centres are critical to organisations that operate and use them; therefore, they
are designed so services are available 24 x 7 x 365. Historically, this was achieved by
________________________________
building a very high degree of resilience at the physical infrastructure level, primarily -
power, cooling and networking.
As________________________________
a result, some Data Centres have a continuous cooling capability to support HVAC
systems from an independent UPS; as they say - it is critical to keep it on, otherwise
________________________________
the temperature would rise beyond limits before cooling could be restored.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The environmental conditions of the Data Centre have direct impacts on IT hardware
operating potential and subsequent power demands. It is essential these operating
________________________________
parameters are reviewed, and the rationale for their given settings fully understood.
This will remove any second-guessing and allow a clear assessment of their operational
________________________________
requirements.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Inefficiencies appear at all levels, with each layer having an impact on the others.
Inefficient software requires more powerful IT equipment to run it.
________________________________
This in turn places demands on power and generates more heat, which needs more
________________________________
cooling, which places additional burden on the power provision, etc.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The machines required to run the latest generation of software are much more fast
and powerful than those used to run previous versions of the same software from just
________________________________
3-5 years earlier.
This means software implementations (the algorithms and data structures that make
up________________________________
software) keep getting more inefficient every year. While hardware gets
progressively more efficient, software engineers have not been as diligent in their
________________________________
algorithm designs; hardware companies are unlikely to object too much, as this creates
the vicious cycle of equipment upgrades.
________________________________
Equally, the connectivity requirements have become more and more complex, with
________________________________
load-sharing and meshed topologies becoming more common.
In most shared infrastructure and virtualisation designs, there is little understanding of
________________________________
which physical devices support which service(s), and large parts of the infrastructure
are ‘mirrored’; this includes replication of data between physical sites on premium
________________________________
storage environments (which also draw a significant amount of power). This layers
inefficiency on top of the mechanical and electrical (M&E) inefficiency, compounding
________________________________
the losses.
________________________________
Most IT devices draw most of their power at zero load, so the splitting of load across
two devices substantially increases power consumption.
________________________________
________________________________
________________________________
Which other systems exist in a Data Centre?
All introduce losses.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Risk cost is a major problem. Your Data Centre is likely to become a constraint, or
require re-investment before it is even full.
________________________________
Low energy costs helped to hide the problems
________________________________
This is no longer the case!
Energy costs are now a significant part of total cost of ownership (TCO)
________________________________
Studies and preliminary evidence suggest there are significant improvements to be
________________________________
made to existing Data Centres without major CAPEX
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Things have changed significantly and the old “We’ve always done it this way” approach
no longer stacks up.
________________________________
Not all requirements rate the same level of service, yet the best in class Tier 3 Data
Centre provides just that. This is normally not questioned, and just taken for granted in
a ________________________________
Data Centre. It would not happen in other parts of an organisation. For example,
would your Chief Financial Officer (CFO) sign off on all company cars being chauffeur-
________________________________
driven Bentleys because the CEO needs one?
________________________________
Another problem is the specification of best in class - What does it mean?
Does it mean:
►________________________________
All racks are shiny and black, the servers are the latest model, the network is
________________________________
running at peak speed and all cabling is concealed under the floor, so the CFO can
show people round?
►________________________________
It is meeting the requirements of the business in the most effective, efficient, flexible
and maintainable way?
________________________________
Best in class should mean best for the business, not a way to spend the most money to
________________________________
deliver the least usable capacity.
________________________________
________________________________
________________________________
Many Data Centre operators are unaware of the financial, environmental and
infrastructure benefits to be gained from improving the energy efficiency of their
________________________________
facilities. Even awareness does not necessarily lead to good decision-making, simply
because there is no framework in place for the operators to aspire to.
________________________________
Making Data Centres more energy-efficient is a multi-dimensional challenge that
requires a concerted effort to optimise power distribution, cooling infrastructure, IT
________________________________
equipment and IT output.
________________________________
Many activities have been initiated within the industry, with numerous vendor-specific
products and services on offer. However, there is a risk of confusion, mixed messages
________________________________
and uncoordinated activities.
As a result, investment can be made and expected savings not realised. In fact, there are
________________________________
many examples where the attempt to improve efficiency actually made it worse!
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Data Centres need a framework that owners and operators can use to set realistic
targets. Identify areas where these improvements can be made, giving guidance on the
________________________________
timescale and order in which these improvements should be carried out, so benefits
are actually realised and costly mistakes are avoided.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Develop a better understanding of what is really happening in the Data Centre.
A single-figure efficiency metric is meaningless for something so complex as a Data
________________________________
Centre, but the actual power usage is real and meaningful.
________________________________
As technology moves on, with more knowledge gained and better models developed, it
will change our view of best practices and how to apply them. Modular thinking
________________________________
enables the Data Centre to respond.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The drive for an all-encompassing Data Centre energy metric has become the Holy Grail
________________________________
for IT professionals. The focus on energy consumption in Data Centres requires facility and
IT managers to begin using some key metrics for monitoring and managing power. To
________________________________
ensure meaningful measurements are properly captured, regardless of the metric(s) used, it
is important to collect data often. Frequent measurements are necessary to capture
________________________________
changes in power utilisation at different times of hour, day or year; most importantly, during
periods of peak and low demand.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Firstly, we need to understand where the metrics we choose actually apply.
________________________________
Next, we need to ask ourselves what we want from metrics.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The points where energy use can be measured in the Data Centre to assist in
________________________________
understanding efficiency, losses, and impacts of changes / improvements can be represented
in three groups:
►________________________________
Simple measurements: These are the most basic measurement, required to deliver a
PUE / DCiE metric of the ratio of utility power to IT power for the facility. These can be
________________________________
initially informative to the operator, and are easy to perform with portable equipment. It
is important the external and data floor set temperatures are recorded, along with the
________________________________
electrical measurements.
________________________________
► Detailed measurements: The next level of assessment, where we specifically
measure each of the points where power is lost in the delivery chain, or diverted to
________________________________
non-IT loads. This provides more effective information on how to improve the facility,
and can directly change relevant behaviour. This level may depend upon the electrical
________________________________
configuration; for example, if the CRAC units are fed from UPS, then CRAC power is a
required measurement.
►________________________________
IT measurements: There is some limited utility in measuring IT device power draw in
situations where a physical device is dedicated to a specific task. This utility disappears in
________________________________
a virtualised or grid infrastructure, or where resources such as network or storage are
shared.
________________________________
________________________________
________________________________
Work has been carried out for several years to suggest and support energy metrics.
________________________________
However, these are not universally accepted; even where they are, they are not consistently
applied. To address these issues, the Global Taskforce was established. In February 2010,
________________________________
some goals and principles for collaboration were published, and on 28th February 2011,
some initial agreements were reached regarding the measurement of PUE.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
DCiE tends to be used in Europe, PUE in the USA. There has been much debate in the Data Centre
community about which one is ‘best’. The debate seems to have swung towards PUE, as the
________________________________
consensus is the Data Centre sector understands it better. But wait, not so fast! Power is an
instantaneous value (kW) and kWh is a measure of energy. So now there are ongoing discussions
________________________________
about whether to call it EUE. Really guys, life is too short.
These traditional metrics are a starting point, but they do not give any insight into what is happening
________________________________
within the Data Centre - just at the input to the facility and at the IT equipment. Even so, many Data
Centres do not even know what their DCiE is.
________________________________
Advantages of this approach are:
► It is relatively easy to implement with comparable measurements (kWh)
________________________________
► It is important to get a baseline, however rough it may be
► Periodic readings account for variations
________________________________
► Valid ‘IT productivity’ metrics are some time away, if they are even conceptually possible (or
possible to measure once defined). Nobody has even suggested valid units for such a metric
________________________________
yet
________________________________
►Problems with this approach include:
► It varies throughout the day, week and year, depending on IT activity and external influences
________________________________
► There is latency between the IT load and the total facility load; accurate real-time
measurement is difficult / meaningless, even with expensive, intelligent power monitoring
________________________________
systems
► There is no standard definition for IT load and facility load, so figures can be massaged by
________________________________
including or excluding elements
► It can be misused to compare one Data Centre with another, completely missing the point
► It may not properly reflect where real improvements have been made, and can even go the
wrong way
________________________________
DCiE / PUE works because it is simple and easy to measure.
________________________________
We are comparing energy with energy, and the units are the same.
For every 2kWh of energy supply, we get 1kWh of IT load. Our PUE is 2.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
First off, we should establish what key things we want from metrics. As shown, we want
________________________________
our metrics to be:
Relevant - they should report something we actually want to know: “Your PUE is 2, that
________________________________
means for every Watt of power you pay for, only half a Watt gets to the IT equipment.”
They need to be applicable to your situation.
________________________________
They should be transferable - if I measure the metric for my Data Centre, then I should
________________________________
be able to measure it in the same way for yours.
Having done this, we should be able to compare how our Data Centres scored on the
________________________________
metric.
________________________________
Finally, if I measure my Data Centre again a year from now, the metric should tell me
whether I have gotten better or worse.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
We have PUE / DCiE, we don’t need any more versions Let’s stop talking and start
________________________________
measuring.
We don’t need metrics that create perverse incentives. We know your PUE gets worse as
IT________________________________
electrical load falls, so targeting on PUE would work against people doing things to
reduce their IT energy consumption – such as virtualisation or power management – as
________________________________
that would harm their PUE.
________________________________
We don’t need metrics that can be easily massaged to give a better answer by subjective
application.
We________________________________
definitely don’t need metrics created to get some public relations exposure: Either you
have something useful to contribute or not.
________________________________
Metrics are not something to support your brand.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
One of the basic problems we face with metrics is the simple vs. useful dilemma.
________________________________
Successful metrics are simple - they provide people with a rapid and clear indication of
‘better’, or how things rank.
________________________________
Unfortunately, as we simplify a metric down to a single number, we destroy the information
that makes it useful. Your PUE is 5! How? Oh, you only turned the Data Centre on last
________________________________
week and it is running at 5% capacity; that is actually pretty good.
________________________________
Why should IT and Data Centre metrics always be a trade-off between simple and useful?
We need to find a way to have both.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
So we want to apply this approach to Data Centre metrics. Most of the current metrics fall
________________________________
on the curve that trades simple for useful. If we include some intelligence in the tool, we
can have both simple and useful.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Insurance price comparison sites understand that it would be meaningless to simply
________________________________
try and rank car insurers from ‘best’ to ‘worst’ as the best insurer for me is unlikely to
be the best for you.
________________________________
These sites collect some simple information from their customers, e.g. age, insurance
________________________________
group of car, etc. and then pass all the data on the insurers to produce a ranking for
that specific customer customised down to how large an excess you are prepared to
________________________________
have.
This allows them to have their cake and eat it, they are simple, insurers are ranked in
________________________________
order best to worst but this is still relevant and useful to the customer because the
ranking is customised to them.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
To help us look at this range of metrics, we will have a dashboard to help us assess each
________________________________
metric. This dashboard is used by Romonet and BCS, and gives a method of positioning
metrics within the Data Centre.
________________________________
On the left, we have the coverage of the metric in terms of the Data Centre stack: From
the energy supply through the infrastructure, the IT equipment is separated into compute,
________________________________
storage and network, then software and finally, the delivered business value.
________________________________
On the right, we have a graph to plot the metric: The horizontal axis is our judgement of
how useful the metric is, and on the vertical axis, how simple it is. We want simple and
________________________________
useful in the top right corner.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Well, it is definitely relevant it tells us how much energy is lost before it can get to the IT
________________________________
equipment, this directly impacts cost.
It is applicable, all Data Centre's have a PUE of some sort, it is also transferable, in most
________________________________
cases the PUE can be measured in the same way.
But we do know that the PUE is not really valid for comparison of Data Centre's without
________________________________
some extra information. If I am choosing between two Data Centre's the near empty Data
________________________________
Centre with a poor PUE could well improve and become more efficient than the already
nearly full Data Centre.
________________________________
Also, as the PUE is not very useful over time as we know that it varies with both
temperature and IT electrical load a change in PUE is not necessarily down to our M&E.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
So the DCiE is well-defined and covers a specific part of our infrastructure – the M&E
________________________________
plant.
We know what it measures, what the weaknesses are, and most people have heard of it.
It________________________________
is fairly simple to measure and report, and quite useful.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
CUE goes a stage further than PUE. It targets the heart of the matter, aiming directly at
measuring the carbon intensity of IT deployments, rather than simply measuring energy
________________________________
efficiency (which can be an imperfect proxy for the amount of polluting emissions that are
produced by a Data Centre).
________________________________
The total CO2 emissions are measured in kilograms of carbon dioxide (kgCO2eq) per
kilowatt-hour (kWh), and total Data Centre energy is the amount of power used, as
________________________________
measured at the utility meter. If the Data Centre is running entirely on electricity sourced
from the power grid, CO2 emissions can be determined by looking at Government data
________________________________
for the region. If part (or all) of the electricity used by the Data Centre is generated locally,
the total CO2 emissions can be found by adding the local source to the fraction that
________________________________
comes from the grid, after multiplying each by its carbon intensity (kg of CO2eq per kWh).
________________________________
To get IT equipment energy, you add the load associated with all IT equipment (such as
storage and network equipment, along with supplemental equipment such as KVM
________________________________
switches, monitors, and workstations / laptops used to monitor or otherwise control the
Data Centre). Everything else in a Data Centre, including cooling, lighting, etc., is excluded
________________________________
from that number.
Once you’ve done this simple math, you will get a CUE number that allows you to
________________________________
compare the CUE of different Data Centres, and track progress in lowering greenhouse
gas emissions.
________________________________
________________________________
________________________________
The units of WUE are litres / kilowatt hours (L/kWh)
________________________________
WUE, or Water Usage Effectiveness, is a metric that assesses the water used onsite in a
________________________________
Data Centre, including humidification, water evaporated onsite for energy production, or
cooling of the Data Centre and its support systems. WUE is expected to be measured and
________________________________
reported
as:
to optimise the operational water use. This can be accomplished by such actions
►________________________________
Reduce IT energy use, thereby lowering cooling demand, thus modulating water
consumption
________________________________
► Ensure the humidity control system is optimised, and the Data Centre is running at the
________________________________
low-end of ASHRAE guidelines (5.5°C dew point)
► Optimise cooling tower operations (if they are used) to increase cycles of
________________________________
concentration
________________________________
► Implement all appropriate best-practice airflow management strategies to improve
cooling efficiency
►________________________________
Operate the Data Centre at or near the ASHRAE-recommended upper limit for
temperature, as this will (depending on the cooling plant) allow warmer chilled water,
________________________________
and require less evaporation of water to produce it
________________________________
________________________________
Energy Reuse Effectiveness (ERE), is a metric for measuring the benefit of reusing energy
________________________________
from a Data Centre, and provides operators with greater visibility into opportunities in
identifying and recovering energy from their facilities. Rather than just showing what
________________________________
energy can be reused within the Data Centre, ERE scores on what energy can be reused
after it leaves. For example, whether energy created by the heat from servers can be used
________________________________
to power other offices, or even nearby homes.
ERE has a reporting range from 0 to infinity. ERE does allow values less than 1.0, as an ERE
of________________________________
0 means that 100% of the energy brought into the Data Centre is reused elsewhere
(outside of the Data Centre control volume).
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
You may argue about the fine details of the scope, but energy metrics are all focused
________________________________
towards the bottom of the Data Centre stack.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Most issues with energy metrics like PUE stem from the fact it varies with IT load, but
________________________________
cannot include IT in the measurement. Most of the effort in developing new metrics is now
centered on trying to define the overall ‘productivity’ of the Data Centre, calculating the
________________________________
‘useful work’ output against the energy input.
________________________________
Useful work metrics try to add extra dimensions to measurement in the Data Centre by
________________________________
assessing some outputs which are higher up the Data Centre stack, rather than further
down the stack. It is a very difficult process to define what should be measured, let alone
________________________________
actually measuring it.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Before we go and examine the metrics in detail, we should dispose of one category
________________________________
which was popular when the industry first started to look for efficiency metrics - the
chained value metric.
________________________________
Corporate Average Datacenter Efficiency (CADE) is a good example where we
multiply a series of component measures to get an ‘overall efficiency’.
________________________________
The problem is these metrics simply don’t work - they are conceptually flawed.
________________________________
The values are not independent. If we improve IT efficiency, the IT utilisation will fall,
unless we reduce the amount of IT equipment; then Data Centre utilisation falls,
________________________________
meaning the DCiE will get worse.
________________________________
Once we understand Data Centre efficiency, it is trivial to find instances where the
reported CADE goes in the opposite direction to the actual impact of a change, and
________________________________
therefore CADE is simply not a Data Centre efficiency metric.
Part of this problem is due to the mixing of capital asset utilisation metrics with
________________________________
operational efficiency metrics. A colocation provider would want a high Data Centre
utilisation, as it would get paid more, whilst a corporate site would strive to keep it
________________________________
down to delay the point at which more capacity and CAPEX is required; CADE
implicitly assumes higher utilisation is better.
________________________________
________________________________
________________________________
So we have looked at the four basic types of proxy being used for the next step, DCEP, and
________________________________
seen how some of them may be quite useful while others are quite useless. Perhaps there
is one you’ll want to tailor and use.
A________________________________
word of warning though - The industry seems to want to agree on one magic metric to
measure Data Centre efficiency, but regulators want one too. There is already one national
________________________________
government trying to develop a Data Centre productivity metric to mandate minimum
efficiency.
________________________________
We all know how targeting on DCiE can create perverse incentives by punishing efficiency
________________________________
improvements in IT.
The current DCEP proposals present a number of issues, if used by regulators:
►________________________________
There would be a substantial reporting burden with most of these proxies
►________________________________
The number of perverse incentives tends to rise with the complexity of the metric
used. DCEP would create a lot of perverse incentives; if you are measured on your
________________________________
CUPS, then buy a lot of low-performance servers
► DCEP (if regulated) would also create many constraints that would work against the
________________________________
optimal design and operation of Data Centres
________________________________
________________________________
________________________________
Now, most of the issues with DCiE stem from the fact that it varies with IT load but
________________________________
cannot include the IT in the measurement.
________________________________
Most of the effort in developing new metrics is now centred on trying to define the
overall ‘productivity’ of the Data Centre and measuring the ‘useful work’ output against
________________________________
the energy input.
The Green Grid calls this Data Centre Energy Productivity (DCEP).
________________________________
The problem is that there is neither a description nor units for ‘useful work’ so the
________________________________
industry is now looking for proxies we can use to estimate the ’useful work’.
These fall into four basic categories which we do not have time to cover in detail in this
________________________________
course, other than to say that it is early days and there is much more work to do.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The Green Grid call this Data Centre Energy Productivity (DCEP). The problem is there is neither a
description, nor units for ‘useful’ work, so the industry is now looking for proxies to estimate the useful
work. These fall into four basic categories:
________________________________
________________________________
Qualitative values: These achieve a very broad coverage of the Data Centre, from input energy
through to business output. The output is a single number per Data Centre, which is simple, but they are
________________________________
very complex to implement. They are not very useful as common measures for the market, but
dependent upon your workloads and business activity, you may be able to define specific / relevant actions
and values, making the metric useful to you. The target is a view for the general market and the green
________________________________
area is where the metric might score for different users.
Measured quantities: The coverage of this type is relatively narrow - both include Data Centre
________________________________
infrastructure. The bits / kWh includes the network and some undefined aspect of compute and storage,
whilst the operating system instances include part of the software layer. These metrics are certainly
________________________________
simple to report and measure, so they score well for simplicity, but their usefulness is rather
questionable.
________________________________
Utilisation benchmark: Currently has a small coverage including only Data Centre infrastructure and
compute equipment. It is quite complex to measure and the output units may not make much sense to
________________________________
the business, so it may be hard to explain. In the general case, this benchmark is not ready to be useful;
but if it maps well, then on to your business and facility. For example, if you are eBay and you have huge
homogeneous farms of compute, running a small number of well-characterised workloads, perhaps you
________________________________
have your own custom benchmarks. It could then be a useful and effective measure for you.
Average of averages: Coverage is (again) just the M&E and compute. CUPS is still relatively complex to
________________________________
measure, as it requires analysis of the server estate, grouping and then collation of processor utilisation
data for those groups. Furthermore, it is hard to explain what a MegaCUPS score actually means to the
________________________________
business. It is hard to argue CUPS is useful for the general market, when it is as easy to use more
applicable benchmark times utilisation methods.
________________________________
A series of submetrics have been designed to reflect the energy-saving efforts of the
________________________________
four main stakeholders involved in the Data Centre. Their goal is to improve energy
efficiency of IT equipment in a Data Centre, such as servers, storage and network
________________________________
devices. DDPE is designed as a metric for indicating productivity per unit energy. Similar
to the concept of DCEP, which uses a series of proxy metrics to establish useful work,
________________________________
formulating Data Centre productivity.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The weakness in DPPE is that it assumes all IT activity is useful work and does not really
________________________________
account for business value.
Another point to think about is what happens if a Data Centre were to have 100% green
________________________________
energy, DPPE would be infinite! To stop this happening, DPPE imposes a maximum value
on GEC of 0.8.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The major issue with current Data Centres is not how much power they consume when
________________________________
delivering peak business output, but how much of that power they consume to do nothing.
If a car used fuel at the same rate whether driving four passengers on the freeway or idling
________________________________
stationary, this would be seen as completely unacceptable; however, this is how most
current Data Centres behave.
________________________________
The Data Centre Fixed to Variable Energy Ratio metric (DC FVER) measures for the first
________________________________
time what proportion of Data Centre energy consumption is variable, i.e., related to the
useful work delivered, versus what proportion is fixed. This allows operators to understand
________________________________
how much of their energy cost is related to the work delivered and how much is a fixed
burden to be eliminated.
DC________________________________
FVER allows each Data Centre operator to select an appropriate set of proxy
measures for useful work, which are meaningful to their business activity. These are then
________________________________
normalised and combined with existing PUE measurement points to allow for a
comparable measure of Data Centre performance.
________________________________
________________________________
Data Centre
Real Data Centre
FVER
11:1
Comment
> 90% of energy consumption is fixed and
________________________________ not related to workload, < 10% is related to
work delivered
________________________________
Ideal Data Centre 0.1:1 < 10% of energy consumption is fixed,
________________________________ >90% related to work delivered
________________________________
The first sample Data Centre represents an old facility without any cooling economiser
________________________________
and a high fixed infrastructure overhead component (poor PUE at low IT load).
The second sample is a newer Data Centre which uses the same UPS, CRAC, chiller,
________________________________
PDUs, etc., but uses airflow containment, an indirect water-side economiser and VFD
control for CRAC fans and water pumps. This site has a substantially lower fixed
________________________________
infrastructure overhead component and a more significant change in PUE with external
temperature.
________________________________
The DC FVER metric leverages both existing measurement protocols developed for
________________________________
DPPE and PUE, as well as the effort expended defining and testing The Green Grid
productivity proxies (DCeP).
________________________________
This alternative approach transforms the issue and thereby provides a high-value
diagnostic metric for the Data Centre, which presents a very strong indication of
________________________________
improvement direction and fewer perverse incentives; it provides a limited ability for
operators to compare their ‘score’ (the major part of PUE’s success, for example).
DC________________________________
FVER is a normalised metric allowing operators to understand and compare how
________________________________
well their energy consumption is linked to the ‘useful work’ (or business value) delivered
from their Data Centre; specifically, how much of their energy consumption is fixed and
________________________________
how much is variable with load. Ideally all would be variable, but this is not realistic in
practice.
________________________________
________________________________
A particular issue seems to be operators accept they didn't want to measure the CPU
________________________________
utilisation of every process on every server in their Data Centre, because that is absurd
and unnecessary.
________________________________
The same people then fall straight down the same hole when looking for the proxies
which suit their business.
________________________________
The whole point of a proxy is it is a substitute for the other thing which is too hard to
________________________________
do; don't go and make the proxy as unachievable as the real measurement – it is an
indicator.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The selection and implementation of proxies focus around two problems – How do I get
a first measurement and how do I get a first win when I change something?
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
DPPE and DCeP proxies have merit, but they are complex to understand and implement.
________________________________
FVER is a good complementary fit with PUE; it is equally simple and shares the common
aim of driving out waste energy.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The extended parameters become useful when there is an objective way of measuring and
________________________________
reporting against them.
RCI and RTI, developed in the US by ANCIS, are a complementary pair of metrics that
________________________________
give an indication of the environmental health of the IT equipment. They are agnostic of
the limits (i.e., the limits are whatever the CDCDP determines them to be). The examples
________________________________
use ASHRAE Class 1 envelopes for indicative purposes. Benchmarking is in progress at
Lawrence Berkeley National Laboratory in the US, and many Data Centres so far are well
________________________________
outside the guidelines, demonstrating poor environments for IT kit despite being
overprovided with cooling capacity.
________________________________
Obtaining the information to determine RCI and RTI can require many sensor points, but
________________________________
the presentation comes by way of a simple dashboard. It is an example that meets the
desired characteristics of a metric; it is both simple to understand and extremely useful.
________________________________
The CDCDP should consider this level of environmental sensor deployment in the Data
Centre.
________________________________
________________________________
________________________________
________________________________
________________________________
As already highlighted, managing density, diversity and evolving network environments has
________________________________
become a major challenge within the Data Centre. This is a problem ASHRAE has been
committed to researching and providing solutions, not only for the US but the global
________________________________
market.
Rack Cooling Index (RCI) metric analyses conditions such that the thermal network
________________________________
environment is defined by equipment intake temperatures. Compliance with intake
specifications is the ultimate cooling performance metric. Cooling effectiveness is gauged
by________________________________
RCI when loads are introduced. By combining CFD modelling in parallel with this
metric, a tremendous amount of data can be processed and presented in an
________________________________
understandable, objective and standardised format.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Return Temperature Index (RTI) is a measure of the air management energy performance.
________________________________
The primary purpose of improving air management is to isolate hot and cold air streams.
This allows elevating both the supply and return temperatures, and maximises the
________________________________
difference between them, keeping inlet temperatures within the recommended limits. It
also allows reduction of the system airflow rate.
________________________________
This strategy allows the HVAC equipment to operate more efficiently. RTI is ideal at 100%,
wherein the return air temperature is the same as the temperature leaving the IT
________________________________
equipment; supply air temperature is the same as the rack inlet.
________________________________
Point of note: Temperature and humidity will affect the reliability and lifecycle of IT
equipment. Modifications or changes to the air management, temperature and humidity
________________________________
settings should be evaluated with metrics such as the Rack Cooling Index (RCI), which can
be used to assess thermal health of the IT equipment.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Before we get into detail, we will have a little diversion.
________________________________
Efficient and effective are often taken to mean the same thing.They are not!
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Everyone can relate to this simple analogy.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A Data Centre is much more complex, but using the same simple argument shows
________________________________
that a Data Centre with DCiE of 40% (PUE 2.5) is likely to be neither efficient nor
effective, as it could run out of capacity before it reaches its design limits.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The purpose of the Data Centre is, first and foremost, to meet the business need. It is
________________________________
then a business decision to determine what (if any) compromises are to be made in
the name of efficiency.
________________________________
The Data Centre must be able to meet ongoing requirements post-commissioning,
and must be designed for operability.
________________________________
Fortunately, a good design can be effective, operable and efficient; one does not always
________________________________
have to be at the expense of the others.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Once the design is handed over, it is imperative for the operator to keep it working
________________________________
within design parameters as it moves through its lifecycle. In order to do this, effective
monitoring must be employed. Retrofitting can be an order of magnitude more
________________________________
expensive, or not even possible; the CDCDP should consider incorporating monitoring
requirements within the initial design.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Fundamentally, the CDCDP has a limited number of principles to address.
________________________________
We should make sure the Data Centre is right-sized for the requirement throughout
its lifetime. This requirement is likely to change, and the Data Centre must be capable
________________________________
of changing with it.
We should select the most efficient systems and equipment to deliver the
________________________________
requirement, operating these at their optimal range.
________________________________
The problem in a Data Centre is with so many different elements contributing to each
area, it is difficult to know where to start.
________________________________
Fortunately, we have some guidance in programmes such as the European Code of
Conduct and the DoE DCEP.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A basic axiom of the CDCDP programme recognises a situation in most Data Centres,
________________________________
where there are several different groups involved, often with different reporting lines in
the organisation. Each group is certain its role is the key one - as a result, they have
________________________________
different targets and objectives which are sometimes in conflict. Equally, they have ideas
how to make ‘improvements’ of one sort or another, but they have no idea how this
________________________________
might affect another group.
This barrier must be overcome from the outset.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
What is needed: A formal forum allowing each of the different groups involved to discuss and
understand the reasons, dependencies and impacts of any proposed requirements and changes.
________________________________
The CDCDP can play a vital role in establishing this principle from the outset, building it into
the operating plan for the Data Centre.
________________________________
The traditional sacred cow of Tiering must be continuously challenged. Nobody would say the
Tier framework is wrong - in fact, it is very comprehensive. However, it is all about resilience
and availability, and it is very often applied inappropriately and expensively by organisations.
________________________________
Not all applications are equal, and they do not all require 99.99% availability. Many enterprise
Data Centres offer the same levels of availability to all applications, certainly at the physical level,
________________________________
when it is not needed. Quite often this was just assumed to be required, and the question
“What can the business live with?” was never asked. The Data Centre could be split into zones
________________________________
with different levels of resilience, depending on criticality to the business.
It is impossible to second-guess the future, yet many designs allowed for statements such as
________________________________
“100% expansion in three to five years”. Think about this statement and what it means: 100% of
what? This resulted in Data Centres that were physically twice as big, equipped with twice as
________________________________
much power and cooling capacity, etc. As we have seen, losses are significant at low loads. A
modular approach allows better flexibility to meet requirements at the correct time; there are a
________________________________
number of ways this can be implemented once the mindset is established in the organisation.
Even within the modular approach, there is potential to improve efficiency. A Data Centre on
________________________________
Day One will not be fully equipped, so support systems should be able to take account of this in
some way. The cooling system is an example; there is no point running pumps and fans at full
________________________________
speed when the IT load is light; all that does is move too much air and fluid around, which
actually reduces cooling efficiency.
________________________________
There are other ways to increase resilience in IT systems, e.g., at the software level. These
should be explored.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It is important that key stakeholders have a broad understanding of the Data Centre,
and are then committed to driving design improvements through the business. Key
________________________________
stakeholder commitment will provide:
► Advocacy from senior managers
►________________________________
Visibility of the issues across your organisation
►________________________________
Impetus for the organisation to implement energy management
► Resources, both human and financial
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
DoE is partnering with the industry to develop a Data Center Energy Practitioner
________________________________
(DCEP) programme to accelerate energy savings in the dynamic and energy-intensive
marketplace of Data Centres.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
FEMP – Federal Energy Management Program
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The programme is structured so all DCEPs can take a Generalist-level qualification,
________________________________
followed by a choice of specialist tracks as appropriate. The Generalist can be
undertaken as an individual component, or combined with a specialist component.
________________________________
Training Overview:
Day 1 – Generalist Training and Exam
________________________________
DoE DC Pro Profiler
IT Equipment
________________________________
Air Management
Cooling Systems
________________________________
Electrical Systems
Commissioning, Monitoring and Controls
DoE DC Pro Profiler Example
________________________________
Generalist Exam
Day 2 – HVAC Specialist Training
________________________________
Training Overview
Environmental Requirements
________________________________
Airflow and Temperature Management
DoE DC Pro Air Management Tool
________________________________
Air Handlers and Air Conditioners
Liquid Cooling
Day 3 – HVAC Specialist Training
________________________________
Chilled Water Plants
Cooling System Controls
________________________________
Modelling Data Centre HVAC Systems
Commissioning
HVAC Specialist Exam
________________________________
Developed by the European Union’s Joint Research Centre (JRC), the Code is an initial
________________________________
response to the accelerated power consumption of Data Centres across the EU. Data
Centre power consumption is expected to double by 2020 (2007 reported use 56
________________________________
TWh, and 2020 projection is 104 TWh); its generation and distribution systems are
already under strain. Power outages, through excessive demand requirements across
________________________________
the world, are not uncommon. Added to this are serious concerns for the environment
with increased carbon-based energy consumption.
At________________________________
this moment in time, the EU strategy is to provide clear and concise direction
aligned with common sense, without resorting to legislation. To enable the programme
to________________________________
be driven forward, the EU has developed a set of best practice guidelines (Guide).
The Guide draws from the experience and expertise of Data Centre professionals,
________________________________
operators, vendors, consultants, academics and industry bodies.
________________________________
The reality is that this is something that will not go away and needs to be addressed
now, not at a later date. The impact of not adopting the EU Code of Conduct will have
________________________________
significant future business implications, as winning and maintaining UK and EU
government contracts will be dependant upon businesses being committed to the
________________________________
Code of Conduct.
________________________________
________________________________
________________________________
There are many instances where the responsibility for a practice will need to be shared
________________________________
between supplier and customer. This might be something as simple as the correct
orientation of IT equipment in the hot / cold aisle of a Data Centre.
In________________________________
this case, both parties should implement the practice themselves and endorse it to
the other party.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The best practices apply to all areas of activity in the Data Centre. There is a section in
________________________________
the Best Practice Supplement covering each of these in detail.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The Data Centre Maturity Model shows the relative steps Data Centres can take to
________________________________
enhance their energy efficiency, while remaining technology-agnostic. Other models,
while valuable, often focus on limited aspects of the Data Centre and promote particular
________________________________
products and solutions. The Green Grid lays out the journey in numerous areas of the
Data Centre, for users as well as the industry as a whole. The model promotes
________________________________
collaboration, innovation, and change, pushing the industry forward without constrained
thinking, emphasising resulting efficiencies rather than prescriptive steps.
________________________________
The ultimate goal is for Data Centres around the world to be able to do more with less,
minimise their environmental impact, and take advantage of each others’ insights to
________________________________
reach efficiencies faster, using the common language of the Data Centre Maturity Model.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Over time, The Green Grid expects organisations to progress through the levels of the
________________________________
Maturity Model. An illustration of progress is illustrated above; the expectation is that
state-of-the-art Data Centres (green line) will progress from Level 2 today to Level 5
________________________________
by 2016. Typical Data Centres (amber line) will move to Level 3, and underperforming
Data Centres (red line) will lag at Level 2. The Green Grid encourages all organisations
________________________________
to move through the levels of the Maturity Model as soon as feasibly possible, taking
into account business and site constraints.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Fully understanding IT infrastructure requirements is a key challenge for Data Centres as
they strive to improve energy-efficiency and productivity. For many, this challenge is not
________________________________
fully understood - having a clearer understanding of IT facilitates the accurate alignment
of both power and cooling requirements.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Devices carrying the Energy Star logo generally use 20–30% less energy than required by
________________________________
US Federal Standards.
The Standard Performance Evaluation Corporation (SPEC) was formed by the computer
________________________________
industry in 1988, to establish industry Standards for measuring compute performance.
SPEC has since become the largest and most influential benchmark consortium.
________________________________
The European Union, Australia, Canada, Japan, New Zealand and Taiwan have also adopted
the EnergyStar program.
________________________________
Many European products are labelled using a different Standard instead of Energy Star -
TCO Certification, a combined energy usage and ergonomics rating from the Swedish
________________________________
Confederation of Professional Employees.
Harmonisation will continue to progress, but designers and IT Managers must continue
________________________________
to challenge the vendors for more accurate power consumption information.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The practice of using nameplate values to estimate power requirements has, in many
cases, produced highly-unreliable results, including undersizing (miscalculating by derating
________________________________
too far), or oversizing (using the label as-is). An example server rated at 475W actually
used 142W at startup, but 133W under normal load – only 30% and 25% of its nominal
________________________________
nameplate rating respectively. Therefore, calculating power requirements based on
________________________________
nameplate ratings alone can significantly bloat power requirements.
These miscalculations do not increase service reliability, but rather increase the cost of
________________________________
operating a Data Centre. The more accurate the calculation of realistic power
consumption, the more efficient and reliable the electrical distribution systems will be.
________________________________
Accurate calculations will assist in improving service availability, and achieve best “value
for money” in operating costs.
________________________________
Remember that missizing the energy demands of IT equipment will have a cascade effect
upon the entire physical infrastructure of the Data Centre. More power equals more
________________________________
cooling, and more capital expenditure.
________________________________
“Time to move forward and become more accurate”
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Extending the envelope of the IT equipment is a major energy saver, providing the
________________________________
equipment will stand it.
Why not put pressure on the equipment manufacturers by telling them the operating
________________________________
environment their equipment must work in, rather than have them tell you what
environment must be provided?
________________________________
Make extended operating parameters a key decision factor in selection and
________________________________
procurement.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Many servers already have power-saving features built into their BIOS, yet these are
________________________________
disabled by default - Use them!
Many servers can already report operating parameters, such as temperature and energy
________________________________
use - Use them!
We can influence the IT equipment manufacturers by including energy-efficiency as a
________________________________
key part of the procurement process.This is already happening in Europe.
________________________________
Standard metrics such as Energy Star and SPEC power are being developed, so we can
use these where available. Alternately, we could use specific measurements against a
________________________________
load / performance cycle, similar to what the equipment will perform in service.
Specify that manufacturers state the maximum / minimum temperature and humidity
________________________________
for the IT equipment, so the use of free-cooling can be maximised. Don’t accept that
ASHRAE is the only answer.
________________________________
Specify and use equipment suitable for use at the power density in the Data Centre;
________________________________
ensure it does not require the cooling plant to operate outside of parameters to
maintain temperature.
________________________________
________________________________
________________________________
________________________________
Processors use power even when idle.
________________________________
Virtualisation is a means to ensure more activity is carried out by running the
processors harder at all times, increasing the power required by the processor, but
________________________________
reducing the number of processors required. In the example shown, it would appear
the processor could carry out the work of around eight similar devices for around 20-
________________________________
25% additional power consumed, a performance / power advantage of around six to
seven.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
In many shared infrastructure and virtualisation designs, there is little understanding of
________________________________
which physical devices support which services, and large parts of the infrastructure are
‘mirrored’.
________________________________
As you have seen, most IT devices draw much of their power at zero load, so the
splitting of load across two devices substantially increases power consumption.
________________________________
Virtualisation has, perversely, made this worse in many cases by making the operators
________________________________
more concerned about the loss of a physical server than ever before. Alongside this,
there is often a perceived requirement to add even more storage to the SAN for virtual
________________________________
servers, which do not have their own physical disk capacity.
Data Centres should thoroughly investigate the hardware resilience levels needed for
________________________________
each application, and provision accordingly. Resilience can be built into the software,
which is another story.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It is easy to think that everything in a Data Centre is similar, but this can catch out the
unwary.
________________________________
The shape, size, weight, power and environmental requirements may be very different.
________________________________
There are potential energy savings in each area.
Traditionally, these different systems were housed in a common computer room in
________________________________
order to minimise cable runs, particularly where copper cable channels are limited to
________________________________
100m. As a result, the environment in the room was dictated by the weakest element
(e.g., tape storage), leading to an over-engineered, monolithic facility.
________________________________
Consideration should be given to a modular approach where diverse equipment with
varying requirements is treated accordingly. This may lead to splitting servers from the
________________________________
SAN, or even splitting high-powered servers from lower-powered servers. Each Data
Centre will have unique, detailed requirements, but the overall approach should be
________________________________
similar.
________________________________
Think Modular, Not Monolith!
________________________________
________________________________
________________________________
________________________________
Lack of domain knowledge – The greatest contributor to poor software quality is the
________________________________
unfortunate fact that most developers are not experts in the business domain served
by their applications
________________________________
Lack of technology knowledge - Most developers are proficient in several computer
languages and technologies. However, modern multi-tier business applications are a
________________________________
complex tangle of many computer languages and different software platforms.
Unrealistic schedules - When developers are forced to sacrifice sound software
________________________________
development practices to ridiculous schedules, the results are rarely good.
Badly engineered software - Two-thirds or more of most software development
________________________________
activity involves changing or enhancing existing code. Studies have shown that half of
the time spent modifying existing software is expended analysing code.
________________________________
Poor acquisition practices - Most large, multi-tier applications are built and maintained
by distributed teams, some or all of whom may be outsourced from other companies.
________________________________
Consequently, the acquiring organisation often has little visibility into (or control over
the quality of) the software they are receiving.
________________________________
________________________________
________________________________
________________________________
________________________________
It is difficult for the CDCDP to influence software, but there are several things to be
________________________________
aware of that may impact the design.
There are many claims from software manufacturers about the efficiency, or green-ness, of
________________________________
their respective offerings. Whilst this is mainly a marketing exercise right now, we should
watch this area with interest, as there is undoubtedly a lot of good and interesting
________________________________
development work going on to improve the efficiency of software.
an________________________________
Another area that is worthy of consideration is how quickly software can recover from
incident, e.g., spinning back up on another virtualised server if the original server fails
________________________________
(VMotion).This could reduce the need for some hardware resilience.
Used in conjunction with software management systems, it could make management of
________________________________
power density simpler by allowing the software to migrate easily, in order to dynamically
reduce and balance loads in the Data Centre.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Software is often inefficient, and is just rehashed legacy software, which requires
________________________________
additional equipment to work around its issues. The first popular spreadsheet program,
Lotus 1-2-3, ran on an Intel 8086 IBM PC with 640Kb of RAM, and it ran reasonably fast.
________________________________
The latest software runs on hardware with many orders of magnitude increase in CPU
speed and RAM, yet it runs slower.
________________________________
Software is feature-rich, but these features are not used by the majority of users. 1E, a
leading industry analyst firm estimates that 22% of all IT spend is on software, which is
________________________________
$726 billion annually worldwide. Knowing that a single piece of software can cost
upwards of several hundred dollars with licensing and support, these findings indicate a
________________________________
significant percentage of IT software spend is currently wasted.
________________________________
The primary reason new software runs slower is due to the fact it’s developed and tested
on the latest hardware available. This means the software engineers are able to get away
________________________________
with much less efficient implementations of algorithms and data structures, which could
be made much faster and much less complicated, if there was an incentive to do so.
________________________________
Software companies profit greatly from the sale of new hardware, which comes with
versions of their operating system that are able to be less-efficiently coded.
________________________________
One approach to this is right back at the procurement decision stage.
________________________________
Can we stop buying badly-written software that needs hardware clusters (or expensive
‘cluster’ virtualisation underpinning) in order to work?
________________________________
Can we choose software designed to be manageable and resilient?
________________________________
Data Centres provide secure storage, often on SAN. It is typical for the SAN to be
________________________________
duplicated, mirrored or striped in some way requiring the storage capacity to be
doubled or tripled. In some cases, the SAN storage capacity can be replicated ten or
________________________________
more times. Virtual servers without their own physical disk can also require the SAN to
provide them with storage.
________________________________
The SAN provides fast access to data - but do all applications require this speed, or can
some data be stored using lower-power (slower) storage systems or even offline?
________________________________
A robust data storage policy should be implemented from the outset, which controls
________________________________
the performance, resilience and availability of data for each business application,
according to the business need.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Virtualisation is the ability to replace a number of physical servers with virtual servers
________________________________
running on a single physical server.
Different operating systems can be run in parallel, as well as instances of applications,
________________________________
e.g., Oracle. There may be issues with licensing and support, though, which need to be
investigated in detail.
________________________________
There are many benefits that can flow from a virtualised IT strategy that are beyond the
of________________________________
scope of this course, but there is a clear impact on the CDCDP, who must take account
the power and cooling provision. The CDCDP may consider separating the high-
________________________________
density equipment into a separate zone.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Server virtualisation is a hot topic within the Data Centre environments, allowing
________________________________
businesses to run multiple applications, operating systems on fewer hardware resources
and greatly appealing to IT managers aiming to improve utilisation. Another major
________________________________
attraction of virtualisation is the potential opportunity to make considerable financial
savings from the numerous benefits it introduces. Some of the more significant benefits
________________________________
are:
► Server Consolidation – By virtualisation of server environments, figures are
________________________________
highlighting increased hardware utilisation of up to 60-80%. Providing consolidation
ratios of 3:1 or 4:1, but some companies have reported higher ratios.
________________________________
► Smaller Footprint – Reduction of the physical number of servers and real estate
being utilised by the business.
________________________________
► Flexibility and Agility – This is a factor that really does apply to virtualisation, as it
________________________________
introduces a quick creation of different operating system environments, allowing
legacy and new applications to operate alongside one another.
________________________________
► Rapid Application Deployment - Allows a rapid infrastructure provisioning that
is considerably quicker than traditional processes, also making back-up and recovery
________________________________
a lot more manageable.
► Increased productivity – Improved service capability
________________________________
► Business Resiliency – Provides easier replication and restoration.
________________________________
► Cost Savings – Fewer hardware components, initiating financial savings on upfront
hardware (and subsequent maintenance costs) when fully virtualised.
________________________________
Consolidation - The ability to get more from less: Less space, power and hardware
________________________________
With businesses continually endeavouring to improve service levels and availability, and reducing
costs by controlling / managing IT environments more efficiently, consolidation is an option
________________________________
enabling businesses to simplify their environments, increasing manageability while adding a
number of significant business benefits. The main categories for business benefits are:
________________________________
► Cost reductions (personnel, building and complexity)
► Maintaining business challenges
►________________________________
Improvement of service levels and availability
► Reducing the impact of external pressures
________________________________
Categorising the various types of Data Centre consolidation opportunities can lead to significant
cost reductions and better manageability, such as:
►________________________________
Servers and storage
►________________________________
► Applications and shared services
Networks
________________________________
► People, resources and processes
Before implementing any consolidation processes, the current status needs to be clear: What
________________________________
you have, where it is located and how it is configured and connected. Then you need to have a
clear logic around where you want to go with the consolidations. Next step is the ability to
________________________________
communicate all proposed plans to designated personnel for approval and implementation. A
successful Data Centre consolidation is having the systems to effectively and efficiently manage
________________________________
and report the status of the project from inception to completion.
________________________________
Many Data Centres were originally designed with major assumptions built into the
power and cooling infrastructures, to account for projected IT growth expectations. This
________________________________
has often resulted in poor utilisation of the installed infrastructure, leading to significant
inefficiencies. Today’s Data Centre designer should consider the feasibility of modular
________________________________
and zonal designs, to closer match cooling and power requirements with the actual IT
________________________________
load, whilst having the flexibility to meet future growth expectations.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Lifecycle Cost Analysis (LCCA) is a method of project cost evaluation - all charges to
own, operate, maintain and dispose of it are considered important to the decision to
________________________________
engage in or approve the project. It is the sum of present values of all costs over the
lifetime of a project, product or measure, including:
________________________________
Investment
________________________________
Installation
Maintenance
________________________________
Capital
________________________________
Energy
Opportunity
________________________________
Financing
________________________________
Operating
Disposal
________________________________
LCCA is particularly suited to assessing design alternatives that satisfy a required
________________________________
performance level, but may have different investment, operating, maintenance, or repair
costs, and possibly different lifespans. The result is generally expressed as the value of
________________________________
initial and future costs, as reflected by an appropriate discount rate. It can be applied to
any capital investment decision, and is particularly relevant when higher initial costs are
________________________________
traded for reduced future cost obligations.
________________________________
With the implementation of a well-defined and executed decommissioning strategy, there
are significant business benefits that can be gained:
________________________________
► Reduction of software licensing, hardware leases and Data Centre costs.
►________________________________
Mitigation of operational risks (limiting future pressure on resource and skill
shortages).
________________________________
► Maintain IT Staff resources to work on ‘in-life’ assets, not wasting time supporting
________________________________
‘life-end’ assets.
► Reclaiming power and cooling capacity
►________________________________
Improvement of the Data Centre efficiency.
________________________________
Continual decommission assessments = Efficiency and cost-effectiveness
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Electrical distribution is usually covered by regulations, so the CDCDP must check Local
________________________________
Codes.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The general principle in all National Regulations and Codes is that the power system
________________________________
shall be safe, resilient, efficient and measurable.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The slide illustrates points in the Data Centre power train where energy use can be
________________________________
measured, and provides an understanding of key areas where energy inefficiencies and
losses occur; this impacts electrical distribution capability and (more importantly)
________________________________
highlights areas of potential energy efficiency improvement.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The resilience of the power infrastructure is obviously a critical element in the Data
________________________________
Centre, but the CDCDP should ask searching questions about what levels of physical
redundancy are really required. There is seldom any reason why all applications require a
________________________________
fault-tolerant Data Centre - only the business-critical ones need that. So why provide it to
every application?
It________________________________
is like every member of staff getting a Bentley because that Chairman has one! The
CDCDP should consider different zones of resilience.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
With their large demand requirements, Data Centres will normally have at least one HV
________________________________
intake, provided by the power company.
The distribution (copper) losses within the Data Centre can be reduced if large blocks of
________________________________
power are distributed at HV to load centres in the computer room.
________________________________
With the substation sited at the centre of its load, the loss and voltage drop in the LV
distribution system will be minimised. The cost may also be significantly cheaper than an all-
________________________________
LV system, due to less copper mass required; however, there may be additional
requirements imposed by Local Codes.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Ideally, distribution transformers and main LV switchboards should be sited at their load
________________________________
centres rather than the building periphery, provided all local supply rules and fire
regulations can also be followed.
________________________________
Traditional location of a transformer room in a building is normally the ground level, with
appropriate access for loading and unloading substation equipment. The main LV
________________________________
switchroom is typically located adjacent to the transformer room, and all sub-main and
feeder circuits (including the rising mains) will be fed from the main LV switchboard.
________________________________
Distribution losses and cost for electrical loads far from the main LV switchboard (e.g.,
roof level) are usually high.
________________________________
Electrical rooms are usually incorporated in the design of Data Centres where all major
________________________________
power and mechanical plant rooms are located. Transformers and main LV switchrooms
could be provided around the data hall floors to minimise LV distribution losses.
________________________________
Problems to consider include separate cable duct provision for HV (11kV) cables, vertical
transportation for transformers and switchgear, fire protection, etc.
________________________________
Substations sited other than at ground floor locations may need to be installed with non-
________________________________
flammable equipment to satisfy fire safety requirements - e.g., SF6 or vacuum circuit
breaker, SF6 or silicone fluid-filled transformers, and LSF / XLPE cables.
________________________________
________________________________
________________________________
Transformer efficiency varies with load, depending on whether it is linear or non-linear.
________________________________
Modern IT load is fairly linear; motors, pumps and fans are non-linear.
The CDCDP should be aware how the transformer best matches the requirement of the
________________________________
Data Centre.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It is necessary to maintain the generator so it will start within seconds when required. This
________________________________
requires regular maintenance, and also means the engine needs to be kept in a state of
readiness.
________________________________
As with a diesel engine in a car or truck, if the engine gets cold, it can be difficult to start; if
the fuel gets cold, it can start to freeze. Heaters maintain the temperature of generator
________________________________
systems.There are various recommendations and common practices:
________________________________
► Although generators may be rated to work in low ambient temperatures, it is typical to
maintain the environment above 5ºC.
►________________________________
It is important to ensure the entire generator is kept above dew point temperature.
► Water jackets are used to keep the generator warm when it is not running; these are
________________________________
typically maintained at around 21ºC (70ºF).
________________________________
► Block heaters are used to heat the engine block, allowing combustion to take place at a
faster rate on initial start up. These circulate coolant around the cylinders and maintain
________________________________
the temperature between 40 and 55ºC (100-120ºF). These heaters can require several
kW of power.
________________________________
We should consider whether we can reduce the set-points for the engine and the
generator. That said, ensure the manufacturer is properly consulted, as any non-authorised
________________________________
changes could nullify the generator warranty.
________________________________
________________________________
Verify UPS efficiency at the load point where the Data Centre is going to operate, and
________________________________
optimise it at that load. UPS efficiency ratings are not static - rather, they follow a curve
dictated primarily by power capacity percent utilisation.
________________________________
UPS efficiency ultimately is calculated by potential, less fixed and proportional losses. While
proportional losses, in the forms of heat-dissipation (“I2R” losses), are tied directly to
________________________________
increases in load, fixed (or “no-load”) losses remain constant, independent of the amount
of current running through the UPS. This means fixed losses will have a more significant
________________________________
impact on efficiency when a UPS is evaluated at partial load.
________________________________
Source: Emerson Network Power
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Energy losses from the UPS generate heat, requiring additional electrical energy
consumed in the cooling system to remove it. The total energy loss from the UPS is a
________________________________
sum of both the energy losses of the UPS and cooling system. UPS energy losses (kW)
are calculated from UPS energy efficiency at the normal operating load conditions (i.e.,
________________________________
50%); cooling losses are calculated using cooling effectiveness. Cooling effectiveness is the
________________________________
ratio of the annual average kW of cooling energy consumption required to remove a kW
of heat from UPS – a 50% cooling effectiveness is a good estimate (cooling effectiveness
________________________________
ranges from 30%-70%, depending on cooling system efficiency and regional environmental
conditions).
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
► Switch to “standard operating mode” when conditions arise
►________________________________
Significant efficiency gains, especially at high loads
► 2% to 8% improvement
________________________________
► Increased UPS efficiency translates to lower power bills
________________________________
► 8% gain translates to yearly financial savings
► What’s New?
________________________________
► Multi-modal UPS with sophisticated control algorithms
________________________________
► Intelligent Paralleling
High-Efficiency Modes and Intelligent Paralleling capabilities provide cost-effective
________________________________
approaches to reducing current and future energy costs.
►________________________________
Some UPSs can still manage power quality while delivering significant power savings
► There will be some trade-offs between availability and efficiency, so consider all factors
________________________________
in your decision.
________________________________
________________________________
________________________________
________________________________
The benefit of eco-mode is the efficiency of the bypass path is typically between 98.0% and
99%, compared to the base UPS efficiency of 94% to 97%. This means there is a pickup in
________________________________
UPS efficiency of between 2-5% when eco-mode is used.
________________________________
Eco-mode represents a potential way to save energy in Data Centres and other UPS
applications. Data Centre operators can expect to see savings on the order of 2-3% in
________________________________
total energy if eco-mode is enabled. Higher percentage savings are possible if the Data
Centre is operated at very light electrical loads. The energy savings associated with eco-
________________________________
mode are getting smaller as newer-generation UPS systems improve in efficiency.
The use of eco-mode does entail risks - it introduces a number of new modes of
________________________________
operation to the Data Centre, and reduces power protection. Current IT equipment is
much more resilient to power variations than the IT equipment of prior generations,
________________________________
suggesting this equipment should operate reliably using eco-mode. However, complex Data
Centre systems comprised of a mix of IT equipment, transformers, transfer switches, and
________________________________
other possible non-IT loads are less predictable in their response to infrequent and
________________________________
unusual power events; their compatibility with eco-mode is less certain. These
considerations have greatly limited the use of eco-mode in real Data Centres in the past,
________________________________
and are likely to continue to do so.
As with the introduction of any UPS infrastructure, careful due-diligence needs to be
________________________________
undertaken to certify critical IT load is protected, ensuring business needs are maintained.
________________________________
________________________________
The DCiE function now varies significantly through the life of the facility with distinct saw-
________________________________
tooth steps.
Although DCiE is significantly improved across much of the IT load range, if this facility
________________________________
were to be measured purely on DCiE, the results might be confusing to the operator, and
difficult to explain to business management.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Any technology using UPS should deploy high-efficiency settings when available.
A________________________________
UPS should be normally operated in its most efficient mode, e.g., without dual-
conversion. Some dual-conversion UPS systems can operate in ‘economy’ or ‘bypass’ mode.
________________________________
Deploy power in a modular fashion: PDUs, cabling, etc. can be fully deployed, but glossy
elements such as UPS and batteries are only brought into service when required to meet
________________________________
the load. In smaller Data Centres this may be via plug-in units, whereas in larger Data
________________________________
Centres, it will be by adding entire UPS systems.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Manufacturers now recommend replacing batteries every 3-5 years, to ensure the
load will be properly supported during power outages, especially during high load
________________________________
conditions.
There are a number of conditions that affect battery life, including temperature, depth
________________________________
and frequency of battery discharges. The shelf-life of batteries decreases with
increased ambient temperature. This is particularly important if the UPS is to be
________________________________
stored without periodically recharging the batteries.
________________________________
For example, at a 40ºC (104ºF) ambient, battery life is reduced approximately 70%
from what would be expected at 25ºC (77ºF).
________________________________
Even when the UPS is in use and the batteries are fully charged, higher ambient
conditions will reduce "float" service life.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Codes require electrical systems to be split into circuits, so distribution of power can be
________________________________
carried out safely.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
In a circuit supplying a load, we are concerned with voltage and current.
________________________________
The voltage presented to the load must be within Code or Regulatory limits.
The circuit must be capable of continuously carrying the design current.
________________________________
The circuit shall be protected by some device to limit and disconnect current flow in
________________________________
the event of a fault; this is usually a fuse or circuit breaker.
________________________________
Ib = Circuit Design Current
In________________________________
= Rating of Protective Device
________________________________
Iz = Is Current Carrying Capacity of the Circuit
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The copper loss of every main circuit connecting a distribution transformer and the main
incoming circuit breaker of a LV switchboard should be minimised by means of either:
________________________________
Locating the transformer room and the main switchroom immediately adjacent to,
above, or below each other
________________________________
Restricting its copper loss to <0.5% of the total active power transmitted along circuit
________________________________
conductors at their rated current.
The cross-sectional area of neutral conductors should not be less than that of the
________________________________
corresponding phase conductors.
In any electrical circuit, some electrical energy is lost as heat which, if not kept within safe
________________________________
limits, may impair the performance and safety of the system. This energy (copper) loss,
________________________________
which also represents a financial loss over a period of time, is proportional to the effective
resistance of the conductor, the square of the current flowing through it and the
________________________________
operational duration of time.A low conductor resistance therefore means low energy loss.
The length of the main distribution circuit conductors connecting the distribution
________________________________
transformer and the main incoming circuit breaker (MICB) of the LV switchboard should
be as short as possible, by means of locating the substation and the main LV switchroom
________________________________
adjacent to each other. A maximum conductor length of 20m (66ft) is recommended.
________________________________
Due to the possibility of large Triplen harmonic currents in the neutral conductor for Data
Centre loads with a large proportion of non-linear equipment, it is not recommended to
________________________________
use neutral conductors with a cross-sectional area less than that of phase conductors in
the main circuit.
________________________________
PDUs are not clearly defined by Standards, but they all have common features and
________________________________
functions. Some have metered inputs and outputs, others do not.
PDUs with the ability to report energy usage can assist in understanding Data Centre
________________________________
power requirements dynamically. In the US, there is an additional point to consider, as
PDUs also have transformers within them.
________________________________
The CDCDP should consider PDU features in selecting the initial design.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
This method is based on the assumption that supply voltages and load currents are
________________________________
sinusoidal and balanced among the three phases in a 3-phase / 4-wire power distribution
system.
________________________________
Extra care must be taken if the 3-phase feeder circuit is connected to nonlinear loads, such
as UPS systems, Variable Voltage Variable Frequency (VVVF) systems and Variable Speed
________________________________
Drive (VSD) motor systems.
________________________________
The design current used for cable sizing must take harmonic currents into account, but this
is outside of the scope for this course.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Maximum THD of current:
________________________________
I < 40A / 20.0%
40A < I < 400A / 15.0%
________________________________
400A < I < 800A / 12.0%
________________________________
800A < I < 2000A / 8.0%
________________________________
I > 2000A / 5.0%
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Motors appear in numerous support systems, and can be a major loss source. Variable
speed drives (VSDs) can have many energy advantages, so the CDCDP should discuss
________________________________
options with the electrical designer.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Cables are sized according to codes and regulations to meet safety and performance
________________________________
criteria. Copper losses in switchgear and cables are relatively small in the overall picture
(<5% typically). However, codes of practice are being discussed for energy-efficient
________________________________
electrical networks, which may impact Data Centres at some point. These seem to be
setting cable losses somewhere around 2.5%, so they may result in designing for shorter
________________________________
cable runs or using larger conductors.
The CDCDP must be aware of developments and discuss these requirements with an
________________________________
electrical engineer.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Poor distribution along the power infrastructure can have a significant cascading effect,
so it is essential to have a holistic understanding of Data Centre energy consumption
________________________________
and subsequent areas of loss. This is easier said than done for some Data Centres with
legacy components still in service.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Electrical switchgear and panels are the lifeblood of a Data Centre. Without them,
equipment failure is immediate. Infrared scanning can identify problems prior to failure,
________________________________
allowing you to take proactive steps to prevent costly (and sometimes dangerous)
electrical outages.
________________________________
If you are dual-powered, or utilising redundant power sources: Anytime you switch
power from primary to a backup power source, infrared scan the switchgear and
________________________________
backup PDU / RPP electrical panel boards and branch circuits for excessive heat.
Testing while under load will tell you if there has been any degradation in system
________________________________
components that could lead to future equipment failure and a possible outage.
Open and scan the electrical panel, then look for hot spots - loose or weak
________________________________
connections between breakers and the main panel bus and other deficiencies;
overloaded circuits, load imbalances, harmonic problems and defective electrical
________________________________
components can all contribute to a thermal scan heat signature.
When a hot spot is identified, amp probe the circuit to measure the draw. Calculate
________________________________
loads to determine if the problem is related to excessive load on one or a number of
circuits. If the draw is less than 80% of circuit capacity, focus your attention on the
________________________________
breaker, or wiring connections to the electrical panel. Check for loose or weak
________________________________
connections between a breaker and the main panel bus. Systematically work through
the components (testing and replacing if necessary), and electrical connections
________________________________
(checking for loose connections and tightening, if possible) until the problem is isolated
and resolved. If the draw exceeds 80% of circuit capacity, backtrack to the draw source.
________________________________
Has additional equipment been added to the circuit?
________________________________
Electrical Safety Inspection ensures power systems and equipment are properly designed
and installed, as well as operated and maintained in a safe and reliable condition. Verify full
________________________________
compliance with regulations and Standards - ensure employees are maintaining electrical
systems and equipment in reliable and safe working condition; follow safe work practices
________________________________
and use electrical protective equipment for energised and de-energised work.
________________________________
Electrical Safety Inspection is meeting electrical safety goals for regulatory compliance.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Start by considering one thing:AIR IS AN INSULATOR.
________________________________
Any system that is based on moving air around inside large open spaces is not going to
break any efficiency records. However, it is where we are today, so we must maximise
________________________________
efficiencies by minimising unnecessary losses.
Direct fluid cooling of the chipsets may be on the way, but it will take some years to
________________________________
become common.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There are many options available for providing HVAC, and these can be combined in
________________________________
different ways as shown. Interestingly, the majority of Data Centres use subfloor air
delivery, DX CRAC and condenser units, and air-cooled heat rejection.
It________________________________
is worth considering the relative efficiencies of the component parts as shown
above. Placing heat recovery close to heat generation (rack or row), and centralised
________________________________
liquid cooling (with water-side optimisation and cooling towers) is an efficient
combination being widely used at the moment.
________________________________
Free air-cooling is not represented in the above.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It is worth remembering the cooling system is a series of loops, each transferring heat to
________________________________
the next.This helps visualise what is happening at each stage.
Also, remember not all loops need be present.
________________________________
What is the effect of raising the temperature to the IT equipment?
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Before implementing any major cooling efficiency measures, it is essential the existing
cooling infrastructure is working to its best potential. It is paramount the integrity of the
________________________________
cooling is correctly maintained. Classic examples, such as:
________________________________
► Failure to decommission and remove redundant servers
► Failure to maintain the hot and cold aisle principle (classic example - perforated tiles
________________________________
placed in the hot aisle)
► Failure to remove redundant underfloor cabling, reducing the effectiveness of the
________________________________
plenum
►________________________________
Blanking panels not being fitted in cabinets
► Cabinet floor entries not being sealed properly (brushes and grommets)
►________________________________
Open cable cutouts in the hot aisle
►________________________________
Cable routes blocking airflow to equipment within cabinets
► Failure to firestop and seal wall or floor penetrations
►________________________________
Poorly maintained raised floors, causing unacceptable air gaps
________________________________
Individually, they may have a limited impact; combined, they will (and do) have a significant
bearing upon cooling efficiency, with unnecessary waste added to additional cost.
________________________________
________________________________
________________________________
Rack Cooling documentation and analysis:
Identify airflow obstructions within, above, and below the rack environment.
________________________________
Identify racks where air-mixing is likely to occur, and where blanking panels should
________________________________
be installed.
Examine rack enclosures for airflow suitability.
________________________________
Determine the airflow pattern (i.e., front-to-back, side-to-side) of racked and non-
racked equipment.
________________________________
Detect and document areas where rack inlet temperatures exceed industry
________________________________
Standards and guidelines, using infrared thermography.
Analyse rack arrangements for air distribution effectiveness.
________________________________
Measure the airflow and temperature of raised floor air distribution systems and
________________________________
overhead grilles, if accessible.
Locate significant obstructions under the raised floor and in the drop ceiling that
________________________________
affect airflow. This will be collected onsite or by using customer-provided mechanical
drawings.
________________________________
Determine areas of significant air leakage in the Data Centre.
________________________________
________________________________
________________________________
Extending the envelope greatly reduces the burden on the Heating, Ventilation and Air
________________________________
Conditioning (HVAC) equipment but, in order to maintain the integrity of the IT
equipment, it is necessary to efficiently manage the airflow.
________________________________
There has been much debate about which is the optimum operating envelope. Whilst it
is undoubtedly true tightly controlling the environment has a cost, letting it range too
________________________________
freely also has issues.
________________________________
Some Telcos and proof of concept sites operate to NEBS / ETSI, but for many other
operators this is not desirable. Some IT equipment can operate reliably at higher
________________________________
temperatures but reduces its processing speed; some operators simply cannot accept
the reduced performance.
________________________________
The debate continues and no formal decisions are published, but it is looking probable
the ASHRAE Allowable Class 1 (not the Recommended) will become a de facto range
________________________________
for Data Centres to operate within.
________________________________
________________________________
________________________________
________________________________
________________________________
Cooling system components can be broken down into roughly two categories.
________________________________
The logical place for us to start looking into improvements is right at the beginning - the
air management collecting heat from the IT equipment.
In________________________________
a traditional hot aisle / cold aisle Data Centre, cold air is supplied either via the
subfloor plenum or by overhead ducting. The hot air return is usually not contained.
________________________________
CRACs and CRAHs are rated for unducted operation under ASHRAE Standard 127,
________________________________
standard air conditioning systems are not.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
This table shows the usual way of controlling CRACs, return temperature and percentage
________________________________
relative humidity is prone to error. The CRACs were measuring temperature and relative
humidity (rh). The thermal dewpoint figures for the CRACs are calculated based on these
________________________________
readings. Humidity sensors are prone to drift and the CDCDP should be aware of this.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The CRAC sensors were incorrectly set up, and therefore duelling was taking place.
________________________________
The readings from the relative humidity sensors in the CRACs vary with the local
temperature at the measurement point, however this example was an open room, and
________________________________
therefore you would expect the thermal design power (TDP) to be relatively constant
throughout.The calibrated reference probes support this argument.
________________________________
Centralising the control or setting one unit as the master (and triggering on dewpoint)
________________________________
would alleviate this issue. Ignoring humidity altogether would do the same.
Mark Hydeman at Taylor Engineering has produced some good technical papers
________________________________
demonstrating that careful initial set-up, accurate recalibration and ongoing trimming of
the Sequence of Operation (SOO) for cooling systems is essential for optimum
________________________________
operation.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Using these basic laws with the picture on the previous slide can give us good insight
________________________________
into affecting efficiency by adjusting flow rates and temperatures.
Of course, this is a simplistic view and the HVAC engineer goes into much more
________________________________
detail.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
CRACs with fixed-speed fans also have fixed losses, and show a characteristic where
________________________________
efficiency drops dramatically at low cooling loads. In a real sample of Data Centres, fan
energy in the systems was found to be up to 7-10% of the total for the Data Centre.
________________________________
Changing some of the fixed losses (fan losses) into variable losses can improve efficiency
across the range.
In________________________________
a Data Centre, it is recommended each zone or partition is operated at a single
pressure, therefore all fans should be set to the same speed.
________________________________
The pressure must be sufficient to supply adequate air to the most demanding cabinet.
________________________________
Providing too little means recirculation and hotspots; too much means bypass, and is
generally a waste of energy and money.
________________________________
Most Data Centres find the optimum pressure varies between 10 to 15Pa (0.04 to 0.06
inches of water), although many were designed for subfloor static pressures of 20 - 25Pa
________________________________
(0.08 – 0.1 inches of water).
________________________________
This static view is not complete though: The Data Centre is a dynamic environment, and
the control system should optimise supply air temperature airflow on a regular basis to
________________________________
match changes, using trim and respond algorithms. Metrics such as RTI and RCI could be
used to control this.
________________________________
________________________________
________________________________
The power used by fans moving air around the Data Centre is a significant contribution
________________________________
to the total, not to mention to the noise levels.
Although this metric is not heavily promoted in Data Centre papers, it is something
________________________________
well-known to the HVAC engineer; the CDCDP can use the values above in discussions
with the HVAC engineer around expected system efficiencies.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Optimising CRACs may not be simple - Water and evaporative-cooled units are more
________________________________
efficient than air-cooled in a new build situation.
Pre-coolers are available for some units, but not others.
________________________________
VSD units tend to be manufacturer-specific and risky to retrofit (even if they are
available), because of possible icing on the coils.
________________________________
Adding air-side economisers may be limited or not even possible, due to building
constraints preventing the necessary inlet and outlet ductwork.
________________________________
Parallel control of CRACs is possible, but sometimes tied to a manufacturer’s standard
control algorithms, rather than control routines customised for the Data Centre.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Optimising CRAHs is easier because the CRAH is a simpler unit. Some of the
________________________________
difficulties remain, such as space for pre-coolers and air-side economisers, but
customised central control and fitting VSDs to the fans is less of an issue.
It________________________________
is also possible to improve CRAH efficiency by using custom coils with a high delta-T
performance.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Optimising the wet-side systems is complex, as there are many interdependencies.
________________________________
We will identify a few key considerations for the CDCDP so they can discuss them
with a mechanical engineer.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It is important to note that air-cooled chillers are significantly less efficient than water-
________________________________
cooled plants (30%-60%, depending on climate)
The chiller’s control used to be sluggish and designed for a constant flow rate: If flow
________________________________
was reduced, the CHW temperature would often fall below safety limits and the chillers
would trip. It was even possible to freeze the evaporator, resulting in severe damage;
________________________________
thus, the old situation that chillers require constant flow at all times.
Today’s chillers have digital controls and can respond to a rapid rate of load change
________________________________
(either through adjustments in flow rate, or entering temperature) without losing
control of the CHW supply temperature.
________________________________
Modern chillers can be operated with variable flow within the manufacturer’s
recommended flow ranges, as long as flow rate changes are not too dramatic.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Consider propeller fan towers.
________________________________
Centrifugal fans use about twice the energy for the same performance.
They should only be used where high static pressures are required (such as an indoor
________________________________
tower with pressure drop from louvres), or in cases where low profile is essential and
there are no propeller fan options short enough.
________________________________
Low noise is not a valid reason for using centrifugal fans. Low-noise propeller blades
combined with oversizing the tower shell to increase efficiency, and reducing fan speed
________________________________
mean sound levels can be as low as standard centrifugal towers.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Install VSDs on all chilled water pumps for more precise control and energy savings.
________________________________
Controlling pump speed in a complex cooling system is a job best left to experts, as
the sequencing of the system is critical to realise efficiencies.
________________________________
There are many papers discussing this topic, but we will not go into depth. The
CDCDP can do further reading if necessary.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Air-cooled and water-cooled chilled water plants can be configured with water-side
________________________________
economisers.These can also be retrofitted quite easily in many cases.
They show a 1 to 4-year payback in temperate climates (such as California) as a
________________________________
retrofit, and even faster payback in a new build.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It is possible to use air-side economisers which introduce external cold air directly
________________________________
into the cooling distribution system, effectively cutting out the rest of the cooling
system.
In________________________________
the off mode, dampers are positioned so hot air from the computer room is
directed through the terminal unit (CRAC/CRAH).
In________________________________
the on mode, dampers are positioned so cold air is brought directly from outside
________________________________
through filters, and hot air from the computer room is vented outside.
In reality, the dampers can be in an intermediate position, so it is not an ‘all or nothing’
________________________________
scenario. A benefit often overlooked is that economisers can act as an emergency
backup in the loss of a cooling plant.
________________________________
ASHRAE Standard 62 recommends the use of MERV 13 filters on units with outside
air economisers. Lawrence Berkeley National Laboratory has shown in field studies
________________________________
that units with MERV 7 filters have been able to keep dust at an acceptable level.
________________________________
Bringing smoke into the Data Centre could cause the fire suppression system to
trigger. Ultra-low leakage dampers and some form of early detection outside of the
________________________________
intakes is recommended.
MERV = Minimum Efficiency Reporting Value ranges from 1 to 16 - The higher the
________________________________
number, the finer the filter.
________________________________
________________________________
The DCiE for a typical Data Centre cooled by traditional methods is shown. It varies
________________________________
with external temperature, but the biggest influence is IT electrical load.
The best case is at full load when it is really cold (around 58%), so we can hazard a
________________________________
guess that the declared design DCiE was slightly more than this.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The DCiE for a Data Centre cooled by free air is shown. It is fairly constant with
________________________________
external temperature.
It has a step change when the traditional system kicks in. In the example, this is at 21ºC
________________________________
(70ºF). At higher IT inlet temperatures, the step would move accordingly. With
evaporative cooling, the step would be around 30ºC (86ºF).
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Diagnostic monitoring can help pinpoint particular performance issues.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Example used to calculate VFD energy savings.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Deadband: On thermostats that automatically control both heating and cooling
systems, a deadband is a temperature range in which neither system turns on. The
________________________________
deadband prevents the thermostat from activating heat and cooling in rapid succession
(called “hunting”). This conserves energy by providing a range of temperatures requiring
________________________________
no energy consumption
________________________________
SAT: Supply Air Temperature
RAT: Return Air Temperature
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A bin analysis may be appropriate to give a more detailed indication of improvement
________________________________
opportunities, such as the performance of water-side economisers.
Bin analysis is a statistical evaluation to calculate energy savings from thermostat
________________________________
adjustments, using system analysis and weather data (8,760 hours utilising dry bulb, dew
point, wet and dry temperatures, or humidity).
________________________________
This approach can help to explain variances, e.g., exceptionally hot or wet periods
________________________________
compared to previous years.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Here is an example of a simple annual average calculation approach for estimating
________________________________
CRAH fan energy savings from implementing physical separation of hot and cold air in
a Data Centre.
________________________________
This allows us to raise the air‐side delta‐T, and therefore reduce the necessary supply
air flow-rate from the CRAH.
If________________________________
we know what change in air-side delta-T we are designing for, we can estimate
________________________________
annual fan energy savings using one of the fan affinity laws.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
If the CCF for a room is between 100% and 110%, there is little to no redundant
________________________________
cooling capacity. Airflow management (AFM) improvements will likely improve IT
equipment intake temperatures, and create an environment where cooling unit set-
________________________________
points can be raised, increasing cooling unit efficiency. But there is no opportunity to
turn off cooling units.
________________________________
If a CCF is 110% to 120%, the number of running cooling units is likely being managed
well. There are approximately 1 to 2 redundant cooling units for every 10 units running.
________________________________
In most cases, this is sufficient to maintain the room temperatures when a cooling unit
fails. It is not recommended that any cooling units be turned off, unless the room has
________________________________
24-hour-by-forever monitoring and staffing.
For rooms with a CCF of 120% to 150%, there is moderate opportunity to realise
________________________________
savings from turning off cooling units. This can often only be done once AFM has been
effectively implemented. This does not require full containment strategies, but does
________________________________
require thorough sealing of raised floor penetrations and open spaces in racks, as well
as best practice placement of perforated tiles and grills.
________________________________
A CCF of 150% to 300% is most common. These rooms have substantial opportunity to
________________________________
reduce operating cost, improve the IT environment, and increase the IT load that can
be effectively cooled. Rooms in this range often have significant stranded cooling
________________________________
capacity that can be freed up by improving AFM. Rooms with a CCF greater than 300%
have great potential for improvement. A CCF of 300% means the total rated cooling
________________________________
capacity of running units at 3 times 110% of the IT load.
Assumptions:
________________________________
Cooling unit fan motors are running at full speed, no variable frequency drives (VFD)
Cooling units are capable of delivering their full, rated capacity
Heat load other than IT load (building, lights, people, etc.) is 10% or less of IT load
________________________________
Simple mathematical models can be constructed in a spreadsheet programme relatively
________________________________
quickly. Creating more precise models in spreadsheets can also be done too, but
requires a much greater investment in time.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It should be noted there are a number of other commercial tool sets now emerging on
the market, providing further analytic options. All of which the CDCDP should evaluate
________________________________
to properly understand their merits and shortcomings.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
ASHRAE has published models and technical papers that present the latest equipment
________________________________
designs and simulation techniques.
The Building Simulation List has lots of information on the latest tools, bug lists and
________________________________
techniques. Equipment manufacturers can (and should) provide performance data on
request.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Project Management has proved to be an effective and efficient way to direct and
________________________________
manage continual business changes. These in-turn place demands on the Data Centre,
________________________________
and therefore need to be controlled and coordinated to ensure operational delivery.
Businesses accomplish this by implementing:
►________________________________
More efficient use of limited specialist resources
► Improved affordability and better value of finances
________________________________
► Identification and management of project activities throughout the lifecycle, from
________________________________
inception to completion
► Development of project staff, improving team skillsets and capabilities to the benefit
________________________________
of the business
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Projects differ from other forms of work in a multitude of ways - They are a transient
________________________________
and unique endeavour. Principal features that characterise projects generally include:
________________________________
Duration is usually predetermined (finite) with static start and end dates
Events occurring during a project invariably affect subsequent events, both inside
________________________________
and outside the organisation
________________________________
The project organisation is often temporary, and can sometimes change throughout
its lifecycle
________________________________
Projects are undertaken in an environment of risk and uncertainty
Projects are seldom carried out in isolation, and can often interact with other
________________________________
projects and organisational entities
________________________________
With Data Centre design projects, complexity dramatically expands, as multiple
specialist subprojects will be undertaken; these may have direct impacts on other
________________________________
project facets, due to interdependency relationships within the Data Centre
infrastructure.
________________________________
________________________________
________________________________
________________________________
All project lifecycles should have the following elements, which should be specifically
________________________________
identifiable in the project plan:
►________________________________
Gates:The decision point prior to starting each phase
► Phases:The period of time during which work is undertaken
►________________________________
Milestones: Delimiting the end of a phase, other significant achievements or control
________________________________
points
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Every project requires an initiation phase to provide necessary direction. This will also establish the
________________________________
accountability and authority of the designated project manager. This is followed by the roles and
responsibilities of suppliers, project teams and key stakeholders. Organisational factors should be
________________________________
defined as follows:
► Limits of Authority: Understanding the limits of authority for any action taken
►________________________________
Communication Framework: Introduction of an effective communication mechanism to ensure
all stakeholders are engaged and aware of project progress
►________________________________
Incentive Mechanism: Internally, this might be represented by bonus and reward schemes;
externally, this might be contractual incentives
►________________________________
Project charter development: This provides the Project Manager with the authority to apply
resources to project activities; the PM should participate in its development. The trigger is a
________________________________
needs analysis, a business case, a response to a customer request or a Request for Quote (RFQ)
► Project purpose
________________________________
► Objectives
________________________________
► Requirements
► Description
________________________________
► Risk
► Milestones
________________________________
► Budget
►________________________________
► Identify stakeholders
Review lessons learned from previous projects
________________________________
The purpose of planning is to ensure the outputs from the project are likely to be
________________________________
delivered in sufficient time, within cost and at the required specification. The more
________________________________
planning undertaken before significant contractual obligations are implemented, the less
risk there is of failing to meet them.
It ________________________________
is also a good practice to plan the next phase of a project in detail towards the end
of the previous phase, whilst maintaining an outlined high-level plan for the entire
________________________________
project. The scope will be further developed and refined during planning, usually
revealing more information concerning task, cost, timescales and risk.
________________________________
Project planning should be an iterative process, as each element of the plan can
________________________________
influence other aspects. To fully appreciate this factor, planning usually comprises the
following activities:
►________________________________
Scope definition
► Work breakdown schedule
►________________________________
Work package activities
►________________________________
Sequence of activities
► Schedule development
________________________________
► Resource identification
►________________________________
Determining cost
► Plan confirmation
________________________________
► Ongoing monitoring
The execution is the implementation of the realisation of the design potential, through
________________________________
the project delivery phases. This is ultimately the critical point of any project, and needs
________________________________
to be undertaken in parallel with the monitoring and control process.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
It is essential the progress of the project is monitored correctly against the designated
________________________________
plan. This will ensure corrective actions are initiated to address variances, or initiate an
________________________________
escalation process. The monitoring and analysis of project data should enable the
project manager to address problems at an early stage and take advantage of
________________________________
opportunities that might benefit the project. Good communication between the
project manager, project teams and stakeholders is essential.
An________________________________
effective monitoring and analysis approach will assist in understanding the whole
project in regard to performance and objectives, plus any potential future impacts. This
________________________________
also ensures the scope and schedule are regularly reviewed, assuring potential risks are
identified and actions initiated at the earliest point. Specific reports may be required to
________________________________
support the monitoring and control process:
________________________________
► Actual costs reported against planned cost and variances
► Time & cost at completion
►________________________________
Earned Value Management (EVM)
________________________________
► Continual quality acceptance
► Progressive acceptance
________________________________
These activities and measures allow the project manager to make appropriate
judgments, implementing timely corrective actions to ensure the planned delivery is
________________________________
maintained.
________________________________
The objective of project closure is to ensure it is closed down in a controlled and organised
________________________________
way, and that all accountabilities relating to it have been discharged or handed over to the
appropriate operators. Closure is a formal endpoint to a project. The project closure should
________________________________
review:
________________________________
► The success of the project in terms of meeting the original time, cost and scope
► Confirmation the outputs of the project have been handed over (where appropriate) to
________________________________
ongoing operational management
► Verification the anticipated benefits have actually been built into the forecasted
________________________________
performance of the business
► Record and communicate any lessons which might be beneficial to future projects
________________________________
There are three key steps to closing a project:
________________________________
► Prepare the closure report
► Formally close the project
►________________________________
Administratively close the project
Following approval to close the project, the project manager should:
________________________________
► Finalise the project closure report
►________________________________
Prepare communication, attaching the approved closure report to the stakeholders,
confirming the decision to close the project
►________________________________
Complete any outstanding closure / handover actions
► Feedback and suggestions to improve designated project processes (i.e., lessons learned)
________________________________
The cornerstones, or triple constraints – in other words, how do we complete this project:
________________________________
► Without going bust
________________________________
► Without going to jail
► Without looking stupid
________________________________
► Time – This refers to the actual time required to produce a deliverable, which (in
this case) would be the end result of the project. Naturally, the amount of time
________________________________
required to produce the deliverable will be directly related to the number of
requirements in the end result (scope), along with the resources allocated to the
________________________________
project (cost).
________________________________
► Cost – This is the estimated amount of money required to complete the project.
Cost itself encompasses various factors, such as: Resources, labour rates for
________________________________
contractors, risk estimates, bills of materials, etc. All aspects of the project that have a
monetary component are part of the overall cost structure.
________________________________
► Scope – These are the functional elements that, when completed, make up the end
deliverable for the project. The scope itself is generally identified up front, to give the
________________________________
project the best chance of success. (Although scope can potentially change during the
project lifecycle, a concept known as ‘scope-creep’). Note that the common success
________________________________
measure for the scope of a project is its inherent quality upon delivery.
________________________________
The major takeaway from the triple constraint, being that it is a triangle, is that one cannot
adjust or alter a side of it without, altering the other sides. For example, if there is a request for
________________________________
a scope change midway through the project, the other two attributes (cost and time) will be
affected in some manner.
________________________________
The identification of key stakeholders who can directly or indirectly impact the design
implementation process is an important factor that needs to be determined.
________________________________
Some of these may have the support or influence either to block or advance a given
process. Some may be interested in what the design process is capable of providing,
________________________________
others may not care. This step enables the project team to easily see which
stakeholders are expected to be blockers or critics, and which are likely to be
________________________________
advocates and supporters of the design initiative.
________________________________
Work out stakeholder support and influence, to focus the design implementation
process on key individuals.These can be mapped onto an influence / support matrix.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
There must be appropriate communication and involvement with internal and external
________________________________
stakeholders throughout a project lifecycle. The involvement of partners (such as
________________________________
suppliers, users and customers) adds considerable value in all phases of a project. The
earlier the stakeholders are involved, the better the result. There are numerous ways
________________________________
to improve communication and interaction:
► Focus groups
►________________________________
Facilitated workshops
________________________________
► Early prototyping
► Simulations and presentations
________________________________
Involving these shareholders is a powerful mover for change, whereas ignoring them
can lead to painful processes, ultimately leading to a failed project.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Specific: Be very specific when setting your goals. The five Ws will help you create
your goal - Who does the goal involve? What is the goal? Why are you setting
________________________________
this goal? Where will you start? Which will you need?
Measureable: Now that you have been specific in setting your goal, it is time to
________________________________
ask how much? How many? How will I know if the goal is accomplished?
________________________________
Achievable: Are your goals achievable? Make sure they are attainable, or they will
not follow through. If you are able to complete your goals, it will encourage you to
________________________________
go a step beyond with your next goal, and so on.
Realistic: Make your goals realistic - Set yourself up for success by making sure
________________________________
they are.
________________________________
Timely: Setting a specific time to accomplish the goal brings clarity to what you
expect of yourself. Don’t skip this step - it’s key!
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Determining what to do is one thing, deciding when to do it is something else. What
you schedule first is determined by its urgency. Many businesses confuse importance
________________________________
with urgency, thinking that important tasks are ones that have to be done right away.
All tasks can be classified as indicated on the grid above. Tasks should be worked on
________________________________
based on their importance, with those listed as urgent and important being first.
Important items have intrinsic value in themselves; urgent items simply have a short
________________________________
timeframe on them. If there are more things to do than you can possibly get done, the
key is to be able to identify priorities and ignore things that are urgent but not
________________________________
important. This concept is important to understand when setting goals, planning, and
________________________________
scheduling.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Is the amount of time you are spending on something in-line with its importance? To
prioritise your tasks, enter them in the grid shown above in accordance with their
________________________________
importance and urgency. Urgent items are those that must be completed this week or
sooner, in order to take advantage of opportunities or results they can provide;
important items are those that will have a significant impact on your goals, or
________________________________
significantly influence your results in a positive way.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The use of a RACI matrix helps put in place a clear structure covering key elements;
these include clear communication and reporting channels, which are linked to the
________________________________
ownership of designated roles and responsibilities. These are all key elements that must
be in place to ensure a detailed and effective audit process is formulated.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
OK, at this point we move on to managing the design process…
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
For the project to be successful, the what, when, how, and by whom questions have to
be answered; they must then be subscribed to by the senior people within the
________________________________
organisation.
Details should be captured in the actions, issues and risks logs, as outlined in the
________________________________
earlier section relating to management methodologies.
________________________________
A road map might also prove helpful.
Circulation of (and request for) comments from the senior rung is a must, providing a
________________________________
sound base for the project to move forward.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A constraint is dealt with in greater detail in the afternoon session ‘Managing the
Design Implementation Process’, but for context’s sake, a constraint is anything that
________________________________
affects the cornerstones of project management – spec / scope, budget and time.
Actions, issues and risk logs are best used for formal and measurable communication.
By________________________________
formalising information in writing, one creates a log to use for lessons learned, as
________________________________
with all documented information.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The Theory of Constraints is a methodology for identifying the most important limiting
factor (i.e., constraint) that stands in the way of achieving a goal, and then
________________________________
systematically improving that constraint until it is no longer the limiting factor. In
manufacturing, the constraint is often referred to as a bottleneck.
________________________________
Identify: Identify the current constraint (the single part of the process that limits
the rate at which the goal is achieved).
________________________________
Exploit: Make quick improvements to the throughput of the constraint, using
________________________________
existing resources (i.e., make the most of what you have).
Subordinate: Review all other activities in the process to ensure they are aligned
________________________________
with and truly support the needs of the constraint.
Elevate: If the constraint still exists (i.e., it has not moved), consider what further
________________________________
actions can be taken to eliminate it. Normally, actions are continued at this step until
________________________________
the constraint has been “broken” (or it has moved somewhere else). In some cases,
capital investment may be required.
________________________________
Repeat: The Five Focusing Steps are a continuous improvement cycle. Therefore,
once a constraint is resolved, the next one should immediately be addressed. This
________________________________
step is a reminder to never become complacent – aggressively improve the current
constraint, then immediately address the next.
________________________________
________________________________
________________________________
Mandatory dependencies
Also referred to as hard logic
________________________________
Contract-based requirement
________________________________
Usually involve physical limitations
Determined by the project management team during activity sequencing
________________________________
________________________________
Discretionary dependencies
Also referred to as preferred logic, preferential logic, or soft logic
________________________________
Determined by the project management team during the activity sequencing
________________________________
Used with care and must be well-documented; may impact later scheduling
options
________________________________
________________________________
External dependencies
Determined by the project management team during the activity sequencing
________________________________
Relationship between project and non-project activities, generally outside project
________________________________
team control
________________________________
________________________________
Multidisciplinary project delivery teams can work brilliantly together or be totally
dysfunctional. It’s hardly surprising that when you assemble a diverse group of people
________________________________
with varied skills into a team, things don’t always go smoothly.
But multiprofessional teams are a fact of life in Data Centre design deliveries, with
________________________________
integration being an essential factor in the successful delivery of the design project.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
You will know from your own experiences there are various groups / tribes and
external groups (Architects, Building Contractors, Fire Chief, etc.). Internal groups
________________________________
include Facilities Management, IT, End-user groups, H&S and HR.
Group dynamics relate to culture and conflict in this context (conflict is dealt with on
________________________________
the next slide). We have all endured the laborious meetings where one tribe is intent
on impressing its self-importance over another, or indeed all others!
________________________________
The role of the Project Manager is to ensure communication is constructive (the
________________________________
‘peacemaker’ if you like); the PM facilitates the project to its end goal, on-time, within
budget and to the right quality (spec / scope).
All________________________________
will have their own views on theoretical methodologies, from great to abysmal. A
method often used in practice is that of Tuckman’s four stages of Forming, Storming,
________________________________
Norming, and Performing (FSNP).
________________________________
________________________________
________________________________
________________________________
________________________________
Forming - Stage 1
________________________________
High dependence on leader for guidance and direction. Little agreement on team aims, other
than received from leader. Individual roles and responsibilities are unclear. Leader must be
________________________________
prepared to answer lots of questions about the team's purpose, objectives and external
relationships. Processes are often ignored. Members test tolerance of system and leader.
________________________________
Storming - Stage 2
Decisions don't come easily within a group. Team members vie for position as they attempt to
________________________________
establish themselves in relation to other team members and the leader, who might receive
challenges from team members. Clarity of purpose increases, but plenty of uncertainties
________________________________
persist. Cliques and factions form and there may be power struggles. The team needs to be
focused on its goals to avoid becoming distracted by relationships and emotional issues.
________________________________
Compromises may be required to enable progress. Leader coaches.
Norming - Stage 3
________________________________
Agreement and consensus is largely formed among teams who respond well to facilitation by
leader. Roles and responsibilities are clear and accepted. Big decisions are made by group
________________________________
agreement. Smaller decisions may be delegated to individuals or small teams within group.
Commitment and unity is strong. The team may engage in fun and social activities. The team
discusses and develops its processes and working style. There is general respect for the leader,
________________________________
and some of leadership is more shared by the team. Leader facilitates and enables.
Performing - Stage 4
________________________________
The team is more strategically aware - the team knows clearly why it is doing what it is doing.
The team has a shared vision and is able to stand on its own feet, with no interference or
________________________________
participation from the leader. There is a focus on overachieving goals, and the team makes
most of the decisions using criteria agreed with the leader. The team has a high degree of
________________________________
autonomy, and requires delegated tasks and projects from the leader. The team does not need
to be instructed or assisted; members might ask the leader for assistance with personal and
interpersonal development. Leader delegates and oversees.
________________________________
Compromise and Collaboration.
Avoidance: This is the most frequently used strategy, along with accommodation.
________________________________
Here, conflict is avoided and when it does appear, the person using this strategy
refuses to engage in the situation.
________________________________
Accommodation: Here you take the conflict and submit.
________________________________
Compete: You play the person at his or her own game and work hard to get your
own way in the conflict.
________________________________
Compromise: A much more useful tactic to use - here you don't give into the
conflict, but work out a solution somewhere between the two sides.
________________________________
Collaborate: The most useful tactic, particularly with extremes of conflict such as
________________________________
bullying. The aim here is to focus on working together to arrive at a solution, where
both sides own and are committed to the solution.
________________________________
You might find the following link of interest:
________________________________
http://conflict911.com/guestconflict/difficultrelsatworkdealwithwpconflict.htm
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Using a predefined approach has been dealt with previously (i.e., the risk log). The
concept of RAG status has been used in the risk log example.
________________________________
The benefits of a risk log include recognising threats to the intended outcome -
formalising and measuring risk. The difference between risk and chance is you can
________________________________
measure risk, but you cannot measure chance!
________________________________
Risk logs are a sizeable contributor to documenting lessons learned at the close of the
project.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Identify risks: Determine those which may affect delivery of the design; these may
come from both internal and external factors.
________________________________
Analyse risks: Break down risks to fully understand their potential impact.
________________________________
Decide risk handling: How to protect your design from negative risk while trying
to take advantage of positive risks. For some risks, you may decide not to take
________________________________
action, but that is still a decision.
Add to or modify plans as necessary: If action is needed, it must be built into a
________________________________
plan.
Take planned action(s) and monitor risk: There is little point in planning
________________________________
actions if you do not take them. If the action is proving insufficient or the risk has
changed, the cycle must go back to analysing the risk, then go back through the risk
________________________________
cycle.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Risk management will be continued and dealt with in greater detail in the next module.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Risk management protects and adds value to the organisation and its stakeholders,
________________________________
supporting their objectives by:
►________________________________
Providing a framework for an organisation that enables future activity to occur in a
consistent and controlled manner
►________________________________
Improving decision-making, planning and prioritisation by comprehensive and
structured understanding of business activity, volatility and project opportunity /
________________________________
threat
________________________________
► Contributing to more efficient use / allocation of capital and resources within the
organisation
►________________________________
Reducing volatility in non-essential areas of the business
► Protecting and enhancing assets and company image
________________________________
► Developing and supporting people and the organisation’s knowledge base
►________________________________
Optimising operational efficiency
________________________________
________________________________
________________________________
________________________________
________________________________
Consider the effect of risk on the overall project.
Consider the combined effect of related risks.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
An issue is something that might affect the project critical path, but isn’t yet likely to
impact its overall success.
________________________________
Issues are managed by the project manager at the team level, and are prioritised with
accountability and due dates.
All________________________________
projects have issues - the larger the project, the larger the issues. Issues tend to
________________________________
take on a life of their own, so keep a historical record of activities associated with an
issue, especially decisions made along the way to resolve it. As a reminder, an issue is
________________________________
only an issue if it can delay the project critical path. A difficult action or work activity is
not necessarily an issue.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Both project management and change management support moving an organisation
from a current state, through a transition state, to a desired future state (for example,
________________________________
the delivery of a Data Centre design project).
Each design project you undertake requires some level of project management and
________________________________
change management.
________________________________
Project management focuses on tasks to achieve programme requirements
Change management focuses on the people impacted by the change
________________________________
Any change to processes, systems, project delivery structures and / or job roles will
have a ‘technical’ side and a ‘people’ side that must be managed.
________________________________
Project management and change management have evolved as disciplines to provide
________________________________
both the structure and tools needed to realise change successfully on the technical and
people side.
________________________________
These two disciplines are tools used to support any changes you undertake during the
design implementation.
________________________________
Anytime you alter processes, systems, organisation structures or job roles, you need an
________________________________
ordered approach to manage both the ‘technical’ side and the ‘people’ side of the
pending change, to ensure the stability of the design project delivery.
________________________________
________________________________
________________________________
The roadmap is the initial communication. Why? Because it outlines the vision or aim
of the project, who is to carry out the required tasks, and the timeframe. Thereafter,
________________________________
the Project Initiation Document (PID) becomes the guiding agreement (along with the
issues and risks log) which provides information for the weekly, fortnightly or monthly
________________________________
management and project team meetings; this relates to the bullet point above, on
formal methods for delivering the relevant message, ensuring it is reinforced. The
________________________________
project plan and organisation chart are also key documents.
________________________________
Miscommunication can be seen in two guises: Poor communication or intentionally
misleading communication, and let’s face it - it happens. The key is, why? Failure to
________________________________
manage cohesively will potentially lead to disgruntled tribes and conflict; out of conflict
falls intentionally misleading information.
________________________________
Manage and motivate the project team, rather than the individual tribes;
________________________________
miscommunication and the unhelpful process that leads to it will be removed. Poor
communication can result from speaking without clarity, or insufficiently conveying
________________________________
information; hence, the importance of the road map, formalising documentation and
circulating it to Workstream Leaders. By setting in motion fundamental project
________________________________
management governance, inadequate information is removed as a risk to the project.
Most would agree poor communication is a sizeable risk to every project.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Project Charter:
________________________________
Acquire authority to assign resources to tasks (to spend some money), identify
________________________________
stakeholders and their needs.Acquire the specification and start planning.
The purpose of the project charter is to document:
________________________________
Reasons for undertaking the project
________________________________
Objectives and constraints of the project
Directions concerning the solution
________________________________
The main stakeholders
________________________________
The three main uses of the project charter:
________________________________
To authorise the project - using a comparable format, projects can be ranked and
authorised by Return on Investment (ROI)
________________________________
Serves as the primary sales document for project ranking - stakeholders have a 1-2
page summary to distribute, present, and keep handy, to fend off other project /
________________________________
operation runs at project resources
As a focus point throughout the project - for example: Project as people walk into
________________________________
team meetings, and use in Change Control meetings to ensure tight scope
management
________________________________
________________________________
________________________________
A Risk Breakdown Structure (RBS) is a method of systematically identifying risk. A
brainstorming of risks should be carried out with as many key project team members
________________________________
as possible.
Quantitative Risk Management is contingency planning. It involves the probability
and________________________________
consequences of each identified risk.
________________________________
It’s not enough to identify just the probability or severity; both are required to
establish a ‘risk factor’. For example, look at volcanic ash from the perspective of a jet
________________________________
engine; we’re told it is unlikely (a very low probability) the engine will shut down.
Good. However, what is the severity? Catastrophic! Everyone dies. Calculating a risk
________________________________
factor enables prioritisation, not just of the risk, but the time of both the Project
Manager and project team.
In________________________________
summary, the probability and severity of risk identifies its magnitude.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Develop the risk mitigation plan early in the project planning process. One way to
avoid risk is to add an assumption in the scope of works to do so (e.g., “Provision of
________________________________
external connectivity services by A. N. OTHER” )
The type of contract chosen can also move risk from one party to another:
________________________________
Avoid – Eliminate, withdraw, or don’t become involved
________________________________
Retain – Accept and budget
Control – Monitoring, re-evaluation and compliance with decisions
________________________________
Deflect – To transfer risk to another organisation, individual, or entity
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The type of contract chosen can also move risk from one party to another.
Time and material puts the risk (but also the reward) with the client.
________________________________
Firm bid does the reverse.
________________________________
Most telecoms / datacoms firms use EF&I – design, furnish and install the solution;
good for sales-driven companies like us.
________________________________
Construction companies are profit-driven (look at the housing market as an example),
and therefore use EPCM. They design a new house, procure materials (at least some of
________________________________
them) and construction manage by outsourcing specific trades. That way, if profit is
threatened (just like the slump in the housing market where prices plummeted), the
________________________________
company contracts and reduces (or stops) production until profit returns. They can
survive, but the contracted-in trades suffer. In other words, for their marketplace,
________________________________
nearly all of the risk is passed to the brought-in trades.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The scope of works is an attempt to capture in writing what the project will deliver.
A good scope of works ensures you and your customer have the same vision of the
________________________________
project. It facilitates change control and a satisfied customer.
Customer Need: Do we all understand what the customer is trying to
________________________________
accomplish?
________________________________
Project Objective: What is our role? What are we actually going to deliver? (Cite
the Standards you plan to adhere to). Question: What is the difference between a
________________________________
Standard and a Code? Standards are mandatory, Codes are law.
Project Size: Is it a small, medium or large project?
________________________________
Project Schedule: When do we need to start and / or when do we need to finish?
________________________________
This will also dictate the number and competencies of the team.
Assumptions: Contains items from our risk analysis, and is really a “NOT TO DO”
________________________________
list. Assumptions are used to ‘build a fence’ around the project, to factor in risk. In
other words, the SOW identifies what you are going to do and what you are not
________________________________
going to do!
Scope Verification: How do we know it is correct? We need to ask the customer
________________________________
- is what we’re planning to deliver what is required?
________________________________
Scope Quality Control: How do we ensure deliverables are correct from an
internal perspective, i.e., specific solutions are to Standard, and carry a
________________________________
manufacturer’s warranty. Delphi technique uses a selected group of experts to
answer those questions and provide feedback.
________________________________
If you get the right team and match each member with the role they excel in, anything
is possible!
________________________________
Organisation Breakdown Structure (OBS) is a useful tool for organising people in a
hierarchical way, which does not necessarily reflect pay-grade or seniority, but fulfills
________________________________
the needs of the project by clearly illustrating the chain of command.
________________________________
The Organisation Breakdown Structure groups together similar project activities or
“work packages” and relates them to the organisation’s structure. OBS is used to
________________________________
define the responsibilities for project management, cost reporting, billing, budgeting and
project control. The OBS provides an organisational rather than a task-based
________________________________
perspective of the project. The hierarchical structure of the OBS allows the aggregation
(roll-up) of project information to higher levels. When project responsibilities are
________________________________
defined and work is assigned, the OBS and WBS are connected, providing the
possibility for powerful analytics to measure project and workforce performance at a
________________________________
very high level (example business unit performance) or down to the details (example
user work on a task).
________________________________
________________________________
________________________________
________________________________
________________________________
Becoming a successful manager requires knowledge of the management processes and
the Data Centre process – Codes, Standards, installation methods, telecommunications
________________________________
design, etc.
It requires management, leadership and communications skills.
________________________________
Select the right team members – based upon talent
________________________________
Set expectations – based upon outcomes
Motivate – focus on strength not weakness
________________________________
Develop the people – help them find the right fit; not just ‘moving up the ladder’
________________________________
Treat team members as they deserve
Praise in public, coach in private
________________________________
Give others credit
________________________________
Accept the blame
Show grace under pressure
________________________________
Don’t shoot the messenger
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
A large project is really a series of smaller ones integrated together.
Project integration management includes managing the execution of the plan.
________________________________
Integration management ensures coordination between the various teams and clean
handovers for each section of the project. The subdivision of project deliverables into
________________________________
smaller more manageable events is called decomposition.
________________________________
Decomposition is a process which:
Subdivides the project work packages into activities
________________________________
Provides better management control
________________________________
Leads to activities, the lowest level of the Work Packages (WPs) in the defined
activities process
________________________________
Leads to the WP, the lowest level of the Work Breakdown Structure (WBS), where
the deliverables are identified
________________________________
Process can involve Team Members. This can lead to better and more accurate
________________________________
results
In practice, the ideal is to break the project down into groups of tasks no longer than
________________________________
four to six weeks. Failure to do this can lead to a false sense of security for workers.
________________________________
________________________________
________________________________
A Work Breakdown Structure (WBS) in project management and systems
engineering, is a tool used to define and group a project's discrete work elements (or
________________________________
tasks) in a way that helps organise and define the total work scope of the project.
A WBS element may be a product, service, data, or any combination. A WBS also
________________________________
provides the necessary framework for detailed cost estimating and control, along with
guidance for schedule development and control. Additionally, the WBS is a dynamic tool
________________________________
that can be revised and updated as needed by the Project Manager.
________________________________
The Work Breakdown Structure is a tree structure, which shows a subdivision of effort
required to achieve an objective; for example, a program, project, and contract. In a
________________________________
project or contract, the WBS is developed, starting with the end objective -
successively subdividing it into manageable components; this done in terms of size,
________________________________
duration, and responsibilities (e.g., systems, subsystems, components, tasks, subtasks,
and work packages), which include all steps necessary to achieve the objective.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
‘On-time’ is a major customer requirement, and one of the triple constraints.
PERT is a method to analyse tasks involved in completing a given project, especially the
________________________________
time needed to complete each, identifying the minimum time needed to complete the
aggregate.
________________________________
PERT was developed primarily to simplify the planning and scheduling of large and
________________________________
complex projects. It was developed by the US Navy Special Projects Office in 1957 to
support the US Navy's Polaris nuclear submarine project. It was able to incorporate
________________________________
uncertainty by making it possible to schedule a project while not knowing precisely the
details and durations of all activities, using three estimates - most likely, optimistic and
________________________________
pessimistic. A calculation is then applied to best estimate the time required to
accomplish a task, assuming everything proceeds as normal (the implication being the
________________________________
expected time is the average duration the task would require if it were repeated on a
number of occasions, over an extended period of time).
E ________________________________
= (O + 4ML + P) ÷ 6
________________________________
If the difference in schedule or budget is greater than 10% between most likely and
pessimistic, you are probably in a high risk situation.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Critical Path Management calculates the longest path of planned activities to the end of
the project, including the earliest and latest each activity can start and finish without
________________________________
making the project longer. This process determines which activities are ‘critical’ (on the
longest path) and which have ‘total float’ (can be delayed without making the project
________________________________
longer).
In project management, a critical path is the sequence of project network activities
________________________________
which add up to the longest overall duration. This determines the shortest time
possible to complete the project. Any delay of an activity on the critical path directly
________________________________
impacts the planned project completion date (i.e., there is no float on the critical path).
________________________________
A project can have several, parallel, near-critical paths. An additional parallel path
through the network with a total duration shorter than the critical path is called a sub-
________________________________
critical or non-critical path.
These results allow Managers to prioritise activities for the effective management of
________________________________
project completion, and to shorten the planned critical path of a project by pruning
critical path activities, by ‘fast-tracking’ (performing more activities in parallel), and /
or________________________________
by ‘crashing the critical path’ (shortening the durations of critical path activities
by adding resources).
________________________________
Whilst all predecessor / successor relationships can be found in all projects, the most
common relationship by far is Finish Start.
________________________________
________________________________
An activity on the critical path has 'free float' if delaying its completion will not delay
________________________________
the next sequential task. An activity on the critical path has 'total float' if delaying its
________________________________
completion will not delay completion of the project. A critical path may have unused
time expressed as total float.
________________________________
Total float is associated with the path. If a project network chart / diagram has 4 non-
critical paths, then that project would have 4 total float values. The total float of a path
is ________________________________
the combined free float values of all activities in a path.
The total float represents schedule flexibility, and can also be measured by subtracting
________________________________
early start dates from late finish dates of path completion.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
The CDCDP now needs to be aware of the financial direction or limitation of a given
________________________________
project, as this will have a direct impact on the cost, time and specification. The key
________________________________
aspects are:
Capital Expenditure (CAPEX) - is money invested by a company to acquire or
________________________________
upgrade fixed, physical, non-consumable assets, such as buildings and equipment, or a
new business
________________________________
Operational Expenditure (OPEX) - refers to the day-to-day costs of operation
________________________________
Return on Investment (ROI) - a profitability measure that evaluates the performance
of a business by dividing net profit by net worth
________________________________
Total Cost of Ownership (TCO) - is the total cost of Data Centre computer assets
throughout their lifecycle, from acquisition to disposal
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Bringing the project in on-budget:
Project Cost Management involves cost estimation, cost tracking and budgeting. Most
________________________________
professional Project Managers use the Earned Value Method of Cost Management.
________________________________
Here’s a problem often faced by Project Managers - It LOOKS like we are well under
our budget; but what if you then found out only 20% of the work required by the
________________________________
project had been completed? The value of that work would be 20% of £120k, or £24K.
You’re not UNDERbudget, you’re OVER by £6K!
To________________________________
more accurately track their budget (and schedule), Project Managers use a method
called ‘Earned Value’.
________________________________
Handover and Progressive Acceptance
________________________________
Would you complete a project in its entirety before you ask the customer if he / she is
happy?
________________________________
Progressive acceptance is essentially good practice to ensure quality and scope are
measured throughout the project delivery - a staged implementation, with sign-off by
________________________________
the customer at strategic points. Thus progressive acceptance helps reduces risk,
ensure quality and helps keep the customer happy!
________________________________
________________________________
________________________________
________________________________
► Planned Value (PV) - Planned cost of the total amount of work scheduled to be
performed by the milestone date.
►________________________________
Actual Cost (AC) - Cost incurred to accomplish the work done to date.
________________________________
► Earned Value (EV) - The planned (not actual) cost to complete work that has
been done. Note this is the cost of the work performed, and it is the budgeted
________________________________
amount, not the actual amount.
It compares the PLANNED amount of work with what has actually been
________________________________
COMPLETED, to determine if COST, SCHEDULE and WORK ACCOMPLISHED are
progressing as planned.
________________________________
Work is ‘Earned’ (or credited) as it is completed; the project acquires value as work
________________________________
progresses.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
And one more just to note: Cost Schedule Index (CSI). CSI = CPI x SPI - the further
CSI is from 1.0, the less likely project recovery becomes.
________________________________
Now we can begin to integrate these two pieces of information and get an overall
picture. The Performance Indices are really just a way of calculating, on a percentage
________________________________
basis, where we are compared to plan.
________________________________
The CPI and SPI are statistically accurate indicators of final cost results. Unlike the
stock market, in the project management world, past performance is an indicator of
________________________________
future results. Once a project is 10% complete, the overrun at completion will not be
less than the current overrun. Once a project is 20% complete, the CPI does not vary
________________________________
from its current value by more than 10%. In other words, you have to get it right at the
start of the project and maintain performance.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Calculate CPI and SPI for weeks 2 and 3 to determine the status of this project:
CPI = EV divided by AC
________________________________
Week 2: 4500 / 5500 = 0.81
________________________________
Week 3: 6000 / 7800 = 0.76
Both weeks <1 = Bad news and worsening - overbudget.
________________________________
SPI = EV divided by PV
________________________________
Week 2: 4500 / 5000 = 0.9
Week 3: 6000 / 7000 = 0.86
________________________________
Both weeks <1 = Bad news and worsening - behind schedule.
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
________________________________
Calculate CPI and SPI for weeks 2 and 3 to determine the status of this project:-
CPI = EV divided by AC
________________________________
Week 2: 5500 / 6000 = 0.92
________________________________
Week 3: 8000 / 8000 = 1.0
________________________________
Week 2 at <1 = Bad news (overbudget). Week 3 = 1.0, therefore we are exactly on
________________________________
budget.
SPI = EV divided by PV
________________________________
Week 2: 5500 / 5000 = 1.1
________________________________
Week 3: 8000 / 7000 = 1.14
________________________________
Both weeks greater than 1 = Good news and improving (ahead of schedule).
________________________________
Week 2 is interesting – it is not always good to be ahead of schedule, as in this
example.The project has gone overbudget to “get ahead”.
________________________________
________________________________
________________________________