Condition Monitoring of Rotating

Electrical Machines
Instrumentation Requirements
Introduction
• One feature of condition monitoring is to detect impeding faults at
an early stage, capturing weak signatures in measurements that are
usually mixed with noise.
• Some impeding faults may manifest themselves in non-electrical
variables that are not normally used in control or protection.
• That the following tasks are essential elements in a condition-
monitoring system (non-intrusive approach):
1. The measurement or transduction task (sensing of primary
variables)
2. The data acquisition task (conversion of sensed variables into
digital data in condition-monitoring system).
3. The data processing task (identifying of information buried in data).
4. The diagnostic task (acting on processed data).
Contd..
• Tasks (1) and (2) are usually carried out while the plant being
monitored is operating.
• Tasks (3) and (4) can be performed off-line and the results of
these tasks may not be fed back immediately to the plant
operators.
• In an intrusive approach in which a signal is injected into the
plant being monitored and the response is used to control the
plant status.
• This is particularly attractive in plant involving power
electronic converters that generate monitoring signals and are
very fast acting.
Sophisticated Condition-monitoring System

• Let us consider the analogy of the engineer collecting their

meter readings, and try to identify the tasks.
• The meter deflections or the numbers displayed on the panels
will naturally be in response to measurements made elsewhere
in the plant under monitoring, and may be picked up by current
transformers, voltage transformers, thermocouples or
accelerometers (transduction task).
• The data acquisition task could be simplified to the action of
the engineer in writing down the series of readings from each
meter, together with information regarding the time, location,
loading and ambient conditions perhaps, of the plant under
monitoring.
Sophisticated Condition-monitoring System
contd..
• The processing task corresponds to the analysis of the readings in the same
way. It may be considered appropriate to average several of the readings for
example, and to present them in a way that allows easy comparison with
other data.
• Many of the more common processing techniques such as use of methods
such as spectral analysis, time averaging, and auto- and cross-correlation
• The diagnostic task operates on the results of the processing task in order
to recommend actions that will hopefully result in improved operational
readiness and performance, and improved maintenance scheduling.
• For example, on the basis of examining the average meter readings and
comparing them with the manufacturer’s operational limits, say, the
engineer decides for the next plant shutdown which items of machinery
must be overhauled or replaced.
• For the manufacturer to set the operational limits, a relationship between
the plant condition and the signature during operation should have already
been established.
• The diagnostic task is still most often carried out with significant input from
the human operator, but with the development of knowledge-based expert
systems or other forms of artificial intelligence.
Temperature Measurement
• The three principal methods of measuring temperature
electronically:
• Resistance temperature detection:
A Resistance Temperature Detector (also known as a
Resistance Thermometer or RTD) is an electronic device
used to determine the temperature by measuring the
resistance of an electrical wire.
• Thermistors
A thermistor (or thermal resistor) is defined as a type of
resistor whose electrical resistance varies with changes in
temperature. Although all resistors’ resistance will fluctuate
slightly with temperature, a thermistor is particularly sensitive
to temperature changes.
• Thermocouples
A thermocouple is defined as a thermal junction that functions
based on the phenomenon of the thermoelectric effect, i.e. the
direct conversion of temperature differences to an electric
voltage.
Resistance Temperature Detection
• The variation of resistance of the
metal with the variation of the
temperature is given as,
Rt=R0[1+α(t-t0)+β(t-t0)2+..]
• Where, Rt and R0 are the resistance
values at toC and t0oC temperatures;
α and β are the constants depends
on the metals (Temp. coefficients).
• For small range of temperature, the
expression can be,
Rt=R0[1+α(t-t0)]
• Platinum and nickel are usually used
because they are retardant to
corrosion and because they have
sensitive resistance versus
temperature characteristics.
Platinum has the temperature range
of 650oC, and then the Copper and
Nickel have 120oC and 300oC
respectively.
Resistance Temperature Detection Contd..
• An RTD can be made by either winding insulated metal wire around
a thin cylindrical former or by evaporating a thin coating of the
metal on an insulating.
• A commercial sensor enclosed in a thin stainless steel tube for
protection. Such devices are widely used for gas and liquid
temperature measurement applications up to 1 000 ◦C and are the
usual choice for electrical machine manufacturers for insertion
between winding conductors in machine slots.
• RTD devices are constructed generally to a base resistance of 100Ω
at 0◦C and have the advantage of being linear over a wide operating
range. They have very good accuracy and precision but have a
relatively low sensitivity.
Four-wire Configuration of Wheatstone Bridge
• The circuit converts the change of
resistance into the change of output
voltage.
• The sensitivity is defined as the ratio
of change of output to the change of
input while the linearity is defined as
the deviation from the straight line
between the two end points of the
whole range, that is from 0–100 ◦C.
• It is clear that increasing the value of
R causes the sensitivity to be
reduced but the linearity to be
improved. The self-heating error is
reduced as the value of R is
increased.
• Wheatstone bridge arrangement
allows the lead resistance to be
easily compensated
Thermocouple
• Temperature measurement using thermocouples is based on the
well-known Seebeck effect whereby a current circulates around a
circuit formed using two dissimilar metals, when the metal junctions
are held at different temperatures.
• If a break is made in one of the wires then a voltage is generated
across the break, which increases with the temperature difference
between the junctions.
• One junction is held at a temperature that can be easily measured to
allow the so-called reference junction compensation.
• By doing this the need for a carefully controlled reference junction
temperature is avoided and a device with effectively only a single
junction results.
• As the thermocouple output is already in the form of a voltage,
there is no need to use additional circuit such as a Wheatstone
bridge for signal conversion
• Multiple thermocouples can be combined to form a thermopile of
higher sensitivity.
Reference Junction Temperature
Compensation Circuit

• The circuit uses an LM35 integrated circuit whose output

voltage is 0 V at 0 ◦C and increases at a rate of 10 mV/◦C.
• A voltage divider is included to cause an approximately 40
μV/◦C voltage gradient across resistor R2 that forms part of the
measured output voltage.
• The voltage across R2 will then compensate the voltage
developed across the cold junction marked as ‘3’ in Figure ,
assuming that the cold junction is at room temperature,
around 18–23 ◦C
Thermistor
• Thermistors are manufactured from blends of metal oxides of
cobalt, iron, titanium and nickel, which are fired like clays into small
discs or beads that may be encapsulated in resin or enclosed in
protective brass tubes.
• The working principle of a thermistor is that its resistance is
dependent on its temperature. We can measure the resistance of a
thermistor using an ohmmeter. If we know the exact relationship
between how changes in the temperature will affect the resistance
of the thermistor – then by measuring the thermistor’s resistance
we can derive its temperature.
• Thermistors exhibit a large change in resistance as a function of
temperature. In addition to sensitivity, they also have the
advantages of high stability, fast response and very small physical
size.
• The size advantage means that the time constant of thermistors
operated in sheaths is small. But the size reduction also decreases
the heat dissipation capability and so makes the self-heating effect
of greater importance.
Thermistor Characteristics

• The nominal values of resistance 10 kΩ, 20 kΩ, 30 kΩ or 100 kΩ are

defined for a room temperature of 25 ◦C.
• The vertical axis is in logarithmic scale implying that the change of
resistance with temperature is indeed significant.
• Since the accuracy of the device is hardly high and the characteristic
may also slightly drift after long-time thermal cycling, the signal
conversion scheme adopted for a thermistor is usually quite simple.
Thermistor Types
• There are two types of thermistors:
Negative Temperature Coefficient (NTC) Thermistor
Positive Temperature Coefficient (PTC) Thermistor
The relationship between resistance and temperature in an NTC
thermistor is governed by the following expression:
RT=R0eβ(1/T-1/T0)
• RT is the resistance at temperature T(K), R0 is the resistance at
temperature T0(K), T0 is the reference temperature (normally 25oC),
β is a constant, its value is dependent on the characteristics of the
material with nominal value 4000.
• When current passes through a device it will cause a small amount
of resistive heating. If the current is large enough to generate more
heat than the device can lose to its surroundings then the device
heats up. In a PTC thermistor, this heating up will also cause its
resistance will increase. This creates a self-reinforcing effect that
drives the resistance upwards, therefore limiting the current. In this
way, it acts as a current limiting device – protecting the circuit.
Fibre-optic Temperature Sensing

• When a laser pulse is injected into the optic fibre, it is subject to

scattering as it travels and the back scattered pulse is returned.
• The back-scattered light consists of a Rayleigh component, a Brillouin
component and a Raman component, which differ in wavelength.
• Thermally influenced molecular vibration causes the Raman-
scattered component to change and therefore it is sensitive to
temperature.
• According to the time when the signatures in the back-scattered light
are received, the temperature along the optical fibre can be
detected.
• The temperature resolution is typically less than 0.1 ◦C in the range
of zero to a few hundred degrees celsius. The response time of fibre-
optic temperature sensing is around 2 s.
Vibration Measurement
• Vibration based condition monitoring refers to the use of in situ non-
destructive sensing and analysis of system characteristics – in the
time, frequency or modal domains – for the purpose of detecting
changes, which may indicate damage or degradation.
• It revolves around the measurement of three quantities that are
related by numerical integration or differentiation:
• displacement,
• velocity,
• acceleration
• Which quantity one should measure depends on the size of the
plant being monitored, and the frequency range in which one is
interested.
• Rotating machinery analysis is primarily concerned with signals
relating to the rotational speed and its harmonics. Orders are
harmonics of the rotational speed. Evaluating orders can be critically
important on many types of rotating machines.
Frequency Ranges For Vibration
Measurements

• As the vibration frequency increases it is likely that

displacement levels will fall but acceleration levels will rise.
• So with increasing frequency it is better to progress from a
displacement device to a velocity transducer, and ultimately to
an accelerometer.
• the process of numerical integration attenuates measurement
noises while numerical differentiation amplifies them.
Transducers for Vibration Measurement
• The vibration monitoring usually consists of Transducers
attached at specific location and at specific direction.
• The reasons behind the excessive vibration include-unbalance,
misalignment, worn gears or bearings, looseness etc.
• These are contact or non-contact type and works on principle
of electromagnet or capacitive or strain gauge.
Displacement Transducers
• Non-contact type device uses high-
frequency current/voltage source
to generate an electromagnetic
field at the probe tip.
• The system energy, due to eddy
currents in the target, is
dependent upon the local
geometry of area surrounding the
probe tip.
• This change in system energy is
readily measured as the change of
voltage/current in the high-
frequency excitation circuit is
related to the displacement of the
target surface from the probe tip.
• Sensitivities in the order of 10
mV/μm displacement are easily
achievable with displacement
probes
Displacement Transducers contd..
• The output of the system will depend not only on the
displacement between the probe and the target, but also on
the material from which the target is made.
• This is because the eddy current reaction of the target, and
hence the system energy, is dependent upon the conductivity
and/or permeability of the material.
• Displacement is most effectively measured at the lower
frequencies even though the frequency range of eddy current
systems can extend above 10 kHz.
• Care must also be taken when mounting the probe to ensure
electrically conducting and magnetic surfaces around the
probe tip do not cause unnecessary disturbance of the applied
high-frequency field.
Non-contact Device with Fibre Optic
• Detects the light reflection
and compares it with the
injected light to determine
the distance between the
reflection surface and the
optic-fibre end.
• The result is converted into a
voltage signal using a
photodiode in the receiver.
• The contact-free
measurement range is up to
0.3 mm with a typical
resolution of 0.01 mm.
• Fibre-optic devices are
intrinsically immune to
electromagnetic noise
problems
Velocity Transducers
• Velocity is most usefully sensed in the
frequency range from 10 Hz to 1 kHz.
• This is usually achieved in an analogue
manner by designing a spring mass
system with a natural frequency less
than 10 Hz, and letting the mass take
the form of a permanent magnet.
• The device is responsive regardless of
the range of displacement involved in
vibration.
• The magnet is then surrounded by a
coil that is securely attached to the
housing.
• Whenever the housing is placed in
contact with a vibrating surface the
housing and coil move with respect to
the magnet and cause an EMF to be
induced in the coil (e=Blv)
Accelerometers
• Accelerometers are rigidly
fastened to the body
undergoing acceleration.
• For each accelerometer there is
a sensing axis aligned with the
direction of the acceleration
that is intended to be
measured.
• Accelerometers produce an
electrical output that is directly
proportional to the
acceleration in the sensing axis.
• The output should be low if the
acceleration is applied at 90◦ to
the sensing axis.
• Accelerometers are full-contact
transducers typically mounted
directly on high-frequency
elements, such as rolling-
element bearings, gearboxes,
or spinning blades
Piezoelectric Accelerometer

• When it is subject to vibration, the seismic mass, which is held

against the piezoelectric crystal element, exerts a force upon it
• This force is proportional to the acceleration.
• the piezoelectric crystal, which is usually a polarised ceramic
material, generates a proportional electric charge across its faces.
• The output can then be conditioned using a charge amplifier and
either incremental velocity or displacement signals recovered by
integration
• Electrical impedance of a piezoelectric crystal is itself high, the
output must be measured with a very high-impedance instrument to
avoid loading effects
Accelerometer Response
• The natural frequency of the
device is designed to be above
the usual operating range
• This limits the useful operating
range to around 30 per cent of
the natural frequency.
• As the output is low at low
frequencies the normal range of
application of accelerometers is
approximately 1–8 kHz
• Small devices may have ranges
extending beyond 200 kHz and
integrated amplifier
• Extremely wide range of
piezoelectric accelerometers
from very small devices that will
measure shocks of high
acceleration (in excess of 106
ms−2) to large devices with
sensitivities greater than 1 000
pC/ms−2
Force And Torque Measurement
• One of the commonest ways of
measuring force is to use a strain
gauge,.
• A simple device that comprises a
long length of resistance wire
formed into a zigzag shape and
securely bonded to a surface that
will alter shape elastically under the
action of the force.
• When the gauge is stressed under
the action of the force, the cross-
section and length of the wire
changes so that its resistance alters.
• The input–output relationship of a
strain gauge is expressed by the
gauge factor, G, which is defined as
the relative change in resistance, R,
for a given value of strain, ε :