Basics in Process Design

Frej Bjondahl
2006
Contents
Basics in Process Design ...............................................................................................1
Contents .........................................................................................................................2
1. Introduction............................................................................................................3
2. Design/analysis of process systems using mass and energy balances ...................3
2.1 Balance boundary, input, output, recirculation ..............................................4
2.2 Batch, continuous, steady state, dynamic process .........................................5
2.3 Mixing, chemical reaction, separation, recirculation, accumulation .............6
2.4 Heating, cooling, heat of reaction, phase change, pressure change ...............7
2.4.1 Example – mass balances.......................................................................7
2.4.2 Example – energy balances....................................................................8
3. Transport systems for of fluids ..............................................................................9
3.1 Pressure loss in piping, pipe element resistance ............................................9
3.2 Local pressure in pipe systems ....................................................................13
3.3 Pipe system characteristic curve ..................................................................13
3.4 Pump types, centrifugal, displacement ........................................................14
3.5 Pump characteristic, work point, power, efficiency, recalculation models .14
3.6 Pump and pipe systems, flow control ..........................................................16
3.6.1 Example pipe system ...........................................................................18
3.6.2 Example pipe and pump system ..........................................................19
3.7 Gas transport with fans in ducts...................................................................21
3.8 Gas transport with compressors ...................................................................21
3.9 Compressor types, centrifugal, displacement ..............................................21
3.10 Compressor calculations ..............................................................................21
3.10.1 Example – compressors .......................................................................24
3.11 Typical flow velocities.................................................................................24
4. Heat exchanger systems .......................................................................................25
4.1 Heat exchanger types ...................................................................................25
4.2 Heat exchanger modeling, counter current, co-current, real world .............25
4.3 Overall heat transfer coefficient...................................................................27
4.4 Condensing steam, evaporation ...................................................................27
4.5 Heat exchanger networks, pinch analysis ....................................................27
4.5.1 Example ...............................................................................................28
5. Economic analysis – modeling and selection ......................................................30

2
1. Introduction
This course deals with some of the tools you can use for analysis and design of proc-
ess systems as well as modeling of some basic and essential components or units used
in such designs. The important part in process design is the system, not its individual
units.
The vocabulary used in process design is also good to know and important words or
phrases are therefore highlighted in the text using italic.
Process design often involves large systems. These will have to be broken down into
smaller parts in order to be able to model and analyze these parts both individually
and together with the entire system. Information and knowledge from many sources
are combined and during the design you generally move from an overall view or goal
into ever increasing detail until you have managed to create a complete process de-
sign.

2. Design/analysis of process systems using mass

and energy balances
Consider this process design problem – you have been asked to design a process for
the production of a certain amount of product. Your task is to find out how to produce
the product and then to design a process for the production.
First you find out what it is that you are supposed to do. You find out from your em-
ployer/customer what they expect from the process.
Using literature, by making experiments and so on you find out what is needed to ac-
complish the process. You can gather information about possible raw materials, nec-
essary raw materials processing, what reactions are taking place, what kind of mixing
and separation equipment can be used, the conditions in the process like required
temperatures and pressures in each step, possibilities/necessity for recirculation of un-
reacted raw material, generated waste or byproducts as well as properties for each ma-
terial that goes through the process, and so on.
Once you have found out what kind of process steps are needed and what the condi-
tions in the process steps should be you have to decide how large each piece of
equipment should be and/or how much raw material, utilities and energy is needed. A
basic tool you can use for this step in the design is the mass- and energy balance.
Once you know the mass flows and energy requirements you can select equipment of
suitable sizes for your process. You can also estimate what the process will cost to
build and operate.
Mass and energy balances can also be used for analysis of processes – you have an
existing process and want to analyze it in some way. You might want to study what
happens if you make changes in the process. Mass and energy balances can be used to
determine the accuracy of measurements, determine missing pieces of information or
to determine where to make additional measurements.

3
2.1 Balance boundary, input, output, recirculation
The first thing you need to do in order to make a mass and energy balance is to decide
where the boundary for the balance is.
You can place the boundary anywhere but some ways of choosing makes it easier to
use the balance.
Let us consider a process where the raw materials A and B are mixed and react ac-
cording to

A + 2B ↔ C

Assuming that the reaction is not complete, some of the raw material remains unre-
acted and it is separated from the product C. The unreacted raw material is recircu-
lated back to the beginning of the process.
In order to visualize the process you can draw a simple block flow diagram (see
Figure 1) that shows the material flows and the process steps. You can also give
names (or numbers) to the flows and the process steps and write known or given data
regarding the flows or the process conditions in the diagram.

Figure 1. Block flow diagram of a simple process. The three major process steps have been
named Mixing, Reaction, and Separation. The flows have not been named but the components in
each flow have been listed.

Some possible ways of choosing balance boundaries for the process in Figure 1 are
shown in Figure 2. These are
a) an overall balance boundary around the entire process (including the recircu-
lation)
b) balance boundaries around each process step (Mixing, Reaction, Separation)
c) partial mass balances for components A, B or C using the same balances
boundaries as in a) and b)
Sometimes you only need one balance to find out what you need while at other times
you might need all these balances as well as additional information. The additional
information needed can be the stoichiometry of the chemical reaction, limiting condi-
tions for the reaction, the extent of reaction, the separation efficiency, the recirculation
ratio and so on. The solution can be found by solving a single equation or by solving a
system of many equations either algebraically or numerically.

4
Figure 2. Balance boundaries around the entire process (red) and around the individual process
steps (green).

2.2 Batch, continuous, steady state, dynamic process

Process systems for flexible production of small amounts of product are often batch
processes. Each step takes place in a vessel for a certain time. The same vessels can
often be used in the production of several products in different processes. The mate-
rial is moved from one step to the next step only after the previous step has been com-
pleted.
Modeling of batch processes can be made as calculations either for one batch as it
travels through the different steps in the process or for time averages for many
batches.
In process systems for continuous production of large amounts of product the material
will flow continuously through the process equipment. Stops are avoided whenever
possible. These processes can either be modeled as steady state processes or as dy-
namic processes.
A steady state process model gives an approximation of the plant up and running
without any disruption to the process. It is an idealization of a real process and it is
used for design specifications/calculations.
A dynamic process model is a more accurate model of the actual process since an ac-
tual process is never at steady state. There are always small changes and disturbances
that make the process conditions and flows fluctuate. The control systems of the proc-
ess are intended to take care of the changes. There can also be large upsets or planned
stops and starts of the process. The effects of all these changes, small or large, go
though the process and they are experienced with time delays at different parts of the
process.

5
2.3 Mixing, chemical reaction, separation, recirculation, ac-
cumulation
Mixing, chemical reaction and separation are some of the most basic steps in a
chemical engineering process that must be taken into account in mass and energy bal-
ances.
The mass balance model for mixing is easy if you assume that no reaction takes place.
All components that enter the mixer come out again. The energy balance might be
more difficult since the flows into the mixer can be of different composition, pressure,
temperature and phase. The modeling becomes more difficult if some kind of reac-
tion, dissolution or change of phase takes place during mixing.
In the reaction step you take into account that some components react and disappear
producing other components. The reaction stoichiometry is used to keep track of the
balance between reactants and products. The total mass flow out is the same as the
flow in if you assume steady-state conditions. The extent of reaction (or conversion)
needed for partial mass balances is more difficult to model since it also depends on
the process conditions (like residence time in the reactor, concentration of reactants
and products, pressure, temperature, presence of catalysts, mixing, particle sizes and
so on). The raw material flows, the heat of reaction, reactor heating/cooling all affect
the energy balance.
Modeling of the separation step is also difficult since the separation efficiency de-
pends on the choice of separation method and the conditions in the process step. Sepa-
ration is rarely perfect, you almost always loose raw materials in the product flow and
you separate and throw away/recirculate product. Splitting, distillation, filtering, crys-
tallization, chromatography, stripping, flashing are all examples of separation meth-
ods.
Recirculation of material in the process does not make the mass and energy balance
modeling more difficult but it can make the calculations more difficult. You quite of-
ten need to make iterative numerical calculations to reach a result. When recirculation
is used you also recirculate impurities that can accumulate in process. In order to keep
the concentration of these impurities low you might have to purge some of the recir-
culated material as waste.
Usually you assume that the process is in steady state i.e. that no accumulation takes
place in the process steps (at least in the long run). Sometimes accumulation takes
place anyway and you have to take that into account. This is especially true when you
model existing processes using measured data.
The mass balance calculations require that you can convert material flows given as
mass, mol and volume between each other depending on the models used. The com-
position of the flows might be given as concentration as well mass, mol and volume
ratio depending on the circumstances when different reaction and separation models
are used together with the mass balance models.

6
2.4 Heating, cooling, heat of reaction, phase change, pres-
sure change
The material flows might also have to be heated or cooled. You might want to start or
stop chemical reactions or change the physical or transport properties of the flows.
The reactions taking place might require heat or produce heat depending on the reac-
tion. The process might give off or take up heat from the surroundings.
The reactions, separation or the transport might also cause or require changes in pres-
sure using pumps, fans, compressors, valves and flashes. All these steps or phenom-
ena affect the energy flows in the process and they are be accounted for in the energy
balances.
You must be able to use thermodynamic models for calculation of enthalpy of com-
pounds, enthalpy of reaction, enthalpy of evaporation, enthalpy of dissolution, en-
thalpy of mixtures, and so on when you calculate the energy flows with the materials
and the necessary external energy flows needed to accomplish changes in these. En-
ergy can also be recirculated within the process when warm flows need cooling and
cold flows need heating.

2.4.1 Example – mass balances

In a factory the production rate of ammonia is 1000 kg/h. Ammonia (NH3) can be pro-
duced from hydrogen gas (H2) and nitrogen gas (N2) using the Haber-Bosch process.

N 2 (g) + 3 H 2 (g) ↔ 2NH 3 (g)

The molar mass of nitrogen gas is 28.01 kg/kmol, the molar mass of hydrogen gas is
2.02 kg/kmol, and the molar mass of ammonia is 17.03 kg/kmol. The conversion of
nitrogen in the reactor (here defined as mol nitrogen gas reacted)/(mol nitrogen gas in)
is assumed to be about 17 %. Unreacted hydrogen and nitrogen is separated and recir-
culated to the reactor.
a) How large are the flows in this process?
b) What is the composition of the flows?
The block flow diagram below shows the main steps in the process.

Mixer Reactor Separator

3 4 5
2

7
2.4.2 Example – energy balances
In the previous example we have calculated the mass flows in the process for produc-
tion of ammonia. We will now continue to calculate the energy flows in the process.
The temperature of the flows into the mixer is assumed to be 700 °C and the pressure
30000 kPa. The temperature in the reactor should be 500 °C and the pressure
6

Mixer Reactor Separator

3 4 5
2

30000 kPa. The pressure of the recirculation is also assumed to be 30000 kPa. The
separation is done by cooling the gases from the reactor to -20 °C which will make the
ammonia liquid and possible to separate from the gaseous nitrogen and hydrogen. The
enthalpy of reaction is assumed to be 92.5 MJ/kmol ammonia. The reactor can be
cooled or warmed in order to keep the temperature correct.
Calculate the energy flows with the mass flows and the heat flows that have to be
added or removed in the reactor and separator. The specific heat capacity at constant
pressure, cp , for nitrogen is about 1.88 kJ/kg°C, hydrogen 14.75 kJ/kg°C, gaseous
ammonia 2.60 kJ/kg°C. The heat of evaporation for ammonia at -20 °C is about
1320 kJ/kg.

8
3. Transport systems for of fluids
The processes in a chemical plant involve a lot of process steps and in particular trans-
port of material between these steps. Sometimes the transport systems are also respon-
sible for changing the pressure of the materials in the process. The design of these
transport systems is a large part of the design of process systems.
The mass balances give the sizes of the flows in different parts of the system. The
conditions required in the different parts of the system gives information for the calcu-
lation of the transport properties of the material flows.

3.1 Pressure loss in piping, pipe element resistance

When liquids or gases flow through pipes they experience a resistance due to wall
friction which slows down the flow. There is also additional resistance in the parts of
the pipe system where the flow has to change direction or velocity like in pipe bends,
valves and other pipe fittings. The resistance causes energy in the form of pressure to
become heat and this is experienced and measured as a pressure loss in the pipe sys-
tem as the fluid flows through it.
It is possible to model the resistance in pipe systems or parts of pipe systems. You can
calculate the total pressure loss in the system when a certain volume flow flows
through the system. You can calculate the local pressure at any point in the pipe sys-
tem. The volume flow (velocity) together with inlet and outlet pressure, inlet and out-
let height, fluid density and viscosity as well as any phase change determine the pres-
sure at any point in the system given the configuration of the system. Pumps (for liq-
uids) and fans (for gases) are used when the pressure has to be increased while valves
are used when the pressure has to be decreased to reach a certain flow.
The energy flow balance for any part of a pipe system (assuming a steady state, adia-
batic system and only one inlet and one outlet) is made up of the flows of potential
energy, kinetic energy, pressure-volume work, inner energy and added pump work.

1 2
m ⋅ g ⋅ z in . + ⋅ ξ in ⋅ m ⋅ win + Vin ⋅ p in + m ⋅ u in + Pwork
2
1 2
= m ⋅ g ⋅ z out + ⋅ ξ out ⋅ m ⋅ wout + Vout ⋅ p out + m ⋅ u out
2

where m is the mass flow through the pipe, g is the acceleration due to gravity, z in

and z out the height at inlet and outlet, win and wout are the average velocities (see
later calculation of average velocity), ξ in and ξ out are correction factors for calcula-
tion of kinetic energy using average velocities, Vin and Vout the volume flow, pin and
 

pout the pressure, u in and u out the inner energy of the fluid and Pwork the work added
to the fluid by a pump, fan or compressor.

9
For liquids (and also for gases at moderate pressure changes) the density does not
change much due to change in pressure and the volume flow in and out will be almost
the same. The equation can the be rearranged into

V ⋅ ( p in − p out ) + m ⋅ g ⋅ ( z in − z out ) + Pwork = m ⋅ (u out − u in ) +



 

1
2
( 2
⋅ m ⋅ ξ out ⋅ wout − ξ in ⋅ win


2
)
The pump work is added to the energy flow as pressure-volume work

Pwork = ∆p pump ⋅ V


The power that is actually provided to the pump depends on the pump efficiency

Pwork
Ppump =
η pump

The change in inner energy is due to friction and the friction causes the fluid that is
being pumped to heat up. This change is almost linearly proportional to the kinetic
energy of the flow and it is usually modeled using friction factors according to

1
m ⋅ (u out − u in ) ≡ ⋅ m ⋅ w2 ⋅ ζ


2
 

where  ζ is the sum of friction factors in this part of the pipe system. Division of
the energy flow balance by the volume flow (possible only if the density is close to
constant as it is for liquids but also for gases at moderate pressure change) makes it
possible to express the energy balance using pressure

( pin − pout ) + ρ ⋅ g ⋅ (z in − z out ) + ∆p pump =

1
2
⋅ ρ ⋅ w2 ⋅
1
(
ζ + ⋅ ρ ⋅ ξ out ⋅ wout 2 − ξ in ⋅ win 2
2
)

If the pipe has the same inlet and outlet velocity (i.e. the same density and pipe diame-
ter) then the kinetic energy in and out is the same and the equation can be simplified
into

1
( pin − p out ) + ρ ⋅ g ⋅ (z in − z out ) + ∆p pump = ⋅ ρ ⋅ w2 ⋅ ζ
2

The average velocity can be calculated using the volume flow and the cross section
area of the pipe.

V
w=
A


10
The sum of friction factors is made up of the sum of the friction factors for different
pipe elements and the wall friction which also depends on the length and diameter of
the pipe

l
ζ = ζ elements + ⋅ζ d
 

The flow resistance depends on the roughness of the pipe and it can be estimated us-
ing

1.094


0.25
10 6
1000 ⋅ k
ζ d = 0.01 ⋅ + 18.7 ⋅



 

Re d


where Re is the Reynolds number and k is the surface roughness. The Reynolds
number can be calculated as

w⋅ d w⋅d ⋅ ρ
Re = =
ν η

where w is the flow velocity, d is the diameter, ν is the cinematic viscosity, ρ is the
density and η is the dynamic viscosity. The flow is considered to be laminar if the
Reynolds number is below 2300 (the correction factor ξ is then about 2). Above that
the flow is usually turbulent (and ξ is about 1.1).
Table 1. Surface roughness for some common surface types. Most clean pipes have very low sur-
face roughness while the surface roughness of dirty or corroded pipes can be up to a 100 times
larger.

Type of surface Surface roughness, k

Smooth pipe 0.02 – 0.05 mm
Galvanized pipe 0.1 – 0.2 mm
Smooth cement 0.2 – 0.5 mm
Rusted iron, concrete 1 – 2 mm

If a pipe or duct is not circular (square, rectangular or any other shape) then the hy-
draulic diameter can be used instead of the diameter. It is defined as

4⋅ A
dh =
lp

where A is the cross section area of the pipe or duct and l p is the circumference of
the pipe or duct.
The pressure loss in any single element in the pipe can be modeled as

11
 1
∆p element = ⋅ ρ ⋅ w 2 ⋅ ζ element
2

where ζ element is the friction factor of a pipe element.

The elements can be pipe bends, pipe elbows, valves, diameter changes, pipe splits,
pipe joints, flow meters etc. The friction factors for a number of common elements are
shown in Table 2.
Table 2. Friction factors for some common pipe elements.

Element Image Resistance, ζ element

Sharp inlet
0.5

Rounded inlet 0.1

Pipe inlet 1–3

Outlet
1.1

Size decrease A
2


1− 1


A2


Size increase 1 A
!

⋅ 1− 2
"#$

"

2 A1 

Venturi 2
*
'

1 A
*
'

⋅ 1− 1
(

5 A2
(

) &

) &

Knee bend 1.2

Elbow bend R
1 2 10
d
ζ element 0.5 0.3 0.2

The flow resistance in valves is usually given as a valve capacity ( K v or C v ) at dif-

ferent valve openings L where L is 100 % when the valve is fully open and 0 %
when it is closed. The capacity is the flow through the valve at a specific pressure
drop. The pressure loss in a valve when a specific volume flow passes through the
valve can be estimated using
2
ρV
-

∆p valve = 100 kPa ⋅ ⋅

./0

kg K v
1 +

997 3
m

The valve capacity C v can be recalculated to K v using

12
 Kv Cv
≈ 0.866 ⋅
567 2 567 2

m /h
3
US Gal/min
34 34

3.2 Local pressure in pipe systems

Given an inlet pressure and a flow into a pipe section you can calculate the pressure at
the end of the section and use that as the inlet pressure in the following pipe section
and thus calculate the pressure at any point in the pipe.
129

128

127

126

125

pressure (kPa)
124

123

122

121

120

119
0 2 4 6 8 10 12
pipe le ngth from inle t (m)

Figure 3. The curve shows the pressure at different positions along the length of the pipe system
shown to the left. The pipe system has been divided into sections (numbered 1 to 7) and the pres-
sure out from each section has been calculated using the inlet pressure to the section and the vol-
ume flow through the pipe system.

If you know the inlet and outlet pressure but not the flow you can use an iterative
method to find the flow that settles in the system due to gravity and differences in
pressure at inlet and outlet.
1. Guess a volume flow through the system.
2. Calculate the pressure change in each section. The outlet pressure from each
section is the inlet pressure (= outlet pressure from previous section) plus the
change in pressure in the section. When you reach the outlet you have an out-
let pressure.
3. If the outlet pressure is higher than the expected pressure, increase the volume
flow and recalculate, if the outlet pressure is lower, reduce the volume flow
and recalculate. The pressure drop is a quadratic function of the velocity (vol-
ume flow). This means that if you calculate the pressure drop for three differ-
ent flows you can use a second order polynomial for interpola-
tion/extrapolation between pressure difference and volume flow.
4. When the calculated outlet pressure is the same as the expected outlet pressure
you have found the volume flow that will develop in the system.

3.3 Pipe system characteristic curve

You can calculate the total pressure loss in the pipe for different flows through the
pipe and draw a pipe characteristic curve of this. An increase in flow causes more

13
friction and larger pressure losses in the system and the curve shows how much the
pressure in the system must be increased (or decreased) in order to reach a certain
volume flow through the system. This characteristic curve can be used together with
similar curves for pumps and fans when choosing pumping equipment.
180

160

140
y = 0.0076x 2 + 0.0074x - 8.3504
120
total pressure loss (kPa)

100

80

60

40

20

-20
0 20 40 60 80 100 120 140 160
volume flow (m 3/h)

Figure 4. The characteristic curve for the pipe system above that shows the total pressure loss as
a function of the volume flow through the pipe system. A second order polynomial has been
drawn through the calculated points and the equation for this curve is given in the graph.

3.4 Pump types, centrifugal, displacement

Pumps are used in pipe systems to increase the pressure and the flow of liquids.
Common pumps are usually divided into two major types – centrifugal pumps and
displacement pumps. Centrifugal pumps accelerate the pumped fluid using open im-
pellers and then transform the kinetic energy (velocity) to pressure in diffusers. In
displacement pumps the fluid is enclosed into compartments and moved from the low
pressure side to the high pressure side.

3.5 Pump characteristic, work point, power, efficiency, recal-

culation models
A pump that works at a certain speed ( ω ) with a liquid of a certain density ( ρ ) can
increase the pressure of that liquid a certain amount ( ∆p ) when it pumps a certain
volume flow ( V ) of the liquid.
8

The flow in a displacement pump depends almost only on the pump speed regardless
of the increase in pressure. It does so since each volume of liquid is closed in a com-
partment and moved (displaced) from the low pressure side to the high pressure side.
The flow in a centrifugal pump depends on the increase in pressure. The higher the
pressure increase is the lower the volume flow. The liquid in the pump is never in a
closed volume since the pump is open from inlet to outlet. The liquid is accelerated
and the velocity is then transformed into pressure by changing the flow cross section
area.

14
A pump characteristic curve can be created for any pump where the increase in pres-
sure generated by the pump is plotted against the volume flow through the pump. The
curve is valid only when the pump pumps a certain liquid at a certain speed.
A similar power characteristic curve can also be created which shows the power the
pump needs at a certain volume flow (and pressure increase). Both types of curves are
shown in Figure 5.
Pressure increase Power

60 3

50 2.5
Pressure increase (kPa)

40 2

Power (kW)
30 1.5

20 1

10 0.5

0 0
0 2 4 6 8 10 12 14 16
Volume flow (m 3/h)

Figure 5. Pump and power characteristic curves for a centrifugal pump.

The characteristic curves are defined for a certain pump, at a certain speed and using a
certain liquid. To generate pump characteristics for each pump for any speed and any
liquid is not very practical. Instead the pump manufacturers provide characteristic
curves for pumps at a certain speed (like 1450 rpm) when they pump a certain liquid
(like water with a density of 1000 kg/m3). Using these characteristic curves it is then
possible estimate the pump characteristic curves for other pump speeds and liquids.
The coordinates in the pump characteristic curves can be recalculated to other liquid
densities and pump speeds using
2 2
d new
; ; ;

ω new ρ new
∆p new ⋅ ∆p original
<=> 9 <=> 9 <=> 9

≈ < 9

⋅ < 9

⋅ < 9

ω original d original ρ original

: : :

3 5
d new
A A A

ω new ρ new
Pnew ⋅ Poriginal
BCD ? BCD ? BCD ?

≈ B ?

⋅ B ?

⋅B ?

ω original d original ρ original

@ @ @

3
d new
H H

ω new
Vnew ⋅ Voriginal
IJK F IJK F

≈ I

⋅ F I F

ω original d original
E E

G G

Note that all three ( ∆p , P and V ) change at the same time! L

The capacity of a pump can also be changed by changing the impeller diameter d
(the moving part of a centrifugal pump is called an impeller). The equations above

15
show how a change in impeller diameter changes the characteristic curves of the
pump.
Only a part of the power delivered to the pump is used for pressure-volume work. The
remainder of the power becomes heat. The connection between power delivered and
power used for pressure-volume work is given by the pump efficiency, η , which is
defined as

V ⋅ ∆p
η=
P
M

The efficiency depends on the volume flow just as the pressure increase and the
power and a corresponding efficiency characteristic curve can be drawn.

3.6 Pump and pipe systems, flow control

The pump characteristic curve can be drawn in the same graph as the characteristic
curve for the pipe system in which the pump is installed. The point where the two
curves intersect gives the work point, i.e. the conditions that develop with this combi-
nation of pump and system, see Figure 6.
Pump Pipe system

100

80
Pressure increase (kPa)

60

40

20

0 W ork point

-20
0 2 4 6 8 10 12 14 16
Volume flow (m 3 /h)

Figure 6. The point where the pump characteristic curve and the pipe system curve cross each
other is the work point for this combination of system and pump.

The position of the work point (i.e. the flow and pressure in the system) can be
changed (controlled) either by changing something in the pipe system (opening and
closing of valves, changing pipe size) or by changing something regarding the pump
(pump model and size, pump type, pump speed).
Consider the pipe and pump system described in Figure 6. Decide how to accomplish
the flows given below and calculate new graphs for these cases.
a) You want to increase the flow to 12 m3/h.
b) You want to decrease the flow to 7 m3/h.

16
In order to increase the flow through the system by changing the pipe system you
would have to increase the pipe diameter and perhaps change or remove the valve.
Recalculate the pressure loss at different volume flows in the pipe system using the
new pipe diameters and/or valve resistances. Find the work points and check if the
diameter or valve resistance must be increased or decreased.
Pump Valve removed Pipe system

100

80
Pressure increase (kPa)

60

40

20

0
New work point

-20
0 2 4 6 8 10 12 14 16
Volume flow (m 3/h)

In order to change the flow through the system by changing the pump you might
change the pump to a bigger or smaller one or change the speed of the existing pump.
Data for other pumps can be found from data given by the pump manufacturers.
If the speed is changed then the coordinates in the pump characteristic curve (volume
flow and pressure increase) are recalculated for the new speed according to the equa-
tions given earlier. Find the work point for the new speed and check if the speed needs
increasing or decreasing. When the correct speed is found, recalculate the power
curve to find the correct power needed by the pump at the new work point.

Pump Speed control - decrease Speed control - increase Pipe system

100

80
Pressure increase (kPa)

60

40

20

-20
0 2 4 6 8 10 12 14 16
Volume flow (m 3/h)

In order to decrease the flow by changing the pipe system you would simply close the
valve until the work point is at 7 m3/h.

17
 Pump Valve control Pipe system

100

80

Pressure increase (kPa)

60

40

20

New work point

0

-20
0 2 4 6 8 10 12 14 16
Volume flow (m 3/h)

3.6.1 Example pipe system

The drain pipe shown in Figure 7 should be able to handle a water flow of at least
20 m3/h without using a pump. The water level in the open inlet vessel must not be
higher than 2 m from the bottom of the vessel to prevent spillage. A valve is placed at
the outlet of the pipe.
How large should the pipe diameter be (see standard pipe sizes in Figure 8)?
Where is the pressure highest in the pipe and how high is the pressure at that
point (when the valve is open and when valve is closed)?
The pipe is made of steel and the pipe bends are elbows. The valve in the pipe is a ball
valve and it has about the same resistance as the pipe when it is fully open.

Figure 7. Drain pipe system.

18
Figure 8. Standard steel pipe sizes suitable for pressures up to 4 bar (extract from Finnish stan-
dard SFS 5563).

3.6.2 Example pipe and pump system

Biodiesel will be moved from a short term storage tank at the production plant
through a pipeline to a larger storage tank.

19
The pipeline is about 560 m long and made of steel pipe of size DN 125 with an inner
diameter of 135.7 mm. The ground below the pipe is flat all the way. There are three
U-turns in the pipe to remove thermal stress in the pipe. Each U-turn is made of 4 el-
bows of standard size (radius 190 mm). All other turns are also standard elbows ex-
cept one which is a T-piece. The flow goes only in one direction from the T-piece.
The valves are all ball valves which have the same inner diameter as the pipe when
fully open.
The short term storage must be possible to empty in half an hour (which actually gives
the pipe size since a suitable velocity for liquids is about 1 m/s). Determine the neces-
sary pressure increase in the pump for this volume flow. Select a suitable pump for
this from the attached pump diagram. The biodiesel has a viscosity of 4.5 mm2/s (wa-
ter 1.3 mm2/s) and a density of 900 kg/m3 (water 1000 kg/m3).

20
3.7 Gas transport with fans in ducts
Fans are used in duct systems to increase the pressure (and flow) of gases that pass
through the ducts. Fans are used when the pressure increase (and the corresponding
change in density) is low. Compressors are used when large pressure changes must be
accomplished.
The small change in density (or specific volume) in fans means that the same equa-
tions can be used as for pumps.

3.8 Gas transport with compressors

Compressors are also used for transport of gases. Compressors are used when large
pressure changes must be accomplished. Large changes in pressure also cause large
changes in density in the gas and the compression causes quite large increases in tem-
perature in the gases.

3.9 Compressor types, centrifugal, displacement

Two main types of compressors are used. Centrifugal compressors are much like cen-
trifugal pumps although they often are made up of many successive compression
stages with cooling between the steps. Displacement compressors are available in
many different designs but they have that in common that volumes of gas are enclosed
in a closed space and then moved from the low pressure to the high pressure side of
the compressor.

3.10 Compressor calculations

Figure 9. Flow diagram of a compressor with inlet and outlet as well as power and heat flows into
and out of the balance boundary.
The energy flow balance for the balance boundary around the compressor indicated in
Figure 9 is, when it is considered to be at steady state

21
 m ⋅ hin + P = m ⋅ hout + Q
N N
N

where m is the mass flow of gas through the compressor, hin and hout are the specific
O

enthalpies of the gas at the inlet and the outlet of the compressor, P is the mechanical
power supplied to the compressor and Q is the heat flow from the compressor.
P

If the balance above is solved for the required mechanical power you get

P = m ⋅ (hout − hin ) + Q
N
N

The enthalpy is defined by the pressure and temperature of the gas. Commonly the
inlet temperature and pressure are known as well as the expected outlet pressure. The
outlet temperature depends on the heat flow but also on the efficiency of the compres-
sor.
The efficiency is usually defined in relation to an ideal situation where no entropy is
generated. The entropy balance for the same balance is

Q
Q

m ⋅ sin + S prod = m ⋅ s out +

Q

Q Q

where sin and s out are the specific entropies of the gas at the inlet and the outlet,
S prod is the entropy production in the compressor and T temperature at the boundary
R

where the heat flow passes through.

If the heat flow is zero (i.e. the compressor is well insulated), then the compression is
considered to be adiabatic. This is not always true for real compressors but it is a
good approximation.
The ideal situation mentioned earlier that will be used for comparison is that when
there is no entropy production in the compressor (isentropic compression) and when
the compressor is considered adiabatic. The necessary mechanical power to the com-
pressor would then be

P = m ⋅ (hout
′ − hin )
S

and the entropy balance could be simplified into

sin = s out
′

where hout
′ is the specific enthalpy and s out
′ is the specific entropy of the gas leaving
the compressor in this ideal case.
If the inlet pressure pin and temperature Tin are known then the inlet specific enthalpy
hin and specific entropy sin are both defined.
The outlet specific entropy s out
′ is, according to the equation above, the same as the
inlet specific entropy. If the outlet pressure pout is known or set then the temperature

22
Tout
′ corresponding to this specific entropy can be calculated and using that also the
specific enthalpy hout
′ .
The adiabatic conditions can, as was said before, quite well approximate the condi-
tions in a compressor, but the isentropic conditions cannot be achieved. The actual
conditions are compared to the ideal case using an adiabatic efficiency.
The adiabatic efficiency of a compressor is defined as

hout
′ − hin
η ad =
hout − hin

If the heat capacity is considered constant and the gas ideal then it is relatively simple
to calculate changes in specific enthalpy and entropy. A change in specific enthalpy
between two states 1 and 2 is modeled using

h2 − h1 ≈ c p ⋅ (T2 − T1 )

and the corresponding change in specific entropy with

T R p
V V

s 2 − s1 ≈ c p ⋅ ln 2 − ⋅ ln 2
WXY T WXY T

W T W T

T1 M p1
U U

where c p is the specific heat capacity of the gas at constant pressure, R is the gas
constant (8.314 kJ/kmolK) and M is the molar mass of the gas. Observe that the tem-
peratures are given in K.
The temperature of the gas leaving the compressor in the ideal isentropic case can be
calculated directly from the entropy balance

T′ R p
V V

s out
′ − s in = c p ⋅ ln out − ⋅ ln out = 0
WXY T WXY T

W T W T

Tin M pin
U U

which gives
R
p out
\

c p ⋅M
Tout
′ = Tin ⋅
]^_ Z

] Z

pin
[

The temperature of the gas leaving the compressor in the real case is calculated using
the definition of the adiabatic compressor efficiency

′ − hin c p ⋅ (Tout
hout ′ − Tin )
η ad = ≈
hout − hin c p ⋅ (Tout − Tin )

which gives

23
 Tout
′ − Tin
Tout = Tin +
η ad

The total power required for the compression then becomes

P = m ⋅ c p ⋅ (Tout − Tin )
`

3.10.1 Example – compressors

Natural gas is transported, often long distances, through pipelines from the production
fields to the consumers. The natural gas is compressed using compressors to decrease
the volume of the gas and the pipe size.
Compressor stations are placed along the pipeline to keep up the necessary pressure.
At one of these stations the pressure of the natural gas must be increased from 15 bar
to 50 bar. The flow through the pipeline is about 36000 Nm3/h.
Assuming that the gas temperature is 5 °C before the compressor station and that the
compressor has an adiabatic efficiency of 70 %, how high will the temperature be af-
ter the compressor?
How much cooling must be provided to decrease the temperature to 40 °C?
The higher heating value of natural gas is about 54.4 MJ/kg. If natural gas is used to
power the compressor with an efficiency of about 35 % how much gas would be used
(percentage of flow in to the compressor station)?
Natural gas is mainly composed of methane (CH4) which has a specific heat capacity
of about 2.48 kJ/kgK.

3.11 Typical flow velocities

The velocity of the flow in pipes or ducts depends on the pipe or duct size you choose
for a certain flow. The velocity in larger pipes is smaller causing a smaller pressure
drop in the pipes. Larger pipes also cost more to buy and install so you have to find a
pipe size that minimizes the overall cost. It is possible to optimize the selection of
pipe size in order to minimize cost but this is generally not done. Most often you just
see to it that the flow velocity in the pipe is within a sensible range. The following
flow velocities can be used for gas and liquid flows in ducts and pipes.
Gas flow – 10-30 m/s
Liquid flow – 1-2 m/s.

24
4. Heat exchanger systems
Whenever you want to move energy in the form of heat from a warmer stream to a
colder stream without actually mixing the streams you need a heat exchanger of some
kind. (Note that if you need to move heat energy from a colder stream to a warmer
stream you need a heat pump instead of a heat exchanger. A refrigerator is a heat
pump).
Heat exchangers can be used to set the correct reaction temperature in a reactor, to
change the transport properties of fluids, to stop the reactions or to cool the products,
for heat recovery after the reactor or to evaporate liquids or condense gases.
There are usually many places where cooling or heating is needed in a process system
and it is often possible, or even necessary, to interconnect these heat flows instead of
using external heating or cooling. Heat recovery can give economic benefits.
The interconnection of heating and cooling using heat exchangers can be designed
and/or analyzed using heat exchanger network models.

4.1 Heat exchanger types

The basic idea of a heat exchanger is simple. A thin wall separates the warmer and the
colder stream from each other and at the same time leads heat from the warmer to the
colder stream. Heat exchangers are commonly used for transport of heat between two
gas flows, between gas and liquid flows and between two liquids.
The construction methods and applications of heat exchangers divides them into many
different types such as shell and tube heat exchangers, plate heat exchangers, jacketed
rectors, steam pipes in furnaces, spiral heat exchangers, condensers and reboilers in
distillations columns and so on.

4.2 Heat exchanger modeling, counter current, co-current,

real world
Heat exchangers are often modeled using two basic setups – co-current and counter
current. Both configurations are shown in Figure 10.

Figure 10. Co-current (left) and counter current (right) heat exchangers.

In a co-current heat exchanger the hot and cold flows go through the heat exchanger
in the same direction while they go in opposite directions in a counter current heat ex-
changer.
A basic model used to describe the heat flow through the wall of a heat exchanger of
any type is

25
 Q = k ⋅ A ⋅ ∆θ ln
a

where k is the overall heat transfer coefficient and A is the area of the heat transfer
surface. The logarithmic mean temperature, ∆θ ln , across the heat transfer surface is
defined as

∆θ1 − ∆θ 2
∆θ ln =
∆θ
ln 1
∆θ 2

where ∆θ1 is the temperature difference between the warm flow and the cold flow at
one end and ∆θ 2 is the difference at the other end.
Real world heat exchangers work in a non-ideal way; they might be partially co- and
counter current, they might be cross current – but never ideal. One way of compensat-
ing is to multiply the log mean temperature with a correction factor F .

Q = k ⋅ A ⋅ F ⋅ ∆θ ln
b

The correction factor depends on the actual type of heat exchanger as well as the tem-
peratures of the flows into and out of the heat exchanger.

Figure 11. Correction factors F for two types of heat exchangers (from Perry’s Chemical Engi-
neers Handbook). R = (T1 − T2 ) (t 2 − t1 ) , S = (t 2 − t1 ) (T1 − t1 )

26
4.3 Overall heat transfer coefficient
The heat flow resistance of the heat exchanger wall is more or less constant regardless
of temperature and flow. The overall heat flow resistance, on the other hand, is not
constant. It changes when deposits are formed on the surface and it changes when the
flow velocity and/or the temperature changes.
The overall heat transfer coefficient is a combination of several parts and can be mod-
eled as

1 1 s fa s s fb 1
= + + + +
k α a λ fa λ λ fb α b

where α a and α b are the convective heat transfer coefficients between liquid and
solid, s fa and s fb are the thicknesses of surface deposits on each side of the heat
transfer surface, s the heat transfer wall thickness and λ fa , λ fb and λ are the corre-
sponding heat conductivities.

4.4 Condensing steam, evaporation

The flows being heated or cooled might change phase in the heat exchanger. Liquids
might boil and become gases while gases might condense and become liquids. The
change in phase also changes the properties of the flow (like viscosity, density, spe-
cific heat capacity) and therefore also how the heat exchanger works.
The temperature on the condensing/evaporating side depends on the pressure since the
condensation/evaporation takes place at the saturation temperature. This is also true
when superheated or subcooled fluids are used. Part of the heat exchanger is used for
cooling/heating the flow to the saturation temperature followed by condensa-
tion/evaporation. This makes it necessary to divide the heat exchanger calculations
into to distinct parts.

4.5 Heat exchanger networks, pinch analysis

There can be many places where heating and cooling is needed in process systems and
the whole heat exchanger system can then form a heat exchanger network. During
design and analysis it can be difficult to decide where recovered heat should be used.
One way to analyze the system is to subject the heat flows to pinch analysis. In pinch
analysis you not only account for the available heat but also for the temperature of the
flows. The result is a graph (see Figure 12) that shows which flows can be connected
with heat exchangers and how much external heat or cooling is needed to make it pos-
sible.

27
 140

120

100

C)
o
80

temperature (
heating
cooling
60

40

20

0
0 1000 2000 3000 4000 5000 6000
heat transfer (kW )

Figure 12. Heating and cooling curves used for pinch analysis. The pinch is where the tempera-
ture difference is smallest. If the curves cross each other then additional heating or cooling is
needed in the system.

4.5.1 Example
A compressor produces 4 kg/min compressed air of 10 bar using 25 °C air. The com-
pressor is a two stage compressor with an intercooler and an aftercooler that cool the
air to 25 °C after each compression step. The air temperature after each compression
step is 169 °C.
The compressor also has a separate oil cooler which cools the lubrication oil in the
compressor. When the oil flow is 4.2 kg/min and 50 °C when it enters the compressor
then it will be 90 °C when it is returned to the cooler. The specific heat capacity of the
oil is about 2 kJ/kg°C.
Cold water (4 °C) has been used for cooling in all three coolers. The flow is set so that
the water temperature out is no higher than 15 °C.
The heat from the coolers could be used to heat the air that comes into the building at
-15 °C to 25 °C. The necessary air flow is about 15 kg/s. Another possible use for the
heat is for making hot water. The water comes in at 4 °C and should be heated to
70 °C. The hot water usage is about 8 kg/min.
So there are three hot flows that have to be cooled: air in intercooler, air in aftercooler
and lubrication oil in the oil cooler. The heat in these flows could possibly be used to
heat the two cold flows: house air, and hot water.
It is of course also possible to use additional heating from the district heating system
and additional cooling with cold water using the existing systems.
Calculate the need for additional heating and/or cooling. Draw a graph with heating
and cooling curves to see if it is possible to realize this heat recovery using a heat ex-
changer network. Data for the hot and cold flows are given in the following table:

28
Id description flow heat capacity inlet outlet
H1 air in intercooler 4 kg/min 1.0 kJ/kgK 169 °C 25 °C
H2 air in aftercooler 4 kg/min 1.0 kJ/kgK 169 °C 25 °C
H3 oil in oil cooler 4.2 kg/min 2.0 kJ/kgK 90 °C 50 °C
H district heating x kg/min 4.2 kJ/kgK 120 °C 50 °C
C1 house air 15 kg/min 1.0 kJ/kgK -15 °C 25 °C
C2 hot tap water 8 kg/min 4.2 kJ/kgK 4 °C 70 °C
C cold tap water x kg/min 4.2 kJ/kgK 4 °C 15 °C

29
5. Economic analysis – modeling and selection
One important part of process design is to make choices. Many of these choices are
made based on economic evaluation of the design or designs that we generate.
When the design has reached the stage, where the major pieces of equipment have
been sized (heat transfer surface, pump capacity, mixer power, reactor size etc.) then
can a preliminary estimate of the cost for the plant be made.
Sinnott (Coulson&Richardson’s Chemical Engineering Volume 6 – Chemical Engi-
neering and Design, chapter 6) shows some examples on how estimates of this type
can be made.

30