Professional Documents
Culture Documents
Chemical Engineering in Practice Second Edition - Sampler
Chemical Engineering in Practice Second Edition - Sampler
IN PRACTICE
Design, Simulation and Implementation
SAMPLER
John E. Edwards
Process Simulation Engineer, P & I Design Ltd
Published by
P & I Design Ltd
2 Reed Street, Thornaby TS17 7AF
www.pidesign.co.uk
Copyright J.E.Edwards 2011
jee@pidesign.co.uk
Printed by Billingham Press Ltd, Billingham TS23 1LF
ISBN 978-0-9568169-1-7
Contents
Section 1
Introduction
Section 2
Fundamentals
Simulation Cases
9
16
Section 3
17
36
Section 4
Thermodynamics
Simulation Cases
53
93
Section 5
95
149
Section 6
165
199
Section 7
227
259
Section 8
273
Section 9
Continuous Distillation
327
Section 10
Batch Distillation
Simulation Cases
351
361
Section 11
Batch Reactors
Simulation Cases
377
403
Section 12
421
464
Section 13
Hydrocarbon Processing
465
Section 14
485
Section 15
Emerging Markets
Simulation Cases
509
511
Section 16
Utility Systems
521
Section 17
543
Preface
This book is based on 50 years experience gained whilst working in the process engineering and
simulation fields. It is not suitable for rigorous academic study, as it is intended for the practicing
engineer. However, certain cases could provide a more practical basis for student process design
projects. There is nothing original here, as the information has been developed from experience
and based on many classic texts in Chemical Engineering literature, not least the following:
This book provides a synopsis of each section topic coupled with a concise data set, gleaned from
many sources. Applications are presented from real situations whether it be design, simulation or
operations.
Any originality may be in the presentation format and as such some would consider this to be a
Handbook or Workbook. Many modern text books in Chemical Engineering do not adopt this
approach which is not very helpful to the engineer challenged with providing a solution in a timely
manner. The text books noted above seem to have a clarity of presentation missing from many
modern text books.
This book has been collated from individual technical papers and guidance notes that have been
written in support of our engineering activities and simulation work. As such, each topic is in the
form of a condensed refresher and provides useful practical information and data.
Each section is numbered uniquely for contents and references, with the appendices and
nomenclature being section specific. The references are not a comprehensive list and apologies for
any unintended omissions.
Process simulation examples are carried out using CHEMCAD simulation software by
Chemstations, Inc. of Houston. www.chemstations.com
Examples are included for each section topic and in many cases are based on actual designs or
situations. Units used are not consistent, between Sections, as selection depended on the preferred
units for a particular study. This should not be an issue as all simulators have powerful unit
conversion facilities.
The simulation cases are available in electronic format. To run the simulations you will need a
CHEMCAD software licence.
To obtain a software license and the simulation cases please contact support@pidesign.co.uk who
will organise a 30 day licence with Chemstations.
2nd Edition
This edition has reorganised and rationalised most sections, including Process Instrumentation and
Control, Fluid Flow and Piping Networks, Emergency Relief Systems and Batch Distillation. A new
Utility System Section has been added and includes cases on CHP Steam Systems, Refrigeration
and Cooling Towers. Several new cases have been added to demonstrate dynamic simulation
capability; in particular Fire Induced H2O2 Thermal Decomposition and Bio-fuel Blending. Additional
diagrams and engineering data have been added where appropriate.
Section 1
Introduction
The process industry covers a broad spectrum of activities that involve the handling and treatment
of gases, liquids and solids over a wide range of physical and processing conditions. The scope of
this variety is shown in the following chart.
The development of a process and the steps taken to achieve an operational full scale unit are
complex and involve many disciplines to achieve a successful result. Typical activities are shown in
the following chart.
Project implementation involves many disciplines, including process, mechanical, electrical and
control engineers, working as a team in a coordinated manner. The main work phases, together
with the documentation required, are summarised in the following charts.
The process engineer, working in this environment, is required to provide safe, practical and timely
solutions to design problems without having the opportunity to study the topic in great depth. This
book is an attempt to provide a comprehensive review of the fundamentals, definitions and design
principles to allow the solution of cases encountered in normal practice. For more complex cases or
when in doubt reference should be made to an appropriate authority or specialist in the field.
The design process should attempt to identify potential problems and errors at the earliest stage
possible in the project life cycle. Many techniques have been developed to achieve this, namely
Hazard Identification (HAZID), Hazard and Operability (HAZOP), Layer of Protection Analysis
(LOPA) together with process simulation. The project cost impact curves show the benefits of stress
testing the designs using these techniques to avoid the increased costs of correction, even if
economically possible, later in the project life cycle.
COST IMPACT
1
0.8
0.6
0.4
0.2
x
0
0
0.2
0.4
0.6
0.8
PROJECT TIME
x Problem Discovery
Design Decisions
Processes can be batch, semi-continuous or continuous and the selection depends on many
factors.
Batch processes are used in the manufacture of a wide variety of fine and speciality chemicals and
involve the following general criteria:
Batch processes are inherently transient in nature and the equipment design needs to be analysed
using the physical properties of the process and service fluids under a wide range of operating
conditions.
Continuous processes are used in the oil and gas processing and petrochemicals industries and
have the following characteristics:
To achieve consistent quality and to operate the process optimally maximising productivity and
minimising raw material and energy costs requires effective control of the key process parameters.
This requires process control and data monitoring systems that matches the process to which they
are being applied.
A successful outcome does not necessarily require an in depth mathematical study of the control
system and its interaction with the process, which is how courses on instrumentation and control
are frequently presented. Very often having a good understanding for the process behaviour and
the application of common sense can provide satisfactory results.
In 1967 Shinskey presented a novel approach to the understanding of process characteristics and
the associated control systems by studying the problem in the time domain and the application of
basic mass and energy balances.
Each process parameter exhibits unique properties that determines the control strategy to be
applied.
Flow and liquid pressure periods are fast and noisy with no dead time, requiring a low gain
response with reset adjustment to maintain a desired value.
Gas pressure exhibits a high capacity with no dead time, requiring a high gain response
which is adequate to maintain a desired value. This is considered to be the easiest
parameter to control.
Liquid level period can be fast or slow and sometimes noisy (boiling) with capacity,
requiring medium gain and sometimes with reset adjustment to maintain a desired value.
Temperature period can be medium or slow with capacity and dead time, requiring variable
gain, reset adjustment and anticipatory response.
The control system should not be used to overcome shortcomings in plant design. To avoid this the
process plant design and the control philosophy should be considered throughout the design
process. For example, if a processing unit takes a long time to achieve stable and optimal operating
conditions the feeds should be held constant by the use of surge tanks in which the level is allowed
to float between acceptable limits.
Steady state and dynamic process simulation proves the capability to achieve stable and
reproducible operating conditions with acceptable product purity, yield and cycle times to satisfy the
commercial requirements and the safety and environmental issues for the regulatory authorities.
The characteristic properties and design parameters of the process are analysed in this book using
Section 2
Fundamentals
Contents
1.0
2.0
3.0
4.0
Introduction
Units
Mass and Energy Balances
Simulation Principles
Appendices
I
Nomenclature
References
1
2
nd
1.0
Introduction
This Section reviews engineering units and their application to prevent misuse in certain
applications. Mass and energy balance principles and the fundamentals of steady state and
dynamic simulation are also reviewed.
2.0
Units
(1)
Molality, mi (mol/kg) denotes the total number of moles of solute species i per 1000g of solvent (not
solution).
Molarity, Mi (mol/litre, kmol/m3) denotes the total number of moles of species i per litre of solution,
sometimes referred to as Normality N which is essentially the same as mol/litre.
Mole fraction, xi denotes the number of moles of species i divided by the total number of moles in a
solution.
xi =
mi MWS
1000
Mole ratio Xi denotes the number of moles of species i divided by the number of moles of species j.
For a binary solution we have:
Xi =
xi = xi
x j 1 xi
wi =
M i MW i
So ln1000
Consider a reactant species A with a flowrate WA (kg/h) and molecular weight MA (kg/kmol) and
inert species B with a flowrate WB (kg/h) and molecular weight MB (kg/kmol), we have:
Mole fraction component A
WA / M A
WA / M A + WB /M B
yA =
yB = 1 yA
M AB =
WA + WB
WA / M A + WB /M B
3
For a system pressure P=1.01325 bar and temperature T K the inlet gas density (g/litre or kg/m ) is
given by:
1 =
M AB 273
22.4 273 + T
The gas inlet concentration of component A can be converted from mole fraction to mg/m3 using:
M 273
y A 10 6 A
22.4 273 x T
Weight fraction is converted to mole fraction as follows:
xi =
w i MW i
w i MW i + 1 w i 18
10
ppm v =
Consider
mg
m
at 20C
ppmv = mf x 106
Pollutants in liquids
Consider for water, where L is a litre and is liquid density:
ppm w =
mg
L x
mg
L
ppmw = wf x106
wf =
ppmw
mf =
10
ppm w
6
10
MWSolvent
MW Conta min ant
For example, trichloroethylene (MW 131.4 g/mol) has a consent limit for drinking water, using
preferred units, of 0.005 mg/L = 0.005 ppm(w).Therefore the consent limit as a weight fraction is
-6
-6
0.005x10 and as a mole fraction is 0.000684x10 .
-5
For water contaminated with 100 ppm(w) trichloroethylene the inlet mole fraction is 1.36 x 10 and
-7
for a target purity of 1 ppm the outlet mole fraction is 1.36 x 10 . We will be looking at this air
(13)
stripping example later in Section 8.5.
The following table shows the relationships between volumes, moles, and mass.
MW
kg/kmo
l
MA
MB
Total
MC
%
v/v
Volume
3
in 1 m
3
3
m /m
PA
VA =
PA
100
PB
VB =
PB
100
PB MB
PC
VC =
PC
100
PC MC
10
0
Mass in
3
1m
3
kg/m
kg/h
Molar
Flow
kmol/h
Mass Flow
% w/w
PAMA
100 P A M A
W PA M A
W PA
2241.5
Pi M i
Pi Mi
Pi M i
100 PB M B
W PB MB
W PB
Pi Mi
Pi Mi
Pi M i
100 PC MC
W PC MC
W PC
Pi M i
Pi M i
Pi Mi
100
2241.5
2241.5
Pi Mi
1
2241.5
11
%
mo
l
PA
PB
PC
100 W
Pi Mi
100
12
We have defined: 1 g-mole of any gas occupies 22.415 litre(dm ) at 0C and 1 atmosphere.
Therefore we can say that the same g-mole of any gas will occupy the same volume giving:
Mole % = Volume %
__
M
3
For a Total Flow of W (kg/h) and a mixture density of G 0 =
(kg/m ) we have:
2241.5
Volumetric Flow
W = 2241.5W
Pi M i
G 0
__
Where average MW
M=
Pi M i
100
(Nm /h)
(kg/kmol)
To correct for temperature and pressure gas density calculations are calculated from:
G =
M W Pf 273 kg
3
/ m
22.415 Zf Tf
3
y CO 2 =
MEA
PT
H 2O
p CO 2
PT
rd
Consider a system at 1.01325 bar (101.325 kPa) with components CO2/MEA/H2O with H2O as 3
component having a constant mole fraction of 0.866.
A single set of data from the TPxy plot, at constant temperature of 50C, gives:
pMEA + pH2O = 0.1242 bar when x=y=0, ie no CO2 in system.
CO2/MEA/H2O
CO2/MEA
xCO2 = 0.0335
xCO2 = 0.25
The equilibrium curve liquid composition axis can have units of mole fraction 0.0335 or mole ratio
CO2/MEA 0.25/(1-0.25) = 0.333. The vapour composition axis can have units of partial pressure
pCO2 = (0.126 0.1242)x101.325 = 0.18 kPa or mole fraction 0.18 / 101.325 = 0.00178.
Note TPxy data gives CO2 mole fractions in vapour as mol CO2/(mol CO2 + mol MEA + mol H2O)
or mol CO2/ (mol CO2 + mol MEA) which can be misused in absorption of species from inerts.
13
3.0
(1, 2)
The law of conservation of mass, which states that mass can neither be created or destroyed,
enforces the requirement that a process design is not complete until all inputs and outputs of the
overall process and individual units satisfy a mass balance for each component.
For a batch process no mass crosses the system boundaries between the time the feed is charged
and the product is removed and for a continuous process inputs and outputs flow continuously. A
semi-batch process combines both characteristics.
Consider the mass balance across a Distillation Unit. At steady state the overall material balance
must be satisfied for each component. Only one product stream can be set independently making
the other product stream a dependent variable. In general, the distillate flow is manipulated, directly
or through varying reflux ratio, requiring the bottom outlet flow to be adjusted to maintain the
material balance which is achieved by holding the reboiler liquid level constant.
The material balance, based on the more volatile component (MVC), yields:
D z x
=
F y x
This shows that the flow ratio D/F determines the relative composition of each product, such that if
feed composition z changes then D/F must be manipulated to maintain constant x and y.
Note how the mass balance equations imposes restraints on the inter-dependency of feed and
product stream flow rates and compositions. This feature is used to develop control strategies such
as feedforward control.
The law of conservation of energy, which states that energy can neither be created or destroyed, is
the First Law of Thermodynamics, discussed later in Section 3. The rate at which energy
(kinetic+potential +internal) accumulates in a system is the net energy transferred (heat and work)
to the system from the input streams minus the output streams. A batch process is a closed system,
in that energy does not cross the system boundaries and a continuous process is an open system.
For the distillation column a simplified overall energy balance, where subscript H refers to stream
enthalpy, yields:
Qr + F H = Qc + DH + B H + losses
The process design objective is always to minimize the use of energy consistent with achieving the
desired product quantity and quality. Consider a batch distillation where the objective is to minimise
the batch cycle time which implies optimising the heat input and minimising the reflux ratio to
achieve the required separation. The process variables are interdependent as determined by the
mass and energy balances and the mode of operation.
14
The heat input, for a variable top composition achieved by setting a fixed reflux ratio, is given by:
Q = (SO S1)(R + 1
The heat input, for a fixed top composition is achieved by varying the reflux ratio to maintain a fixed
top temperature at constant pressure, is given by:
Q = (SO S1) f (R )d R
R1
RO
Where
S0
S1
R
Both relationships indicate that the reflux ratio must be kept to a minimum, subject to satisfying the
requirements for the desired separation specification, to optimise the heat input.
Now consider an absorber mass balance, where the gas inlet stream condition and the liquid inlet
composition are determined by the upstream process. The liquid flow rate is the only variable to
adjust for the dependent outlet streams compositions.
The gas stream, G (mol/area-time), at any point in the tower consists of the diffusing species, mole
fraction y, partial pressure p or mole ratio Y and the non-diffusing relatively insoluble gas GS and
are related as follows.
y
p
Y = 1 y =
Pt p
G
G s = G (1 y ) = 1 + Y
The liquid stream L (mol/area-time), at any point in the tower consists of the dissolved species,
mole fraction x or mole ratio X and essentially non-volatile solvent LS and are related as follows:
x
X = 1 x
L
Ls = L (1 x ) = 1 + X
Since the inert gas GS and solvent liquid LS remain essentially unchanged in the tower the overall
material balance, based on them, yields:
4.0
Simulation Principles
Simulators provide process engineers with a powerful tool to assist in the design, optimisation and
operation of processes. The simulator characteristics are determined by the process. Batch and
semi-batch processes require dynamic simulation whereas continuous processes are studied with
both dynamic and steady state simulators.
Dynamic simulators are based on real time or accelerated dynamic simulation principles and can
provide a true representation of the plant behaviour, including retention times and thermal lags.
Continuous steady state with varying throughput, start up, shutdown, feed change, composition
change can all be studied. Dynamic simulators can also be used for operational plant
troubleshooting, control loop tuning and real time optimisation. Dynamic simulations require more
information and are more difficult to set up.
Steady state simulators model plant behaviour using continuous, time interval based, steady state
approximations and require the least effort to set up and calibrate the flowsheet. Dynamic
simulation can also considered to be a series of steady state simulations.
A process simulation involves taking the input stream flow rates, compositions and thermodynamic
conditions, performing a series of iterative calculations as the streams are processed through Unit
Operations and recycles, finally leading to the output stream flow rates, compositions and
thermodynamic conditions.
Products
Feeds
Recycles
Prior to the introduction of process simulators, this activity involved a team of engineers working
many man hours and maybe not achieving a definitive result.
16
Appendix I
Nomenclature
G
GS
L
LS
F
D
B
S
x
y
z
mol/s-m
2
mol/s-m
2
mol/s-m
2
mol/s-m
kg/h
kg/h
kg/h
kg
mole fraction
mole fraction
mole fraction
Subscripts
0
1
2
H
initial condition
tower bottom, gas inlet, liquid outlet or final condition
tower top, gas outlet, liquid inlet
enthalpy
Description
Tutorial A1
This tutorial introduces UnitOps Heat Exchangers, Pipe Simulator, Flash Unit, Control Valve and
SCDS Column and the handling of a recycle stream. The objective is to achieve a column bottoms
propane concentration target of 0.01 mole fraction.
17
Section 3
Process Instrumentation and Control
Contents
1.0
2.0
3.0
4.0
4.1
4.2
5.0
6.0
7.0
7.1
7.2
Introduction
Control System Basics
Controller Basics
Control Valves
Control Valve Characteristics
Control Valve Calibration
Controller Tuning
Dynamic Simulation Considerations
Measurement Sensors
Flow Measurement
Level Measurement
Appendices
I
II
III
1.
2.
3
4
5
Simulation Cases
In this section cases are presented to provide an understanding of CHEMCAD dynamic simulation
and its capabilities for sizing orifice plates and control valves and for the study of control systems.
The reference Process Control Systems by F.G.Shinskey Chapters 1 and 2
the cases presented.
(1)
For practice you can build the models or use the models provided as downloads from electronic
media available. It is strongly recommended that you work with a copy of the jobs.
Case/File Name
3.01
3.02
3.03
3.04
3.05
3.06
3.07
3.08
3.09
3.10
3.11
Description
Control of Dead Time
Control of Capacity - Tank Level
Two Capacity Tank Displacer Level
Tank Balancing
Tank Level Controller Tuning
Gas Pressure Control
Flow Control with Level Cut-back
Level Control Training Simulator
Valve sizing - liquid
Valve sizing compressible fluid
Orifice sizing liquids, gases and steam
18
1.0
Introduction
A basic control loop consists of an instrument measuring a process variable and converting it into a
4 to 20 mA dc signal for transmitting to a controller. The controller produces a 4 to 20 mA dc output
signal, based on the error between measurement and set point and the control settings
Proportional, Integral and Derivative(PID). The output signal is connected to a regulating device,
such as a control valve or variable speed drive. A typical single loop is shown:
:
Transmitter
Control Valve
TT
CV
4-20 mA
4-20 mA
XIC
The key points to consider when setting up a control loop with a controller and control valve are:
1.
2.
3.
The controller set point value is the desired process value in the same units as specified
for the measurement.
4.
Controller Action determines the controller output (OP) in response to the error between
Set Point (SP) Measured Variable (MV). We desire feedback control which behaves as
follows:
For Reverse Action:
For Direct Action:
It is essential that this concept is understood because an incorrect selection will result in
Positive Feed back with MV continually increasing or decreasing.
5.
Control Valve failure action is determined by the process e.g. a heating valve usually fails
closed and a cooling valve usually fails open. A fail closed valve requires an increasing
controller output 4 to 20 mA signal to open (Signal to Open) and a fail open valve requires
an increasing controller output to close (Signal to Close).
6.
An incorrect match of controller action and valve action will result in Positive Feedback.
7.
Tuning parameters for noisy control loops e.g. liquid flow, liquid pressure and level use a
wide Proportional Band (PB in %) as controller gain is 100/PB. In this case, short Integral
Action Times (Ti) are used to eliminate SPMV error. Derivative is never used on noisy
signals as Td operates on rate of change of error de/dt where rapid changes in error will
result in large rapid changes in controller output leading to control valve slamming and
damage.
19
3.0
Controller Basics
The controller algorithm in CHEMCAD is representative of a model which does not exhibit
interaction between control modes:
Controller output =
Where,
P
Ti
Td
Po
e
=
=
=
=
=
de
100
1
e + edt + Td
+ Po
P
Ti
dt
proportional band, %
integral action time, min
derivative action time, min
steady state controller output
(Xset X), the error between process measurement and set point
(Heating valve with fail closed valve, cooling valve with fail open valve)
In CHEMCAD the controller output can be directed to one valve or two valves, in split range
operation, or another controller set point for cascade operation.
The measurement can be derived from a Stream or UnitOp variable. For UnitOp variable
measurements always use the Calculated Value. Set the steady state controller output (Po) at the
simulation start value, which is derived from running the model in steady state with correctly sized
control valves, using valve sizing feature.
20
The basic control loop consists of a field mounted transmitter, with the process variable (PV) range
scaled to represent a 4-20 mA signal, a control room mounted controller and process parameter
regulating device such as a control valve or variable speed drive. Common control loop
configurations are shown below:
Single Loop
Tx
Cascade Loop
Tx
Tx
Ratio Loop
Tx
Primary
Tx
Tx
XIC
XIC
Ratio
RR
XIC
XIC
XIC
Hi Lo Select
Secondary
CV
Single Valve
CV
XIC
XIC
XIC
CV
CV
Ratio PV
control to wild
flow
< or >
CV
Auto PV
selection for
safe control
All the above configurations can be achieved in CHEMCAD using the Excel mapping tool.
The diagram below shows the location and component configuration of a typical modern system:
21
(3, 4)
The selection of a control valve, with an appropriate operating characteristic, valve size and
actuator power, is essential for satisfactory control loop performance. Appendix III shows details of
control valve and actuator assemblies.
The control valve must be sized to cover the anticipated maximum and minimum operating limits of
the process. Valve rangeability, maximum flow to minimum flow ratio, of 50:1 is normal. The
maximum flowing quantity should be set at 15% to 50% above the maximum flow required by the
process. An initial estimate would be to set a full size trim valve at half the line size. Line sized
valves, fitted with reduced trims, can avoid the use of piping reducers which create an additional
pressure drop. This simplifies the piping design and installation but these benefits need to be offset
against the increased valve cost.
The gain of a valve is the rated flow under nominal process conditions at full stroke. The flow
characteristic is the relationship between the flow through the valve and the % rated stem travel as
it is varied from 0 to 100%. An equal % characteristic produces a change in flow by a certain % of
present flow for a given increment in stem position and compensates for gain variations elsewhere
in the control loop. An equal% characteristic has an interesting feature in that changing the valve
size does not affect the loop gain, which is directly proportional to flow, making valve sizing less
critical. Refer to section 4.1 for more details
The liquid flow through a valve is a function of the available pressure, size and opening:
F = Cv
Where
Simulation Cases 3.09 and 3.10 review the sizing methods for liquids, steam and gases. The
consequences of an undersized valve are self apparent. An oversized valve will only operate over
part of its travel, requiring a wider proportional band, which results in poor control and at worst case
no control.
The selection of the valve pressure drop, for design, needs to consider the maximum and minimum
inlet and outlet pressures that will be experienced. In many cases, the process will set the inlet and
outlet pressures and the valve must be sized accordingly. However for liquid transfers using
centrifugal pumps a valve pressure drop of ~30% of total system pressure, at full flow, will give
reasonable control. For valves installed in extremely long or high pressure drop lines at least 15%
up to 25% of system pressure drop should be taken.
If the pressure in the valve vena contracta drops below the liquid vapour pressure, bubbles will form
and limit flow through the valve. If the outlet pressure remains below the vapour pressure the
bubbles remain, and the liquid is said to have flashed. If the outlet pressure rises above the vapour
pressure the bubbles will implode, producing cavitation, which sounds as if sand is flowing through
the valve. Flashing and cavitation can cause severe damage to the valve and even pipework and
should be avoided if possible by increasing the inlet pressure or selecting a valve type that has a
larger critical flow factor, Cf. A change in flow direction by installing a streamlined angle valve will
achieve a significant increase in Cf.
The valve actuator transfers the controller output signal to an equivalent % stroke position. The
actuator must be sized to have sufficient power to overcome out of balance forces, which for an
unbalanced valve, will occur at maximum inlet pressure condition. The actuator must fail safe, on
loss of power or controller signal to drive the valve open or closed or stay-put. The fail safe
condition is determined by the process and in some cases equipment protection conditions. A
single acting actuator has a spring that drives the valve open or closed; the stayput condition
requires auxiliary control devices. A double acting actuator requires a positioner to direct the power,
equivalent to the control signal, to extend or retract the valve stem and requires auxiliary control
devices to achieve the desired failure condition.
22
Pmax / Pmin
<2
Constant
<2
< 2
> 2
(Note 1)
< 2
(Note 2)
> 2
Characteristic
Equal %
Linear
Equal %
Linear
Equal %
Linear
Equal %
The inherent characteristic of a control valve is determined by the valve type, for example quick
opening for a butterfly valve and equal % for a globe valve. This characteristic can be modified by
using different types of valve trims or by characterisation of the controller output signal using
various techniques. Refer to Appendix III for further details.
23
4.2
ACTION
FAIL
CLOSED
FAIL
OPEN
Open
FAIL
CLOSED
FAIL
OPEN
Open
20
-0.0625
1.25
Open
V
a
l
v
e
Controller Output(mA)
Closed
20
(mA)
Av = 0.125
Bv = -1.5
Open
V
a
l
v
e
Split Range
Valves
Closed
-0.25
Av = -0.125
Bv = 1.5
V
a
l
v
e
0.0625
Fail Closed
Signal to Open
ACTION
Bv
Av = 0.0625
Bv = -0.25
Fail Open
Signal to Close
Closed
4
Av
Heating
Av = -0.0625
Bv = 1.25
V
a
l
v
e
COEFFICIENTS
12
Closed
4
20
Av = 0.0625
Bv = -0.25
Av = -0.0625
Bv = 1.25
3 Way Valve
12
24
20
5.0
Controller Tuning
When tuning control loops it is important to understand the impact the tuning parameters have on
the process. The safe setting values are:Proportional Band
P %
highest value (minimum gain)
longest time
Integral Action Time
Ti min
Derivative Action Time Td min
shortest time (or off)
The following ground rules are given for guidance:
Proportional Band (P)
Wide P (50 500%) on fast and noisy processes such as Flow, Liquid Pressure and Composition.
Narrow P (1- 50%) on slow processes such as Gas Pressure, Temperature and Vapour Pressure.
Integral Action Time (Ti)
Integral Action Time changes the controller output at a rate proportional to the magnitude of the
error(e). It should never be set so low that the resulting load change imposed exceeds the load
response characteristics of the process. (Consider approaching a corner too fast in a car).
Use low Ti (0.05 2 min) on fast and noisy processes such as Flow and Liquid Pressure and high Ti
(2 120 min) on slow processes such as Temperature, Vapour Pressure and Composition.
Ti is unnecessary on Gas Pressure and is seldom used on Liquid Level as steady flow is usually the
required condition e.g. surge tanks between process units.
Derivative Action Time (Td)
Derivative Action Time changes the controller output at a rate proportional to the rate of change of
error
de
and provides an anticipatory characteristic.
dt
It should never be used on noisy processes such as Flow, Liquid Pressure and Liquid Level. Td is
used on Temperature, Vapour Pressure and Composition and should be set conservatively(short
Td) to protect against control instability resulting from unanticipated load changes.
Td set in the range
Ti
T
< Td < i
4
2
is considered reasonable.
Controller Type
P only
P+I
P+I+D
Td min
Tu / 8.0 = Ti /4
This procedure can be used as a starting point for tuning control systems on batch processes, Tu
being established at a stable condition. Typical periods (Tu) for batch reactors vary in the range 15
to 60 minutes.
25
6.0
The following provides a summary of key points to achieve rapid results from the dynamic models.
Run time and step size must be selected to ensure consistency with real time behaviour. Dynamic
simulation is provided with a speed up facility called On-line Training System(OTS) which allows for
simulation speed up over real time by a factor up to 50. Under Run Convergence use Run One
Time Step for Dynamic Simulation to allow study of model behaviour.
Control valve rangeability based on the ratio of the maximum to minimum controllable flows
depends on the entire control loop, use 50 for equal % characteristic and 10 for linear characteristic.
The position of the control valve in the model is critical and the rules can be summarized:It requires inlet (Pi) and outlet (Po) pressures to be specified where Pi is taken from the inlet stream.
Po can be specified directly or overtaken indirectly by an assigned UnitOp in the model.
A mass balance must be maintained around the control valve. To achieve this, locate immediately
after a Feed Arrow (direct flow adjust), after a Stream Reference (SREF) or Divider (DIVI) (indirect
flow regulation). Dynamic Vessel, Dynamic Reactor and Dynamic Column can also be assigned for
indirect flow regulation. The optimal mass flowrate transfer is used to set up Indirect Flowrate
Regulation.
If a value for valve flow coefficient Cv is not known use the Equipment Valve Sizing option which will
provide a value at stated stream conditions. The CHEMCAD valve data base smallest valve Cv is
9.0; a smaller valve Cv requirement is determined by separate calculation which can be linked to the
model using the Excel Data Map facility. An incorrectly sized valve will result in dynamic model
convergence problems and an unstable condition in the model.
If the controller input goes out of the calibrated range the control output continues to change
resulting in unreal conditions.
The control valve can be operated in various operating modes which allows the simulations to be
driven to the desired condition. These features allow control valve positions to be set manually or
automatically from an assigned controller.
26
7.0
(2)
Measurement Sensors
This section has been included as an introduction to the process engineer who is not familiar with
process measurement technology and its application. Reliable and accurate measurement sensor
selection is one of the key parameters in the successful operation of a process plant.
Measurement sensor technology is a key driver in the development of modern industrial processes.
The technologies, to measure and transmit process parameters, such as flow, level, temperature
and pressure, have developed significantly since the 1960s.
Impulse lines, used to connect instruments to the process, are found less frequently on new
installations and are being replaced on existing ones. Where used, they require specialist
knowledge during design, installation and maintenance for reliable measurement. Modern
techniques have simplified maintenance and enhanced plant performance.
Data transmission
Transmission technology development has allowed universal application of self-powered two wire
4-20 mA dc signals. This has eliminated power supply and special sensor cabling for magnetic flow
meters, thermocouples and resistance bulbs, reducing cost and simplifying installation.
In the 1980s, microprocessors facilitated the transition from signal to information based process
automation. Smart transmitters provided bi-directional digital communication and diagnostics
capability with HART (Highway Addressable Remote Transducer) protocol. The 4-20 mA and
HART digital signals are transmitted over the same wiring, providing a centralised capability to
configure, calibrate, characterise and diagnose devices in real time together with reporting
capability. Data can be captured from multi-parameter devices without additional hardware,
providing predictive maintenance capability.
Development in fieldbus digital communication allows field devices to be connected using a single
cable bus structure, reducing cabling, installation time and cost. Fieldbus is a device level network
that sacrifices speed for security.
There are several protocols available with Modbus , PROFIBUS PA and FOUNDATION being the
most common. Modbus was the earliest protocol and has connectivity with Ethernet and other
fieldbus protocols. PROFIBUS PA was developed in Europe and with PROFINET has Ethernet
connectivity. FOUNDATION fieldbus was developed in North America by ISA, suppliers and
users.
Fieldbus technology is more complex and costly, requiring suppliers to provide sensor options to
meet the different standards. Fieldbus selection is guided by plant layout, sensor interface
capabilities and data management infrastructure. An industry working group is currently developing
the capabilities of Electronic Device Description (EDD) technology.
27
7.1
Flow Measurement
(2)
Accurate flow measurement is a key element in process productivity. Various types of flow meters
are used, with the orifice plate and differential pressure transmitter being predominant prior to 1965.
New materials, manufacturing techniques and micro-electronics have resulted in significant
developments.
Measurement accuracy is quoted for a given turndown ratio (max/min reading) as % span (max
min reading), % full scale or % actual, with the latter being the highest specification for same value.
Accuracy is important for stock and custody monitoring, with reproducibility, the ability of the sensor
to reproduce its reading, being important for control.
The positive displacement flowmeter, one of the earliest meters, is based on a fixed volume of fluid
causing a known mechanical displacement to generate a calibrated pulse. Configurations based on
vanes, gears, pistons, or diaphragms are available. Oval gear meters have reduced slippage,
achieving accuracies of 0.05% rate for = 0.2 - 5 cps and 0.25% rate for > 5cps over 10:1
turndown. Advancements in accuracy from 0.5% rate have been achieved by temperature
compensation, multivariable flow computation and HART, making meters suitable for custody
transfer. Fluid viscosity affects the pressure drop and presence of solids or entrained air can cause
mechanical damage.
The axial turbine flowmeter consists of a rotor, driven at a rate proportional to the fluid velocity, to
generate a calibrated pulse. Developments in materials have improved bearing design to extend
the application and improve robustness, achieving accuracies of 0.5% rate for = 0.8-2 cSt over a
15:1 turndown, depending on size and conditions. Installations require an upstream strainer and
straight pipe section with custody transfer requiring upstream flow straighteners. Meter is suitable
for bi-directional flow.
The magnetic flowmeter measures the voltage generated by an electrically conducting liquid flowing
through a magnetic field, which is proportional to fluid velocity. Fluid contact electrodes can
measure liquids with conductivities as low as 2 S/cm and with non-contact capacitive signal pickup electrodes down to 0.05 S/cm. Meters can achieve accuracies of 0.2% rate over a 10:1
turndown depending on size and conditions and contribute no additional pressure drop. Installation
requires minimum 5D upstream straight pipe section and is suitable for bi-directional flow. Preferred
method for conducting liquids, including corrosive liquids, slurries, sludges, liquids with abrasive
solids.
The vortex shedding flowmeter measures the frequency of vortices formed by a fluid flowing across
a bluff non-streamlined body, which is proportional to fluid velocity above the minimum flow
condition at which vortex shedding ceases. Typical accuracies claimed for liquids are 0.5% rate for
Reynolds Number (Re) >20,000 and for steam/gases 1.0 % rate for velocity < 35 m/s.
Advancement in accuracies are being achieved by compensating for varying Re and process
conditions, with reduced bore meters extending the measuring range. Installation requires a
minimum 15D upstream and 5D downstream straight pipe section with pressure drop being ~2
velocity heads. Preferred method for steam and with integral temperature compensation gives mass
flow of saturated steam.
The Coriolis flowmeter measures the twist created by fluid flowing through an oscillating single or
dual tube assembly, which is proportional to mass flow and is independent of fluid viscosity and
density. This is a one for all multi-variable meter and provides density and temperature
measurement allowing derivation of other variables such as volume flow, solids content and
concentration. Typical accuracies claimed for liquids are 0.1% rate and for gases 0.35 % rate
over a 20:1 turndown, and is suitable custody metering. There are no special installation
requirements but pressure drops can be significant. If capital cost is not an issue, this is preferred
method for non-conducting liquids.
The thermal dispersion flowmeter measures the cooling effect of a gas as it passes over a heated
transducer, which is proportional to the mass flow, thermal and flow properties of the gas. Typical
accuracies claimed are 1.5% reading over a 10:1 turndown. Installation may require up to 20D
upstream straight pipe and 5D downstream no additional pressure drop. This is preferred method
for pure gases and constant composition gas mixtures.
28
The ultrasonic time of flight flowmeter measures the time difference between the paths of two or
more ultrasonic signals beamed in opposite directions. Meters can be clamp on or insertion type.
The accuracy is enhanced by characterising the varying velocity profiles due to changes in Re and
process conditions. Multi-path chordal insertion meters can achieve accuracies of 0.25% rate over
a 10:1 turndown whilst clamp on meters claim 1.0% rate. Installation may require up to 20D
upstream straight pipe without a flow conditioner and 3D downstream. The meters can be used on
bi-directional flow and contribute no additional pressure drop. Ultrasonic meters are now gaining
acceptance for liquid custody transfer. Time of flight ultrasonic flowmeters are not suitable for use
on liquids containing entrained gases or solids.
The following sequence gives an indication of comparitive costs with the lowest cost first:
Turbine Magnetic Vortex Thermal Ultrasonic Positive Displacement Coriolis.
The chart shows typical flow meter selection criteria and applications:
29
7.2
(2)
Liquid level measurement is key to reliable and safe process plant operation. Normally flows are
held steady whilst levels are allowed to change within limits, requiring reproducibility. Accuracy is
important for tanks used for stock and custody control.
The hydrostatic continuous, indirect, level method measures the pressure due to liquid level and
density plus over-pressure. The sensor measures the pressure difference between this pressure
and a reference pressure, normally atmospheric, so is not preferred for vacuum and pressure
service. Instruments can be flanged mounted or rod insertion type, the latter not being
recommended for turbulent conditions. Typical accuracies claimed are 0.2% reading, and is
dependent on process fluid properties and conditions.
The displacer continuous or point level method measures the change in buoyancy via a torque tube
or lever arrangement. The continuous measuring range is set by the displacer length immersed in
the tank or external cage, which is preferable on noisy applications. The point method uses a float
with the range being limited by the length of the float arm.
The nucleonic point or continuous, non-contact, level method measures the signal strength of a
radioactive source beamed across a vessel. Independent of fluid properties and has typical ranges
of 0.24 to 3.36m. Typical accuracies claimed are 2% reading. This is the preferred method for
controlling level in flash vessels and reboilers under all temperature and pressure conditions
.
The radar point or continuous level method measures the travel time of an impulse transmitted and
reflected from the liquid surface. Interference echoes resulting from tank internals and agitators are
suppressed and signals characterised to give liquid volume. The sensor has no contact with the
liquid but is exposed to head space conditions, which do not affect the measurement. The
reflectivity requires the liquid dielectric constant(R) to be 1.4 (hydrocarbons 1.9 - 4.0, organic
solvents 4.0 - 10 and conductive liquids >10). The antenna and signal conditions are adjusted to
suit the process, with guided radar being used for low R and turbulent conditions. Method is
suitable for custody transfer with accuracy 0.5mm being claimed.
The capacitance point or continuous level method is suitable for liquids which can act as dielectrics.
The measurement is more sensitive when the difference R of the liquid and the vapor space or
between the two liquids are higher. Special designs, involving coated and twin probes, are used
when R < 1.0, conductivities > 100 mho, coating effects or vessel material is non conducting.
Typical accuracies claimed are 0.25% span and is dependent on fluid properties, so is not suitable
for changing conditions. Maximum conditions 200C at 100 bar and 400 C at 10 bar.
The ultrasonic point or continuous level measurement is based on the time-of-flight principle. A
sensor emits and detects ultrasonic pulses which are reflected from the surface of the liquid. The
method is non-invasive, with some types being non-contact, and is not affected by R, conductivity,
density or humidity. Maximum conditions 150C at 4 bar.
Load cells, based on strain gauge or piezoelectric, measure the weight of the process vessel plus
contents. Individual load cell accuracy of 0.03% full scale is achievable but overall performance is
dependent on correct installation practices preventing external forces, due to associated piping and
equipment. For vessels with jackets, agitation and complex piping it is difficult to obtain an
acceptable accuracy. When the container can be totally isolated, as in final dispensing and filling
applications, precision weighing can be achieved.
The vibrating tuning fork principle is used to detect point liquid level but is unsuitable in viscous and
fouling applications. Maximum conditions 280 C at 100 bar.
The conductivity point level method requires a liquid conductivity > 0.1 mho and is frequently used
on utility and effluent pump control systems.
The following sequence gives an indication of comparitive costs, with the lowest first:
Conductivity Capacitance Tuning fork Hydrostatic Displacer Ultrasonic Load cell
RadarNucleonic.
30
ELECTRONIC
RADAR
POINT
TYPE
MECHANICAL
MECHANICAL
TYPE
DISPLACER 1
(Interface)
(Interface)
NUCLEONIC 3
SERVO
FLOAT 1
(Interface)
ELECTRONIC
TYPE
FLOAT 1
RADAR 2
MAGNETIC
FLOAT 1
NUCLEONIC3
TUNING
FORK1
CAPACITANCE 1
CAPACITANCE 1
(Interface)
ULTRASONIC 2
ULTRASONIC 2
HYDRO
STATIC 1
PROCESS
MIXED PROPS
(K >1S/cm)
APPLICATION
CLEAN
STORAGE
TANK FARM
FLUID
DAY/SURGE
DUTY
RADAR 1, 2
GUIDED
400C, 399barg
RADAR 1, 2
400C, 399barg
NUCLEONIC 3
NUCLEONIC 3
(Noisy, flash)
(Noisy, flash)
2% span
2% span
RF CAPACITANCE 1
260C , 690barg
(External cage)
ULTRASONIC 2
150C , 2barg
CONDUCTIVITY 1
RADAR 1
SERVO
FLOAT 1
DISPLACER 1
450C 0.5% r
RADAR 1
(Inventory)
NUCLEONIC 3
2% span
RF CAPACITANCE 1
RF CAPACITANCE 1
(Inventory)
HYDRO STATIC 1
(Atmospheric)
Mixed Props:
Flashing, boiling, foaming
Surface fouling, precipitates
Varying density, viscosity
LOAD
CELLS 3
31
ULTRASONIC 2
HYDRO STATIC 1
(Atmospheric)
Appendix I
Flow Meter Considerations
To model piping systems, involving special items such as flow meters, CHEMCAD provides a
facility under the Valve Data entry Window in the Pipe UnitOp for resistance parameters to be
entered in various formats. The following guidelines should be considered when selecting flow
meter sizes.
74
3
to 79 m/s density kg/m
50
3
to 260 ft/s density lb/ft
Process fluid viscosity requires the Reynolds Number to be greater than 20000
Linear performance is achieved for Reynolds Number in the range 20000 to 7.0 E06
32
Appendix I
Flow Meter Considerations
Differential Head Flowmeters
The differential pressure measured and unrecovered pressure loss across a square edge
concentric orifice plate is dependent on the pressure tap location; as shown in the diagrams below.
It can be seen that full flow taps (2D and 8D) measures the permanent pressure loss and should
be used for restriction orifice calculations.
For liquids
Q = K d2 C
h
f
W = K d2 C h
For gases
W = K d2 C
h pf M
Tf
The diagram below shows the dependency of flow coefficient C on Re and d/D ().
d/D () ratios 0.6 are preferred. For > 0.6 viscosity effects are magnified combined with
increased sensitivity to upstream piping configurations.
33
Appendix II
Flow Meter Types
Positive Displacement Meters Mechanical and Rotating
www.omega.co.uk
Turbine Flowmeter
www.omega.co.uk
Coriolis Mass
Flowmeter
34
(3)
Appendix III
Control Valve and Actuator Types
http://www.documentation.emersonprocess.com/groups/public/documents/book/cvh99.pdf
The referenced Fisher Handbook, first published in 1965, provides a comprehensive and detailed
coverage of control valve and actuator technology.
Globe Valve with modified cage trims
35
(3)
Appendix III
Control Valve and Actuator Types
Control Valve Body Types
Reverse Double-Ported Globe
Angle
Reverse Acting
36
Dead time(transport lag or distance-velocity lag) is the interval after the application of a control
device output change for which no process(measurement) response is observable. It occurs when
mass or energy is transported along a path length with velocity constituting the delay.
For a 1m dead time the natural period is fixed at 2m and a proportional band of 200% will give
amplitude damping. Dead time is a difficult element to control and process design should
concentrate on minimising or eliminating if practical.
37
(1, p18)
Capacity is a UnitOp where mass or energy can be stored such as in a tank with volume V. The
rate of change of tank contents equals the difference between inflow and outflow.
dv
dt
The non self regulating case, having a fixed outflow using a metering pump, will result in the tank
overfilling or emptying for any change in inflow; an integrating process cannot balance itself.
Replacing the pump with a valve will result in a self regulating process.
38
A common method used in the measurement of tank level is a displacer mounted in an external
chamber with the displacer being linked to the transmission system by means of a torque tube.
This measuring chamber creates a two capacity process, that introduces a lag, with a time constant
of chamber volume divided by the maximum flow rate at which liquid can enter.
This arrangement creates hydraulic resonance with a period that varies with the distance between
the bounded surfaces. The natural period for this arrangement can be derived:
o = 2
L1 + L 2
2g
Throttling the chamber connecting valve will reduce the amplitude of this resonance but not the
period. However this enables the use of a narrower proportional band and at the same time
reducing the valve movement due to waves.
39
40
41
42
43
2
P < Cf (Ps )
Gf
P
C v = 1.16 q
Where
q
Cf
Gf
P
Cv =
1.16 q
Cf
Gf
Ps
Ps = P1 0.96 0.28
Where
P1
P2
Pv
Pc
P v
Pv
Pc
Laminar flow can result at high viscosity or when the valve P or Cv is small.
Calculate turbulent flow Cv and laminar flow Cv and use the larger value as the required Cv.
For laminar flow we have:
q
C v = 0.032
P
44
0.667
The CHEMCAD Control Valve Sizing Liquid model is set up with streams suitably configured for
liquids as shown below. A dummy stream is used to determine liquid vapour pressure. Data Map is
defined to interface with the spreadsheet.
The Divider UnitOp 1 allows for transfer of Control Valve calculated flow to Stream 2 to maintain
the mass balance around the Divider. A globe control valve can be sized by selecting Sizing
Control Valve on the main Toolbar. Sizing is be carried out using the stream properties of the
selected stream. The data entry Window is as follows:
This facility has limited use as it only applies to globe control valves of sizes 1 in and the noncritical flow condition.
For a more rigorous design the user should enter manufacturers data into the Control Valve data
entry screen.
45
Results
Sizing spreadsheet Control Valve Liquid Sizing has been created to analyze the CHEMCAD model
calculation results and to obtain Physical Property Data to allow for validation of control valve
results. Sizing parameters are calculated using the relevant equations.
The sizing spreadsheet for liquid control valve sizing is shown below:
The spreadsheet is configured to facilitate the sizing of most types of control valve under noncritical, critical and laminar flow conditions. It also allows for the entry of valve characteristic, critical
flow factor Fa from manufacturers data and for the effect of reducers.
The spreadsheet allows for the position of Control Valve UnitOp 2 to be adjusted to obtain the Cv
at the specified flow conditions.
46
(5)
CHEMCAD provides facilities for the sizing of concentric orifice plates used in the measurement of
fluid flow rates. The methods are based on Principles and Practice of Flow Meter Engineering by
L.K.Spink, Foxboro Company, 1967.
The fundamental equations are presented as follows:
The equation for non-viscous liquid flow is given by:
Wm
N D Fa Fm Gf h m
S =
Where
Wm
D
Fa
Fm
N
Gf
hm
Fa = 1 + 2 (t f 68 )
Where
tf
Kr =
Where
C
d
2
1
C =
2
C
Cd
(1 4 )
0 .5
S =
Wm
N D Fa Fm Fc Gf h m
2
The application of the viscosity correction factor Fc for plant operational measurements and control
is rarely justified. Viscosity limits for 1% calculation tolerance vary in the range of 1 to 8 cps
depending on the ratio, keeping <0.6, and pipe size. Fc can vary in the range of 1.0 to 1.09. Refer
to L.K.Spink Flow Handbook for more information.
47
(5)
S =
Where
Wm
359 D Fa Fm Fc Y f h m
f =
Fc
mw
Pf
Tf
Y
m w pf
10.73 T f zf
2
3
2
S = 0 . 598 + 0 . 01 + 0 . 00001947 10
4 . 425
The sizing procedure is to determine an initial S and then d/D assuming Fc = 1 and Y =1. Then use
alignment chart to determine Y and obtain new d/D from modified S/Y value.
48
Section 4
Thermodynamics
Contents
1.0
2.0
2.1
2.2
2.3
2.4
3.0
3.1
3.2
3.3
4.0
4.1
4.2
5.0
Introduction
Thermodynamic Fundamentals
Thermodynamic Energies
Gibbs Phase Rule
Enthalpy
Thermodynamics of Real Processes
System Phases
Single Phase Gas
Liquid Phase
Vapour liquid equilibrium
Chemical Reactions
Reaction Chemistry
Reaction Chemistry Applied
Summary
Appendices
I
II
III
IV
V
VI
VII
VIII
IX
X
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
1.0
Introduction
The selection of a suitable thermodynamic model for the prediction of the enthalpy H value and
the phase equilibrium K value is fundamental to process modelling. An inappropriate model
selection will result in convergence problems and erroneous results. Simulations are only valid
when the appropriate thermodynamic model is being used. The selection process is based on a
detailed knowledge of thermodynamics and practical experience. Most simulators are provided with
Thermodynamic Wizards to aid selection which should be used with caution.
The selection process is guided by considering the following:Process species and compositions.
Pressure and temperature operating ranges.
System phases involved.
Nature of the fluids.
Availability of data.
There are four categories of thermodynamic models:Equations-of-State (E-o-S)
Activity coefficient ()
Empirical
Special system specific
This review is not intended to be a rigorous analysis of the methods available or in their selection
but is offered as an aide memoire to the practicing engineer who is looking for rapid, realistic
results from his process models.
The study of complex systems invariably involves extensive research and considerable investment
in manpower effort by specialists. There are extensive sources of physical property data available
from organisations such as:
DECHEMA
DIPPR
TV NEL Ltd
www.dechema.de,
www.aiche.org/dippr/ ,
www.ppds.co.uk .
This review presents selection methods developed in discussions with engineers in the field. The
validity of the thermodynamic models being used should be tested against known data whenever
possible.
50
2.0
2.1
Thermodynamic Fundamentals
(1)
Thermodynamic Energies
The thermodynamic fundamentals of fluid states in relation to energies and phase behaviour needs
to be thoroughly understood.
Four thermodynamic variables determine six thermodynamic energies:
Intensive variables
Pressure
(P )
Temperature (T)
Intensive variable
Extensive variable
V or S
P or T
U =T S +F
G =F +P V
Enthalpy
H =T S +F +P V
H =U +P V
Capacity Factor
0
dU = i i dni
Where:
i dni = 0
For equilibrium
dF = 0 (constV & T)
dU = 0 (constS & V )
It can be shown that
dG = 0 (constP & T )
dH = 0 (constS & P )
G = i ni
i
51
2.2
(2)
Intensive variables
The number of intensive variables that can be independently specified for a system at equilibrium is
called the number of degrees of freedom F and is given by the Gibbs Phase Rule.
In a system involving no reactions this is given by:
F = 2 +m p
Where:
m
p
= no of chemical species
= number of system phases
F = 2 +m r p
When defining a stream condition in a simulation the phase rule applies.
Consider a single component liquid in equilibrium with its vapour and an inert.
This gives
m = 2
p = 2
F = 2
Two variables P and T or Vapour fraction (v) with T or P will define the stream.
For a binary liquid system one degree of freedom is consumed by the composition leaving either P
or T to be specified. In a VLE system it is preferable to specify P which then allows system analysis
using Txy plots.
When setting up a Flash UnitOp applying the phase rule will ensure that the relevant flash
conditions are being set. The stream flash calculation can be used to determine the boiling point
and dew point of mixtures with and without inerts present by applying the following:
The bubble point of a liquid, at the given pressure, is determined by a flash calculation at a vapour
fraction of 0.
The dew point of a vapour, at the given pressure, is determined by a flash calculation at a vapour
fraction of 1.
Note that for a pure component the bubble point and the dew point are identical so a flash
calculation at a vapour fraction of 0 or 1 will yield the same result.
52
The figure shows the Txy diagram for Benzene/Toluene, a near ideal mixture. The bubble point for
a given composition is read directly from the liquid curve and the dew point is read directly from the
vapour curve.
The bubble point of a mixture is determined by trial and error from value of Tbp that satisfies:
( )
*
P = xi pi Tbp
The dew point of a mixture is determined by trial and error from value of Tdp that satisfies:
y
iP
*
p Tdp
i
( )
53
=1
2.3
Enthalpy
Enthalpy is the sum of the internal energy U and the external free energy PV where:
H =U +P V
The heat supplied is given by:
dQ = dU + P dV
The sign convention should be noted and is + for heat added and dU gain in internal energy:
dU = Cv dT
The specific heat at constant pressure Cp is related to heat input:
dQ = Cp dT
The adiabatic index or specific ratio is defined:
Cp
Cv
Cp Cv = R
The heating of a liquid at constant pressure e.g. water is considered in the figure below. The figure
shows the relationships between the enthalpies in the different phases namely the sensible heat in
the liquid phase, the latent heat of vaporisation during the vapour liquid equilibrium phase and the
superheat in the gas phase.
54
Enthalpy is calculated using Latent Heat (LATE) in the liquid and vle phases and E-o-S (SRK) in the
superheated or gas phase.
Appendix I reviews the calculation methods adopted in CHEMCAD. A standard reference state of
298K for the liquid heat of formation is used providing the advantage that the pressure has no
influence on the liquid Cp. It is worth noting that if water is the only component, the simulation
enthalpies will match steam table values.
The enthalpy method used will depend on the K-value method selected as detailed in Appendix II.
CHEMCAD forces the following H-values from K-value selected.
Equilibrium-K
Peng Robinson (PR)
Grayson-Streed-Choa-Seeder(GS)
ESSO
Benedict-Webb-Rubin-Starling (BWRS)
AMINE
Enthalpy-H
PR
Lee Kessler (LK)
LK
BWRS
AMINE
55
2.4
(7)
G = H T S
The values for H are determined from the heats of formation of the components and for S from
0
thermodynamic property tables. Superscript
indicates materials present in standard state at
298K.
For isothermal processes at low temperature the H term is dominant. At absolute zero S and T
are zero so G = H. The relationship shows S becoming of increasing importance as the
temperature increases.
Adsorption Processes
The enthalpy change, where G being necessarily negative, is:
H = G + TS
All adsorptions with negative entropy change, which comprise all physical and the great majority of
chemical adsorptions, are exothermic.
Evaporation Processes
When a liquid boils the vapour pressure is equal to the atmospheric pressure and the vapour is in
equilibrium with the liquid. If there is no superheat the process is reversible and G = 0 and the
entropy change can be calculated:
H vap
TB
Entropies of vaporisation, at these conditions, have values near 88 J/molK, and substitution in the
above gives Troutons rule. However in the case of water, due to significant hydrogen bonding, the
entropy change on evaporation is larger at 108.8 J/molK.
Endothermic Chemical Processes
The link between Gibbs Free Energy and the reaction equilibrium constant K is represented by the
equation:
G = -RT log K
A reaction will proceed provided G is negative. The reaction temperature can alter the sign and
therefore the process feasibility.
Chemical Equilibrium
For a reaction at equilibrium (all reactions can be considered equilibrium since no reaction goes to
completion) there is no net reaction in either direction and we have:
G = 0
In CHEMCAD the Gibbs reactor is based on the principal that at chemical equilibrium the total
Gibbs free energy of the system is at its minimum value. The Gibbs reactor can be used in the
study of combustion processes including adjustment of air to fuel ratios and calculation of the heats
of reaction.
56
3.0
System Phases
There are three phase states namely solid, liquid and gas. Processes comprise either single phase
or multiphase systems with separation processes involving at least two phases.
Processes involving solids such as filtration and crystallisation, solid liquid systems and drying,
solid gas system are special cases and receive no further consideration here.
The primary area of interest for thermodynamic model selection involves two phases.
Liquid liquid systems, such as extraction and extractive distillation, where liquid liquid
equilibrium (LLE) is considered
Vapour liquid systems, such as distillation, stripping and absorption, where vapour liquid
equilibrium (VLE) is considered.
The figure below shows the inter-relationships between the system phases for a series of isotherms
based on the Equation of State (E-o-S) due to van-der-Waal. This figure provides the first indication
of the validity of making a thermodynamic model selection for the K-value on the basis of the
system phases namely single phase gas by E-o-S and VLE by activity coefficient.
Simulator thermodynamic settings include the Global Phase Option which can be set for
Vapor/Liquid/Solid or Vapor/Liquid/Liquid/Solid. When carrying out azeotropic distillation or solvent
extraction the Vapor/Liquid/Liquid/Solid option should be selected.
57
Section 5
Fluid Flow and Pipe Networks
Contents
1.0
2.0
2.1
2.2
3.0
3.1
3.2
3.3
4.0
4.1
4.2
4.3
5.0
5.1
5.2
5.3
5.4
5.5
5.6
5.7
I
II
III
IV
V
1.
2.
3.
4.
5.
6.
In this section cases are presented to provide an understanding of CHEMCAD simulation and its
capabilities for sizing of piping networks and the optimisation of flow distribution.
For practice you can build the models or use the models provided as downloads from the electronic
media available. It is strongly recommended that you work with a copy of the jobs.
Case/File Name
5.01
5.02
5.03
5.04
5.05
5.06
5.07
5.08
Description
Pipe size changes and fittings
Control valve sizing and flow split
Piping design
Pump sizing and pipe branches
Cooling water distribution
Gravity flow between reservoirs
Piping Flow Split
Complex pipe network
58
1.0
Fluid flow through a pipe, valve, fitting or nozzle will result in a pressure drop in the direction of the
flow. Process and manufacturing facilities require extensive piping systems to distribute services
and process fluids at acceptable flows and pressures for satisfactory equipment operation. Pipe
networks can involve complex dependencies between process equipment, which can result in
serious restrictions on plant throughputs if poorly designed. Modelling of networks, to represent the
fluid flows and pressures through the pipes, valves and fittings under all anticipated operating
conditions, can avoid these bottlenecks..
The flowrate through a component may be calculated as a function of the inlet and outlet pressures,
giving three variables. If two of the three variables are specified the third is dependent. Specification
of pressure at various junctions in a pipe network allows the system to be represented as a system
of dependent equations. If sufficient flowrates and pressures are specified in the network, the
unknown flows and pressures can be calculated.
The network model allows for the simultaneous solution, if sufficient constraints are specified, to
converge on the unknown pressures and flowrates. A simple flow system is shown below:
The known variables are the pipe geometry, pipe exit pressure, and flowrate. A single equation can
be used to solve for the pipe inlet pressure as a function of the known variables.
Fluid mechanics enables the calculation of fluid flowrate through a pipe, valve or nozzle as a
function of the inlet and outlet pressures. Performance curves allow for the calculation of fluid flow
through a compressor or pump as a function of the inlet and outlet pressures. The diagram below
shows equipment that calculate flowrate as a function of inlet and outlet pressure. These UnitOps
are referred to as flow scaling UnitOps.
59
A node represents a point in the piping network where a change in pressure occurs due to elevation
change, flow through pipe and fittings or flow through equipment that changes pressure such as
pumps, compressors or valves.
For the design of a piping network it is necessary to determine pressure between all UnitOps that
calculate pressure as a function of flowrate. The node sets the pressure on one side of a UnitOp
that calculates the pressure as a function of flowrate.
The pressure at a node may be specified or calculated by the simulator. The flowrate in and out of a
node may be specified or calculated and can be specified at the node or calculated as dependent
on adjacent UnitOps.
The node can set a fixed value on the flowsheet. For piping network calculations there are points on
the flowsheet where either the pressure or flowrate is known.
A node may be used as a divider. Outlet streams from the node will be at the pressure of the node.
Outlet streams will all have the same temperature and composition but flowrates may differ. The
flowrates may be specified, as set by pipe/valve, or at fixed flowrates. Only one outlet stream
flowrate may be a free outlet.
A node may be used as a mixer. The inlet streams to the node will all have the same pressure as
the node. Only one inlet stream flowrate may be a free inlet.
Steady state UnitOps may be used on a piping network. A constant pressure drop may be entered
for a non-scaler, steady state UnitOp, such as a heat exchanger. The inlet and outlet nodes will
recognize this pressure drop.
Between two nodes there must be one flow rate scaler. A heat exchanger and a pipe can be
between two nodes. A constant pressure drop may be specified for the heat exchanger and it will
affect the pressure drop between the two nodes. A heat exchanger cannot be the only UnitOp
between two nodes, as the heat exchanger does not have an effect on pressure.
60
Section 6
Emergency Relief Systems
Contents
1.0
2.0
2.1
2.2
3.0
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
4.0
4.1
4.2
4.3
4.4
4.5
5.0
Introduction
Relief Sizing Fundamentals
Vessel Models
Vent Flow Models
Definitions and Assumptions
% overpressure
Friction factors
Relief Device Discharge Coefficients
Equivalent Lengths
Discharge Pipe Considerations
Inlet / Outlet Pipe Sizing Rules
Environment F Factor
API Valve Selection
Exposed Area Considerations
Methods for Evaluating Gas Density
Sizing Methods Basis
Relief System Sizing General
Relief System Sizing Flow
Heat Models
Sizing Methods for Single Phase Flow
Sizing Methods for Two Phase Vapor Liquid Flow
Nomenclature
Appendices
I
II
III
IV
V
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
J. Wilday and J. Etchells, Workbook for Chemical Reactor Relief System Sizing
HSE Contract Research Report 136/1998.
Emergency Relief System Design Using DIERS Technology
The Design Institute for Emergency Relief Systems Project Manual, 1992.
Sizing Selection and Installation of Pressure Relieving Devices in Refineries
th
API 520, 7 Edition, January 2000.
th
Guide for Pressure Relieving and Depressurising Systems, API 521, 4 Edition, March
1997.
th
Venting Atmospheric and Low-Pressure Storage Tanks, API 2000, 5 Edition, April 1998.
D.K.Das, Specialist Consultant, Chemstations Private Communications.
H.K.Fauske, Pressure Relief and Venting, IChemE Symposium Series No 102, 1987.
L.K.Spink, Principles and Practice of Flowmeter Engineering
The Foxboro Company, The Plimpton Press, 1967.
Flow of Fluids Crane Company, Publication 410M, 1998.
J.C. Leung, Simplified Vent Sizing for ERS in Reactors and Storage Vessels
AIChemE Journal, Vol 32, No 10, October 1986.
S.S. Grossel, Design and Sizing of Knock-Out Drums/Catchtanks for Reactor Emergency
Relief Systems, Plant Operations Progress (Vol 5, No.3) July 1986.
J. Scoville, The Use of Certified KR for Rupture Disks, Technical Note, Oseco Ltd.
61
1.0
Introduction
To achieve safe operation of chemical processes, particularly when involving exothermic reactions,
requires a combination of preventative and protective measures.
Preventative measures minimise the occurrence of an event and include: Automatic control systems including the use of independent hardwired alarm and trip systems
based on Layer of Protection Analysis (LOPA) and Safety Integrity Level (SIL) analysis.
Provision for appropriate manual intervention.
Protective measures mitigate the consequences of a runaway reaction and include: Emergency pressure relief.
Crash cooling.
Reaction inhibition.
Drown out.
This section reviews the techniques associated with the design of emergency pressure relief and
blowdown systems. An emergency relief system comprises the following key components: Primary pressure relief device.
Relief system piping and headers.
Containment system.
Treatment system.
Atmospheric dispersion.
Specific design techniques are required for each component of the system and process modelling
of the total integrated system can demonstrate the adequacy of the total design.
A pressure relief device is either a relief valve, a bursting(rupture) disc or a combination of both. On
chemical reactors the bursting disc is the most favoured due to being able to handle the following
conditions:
Rapid pressure rise with full relief area available, except when vacuum support used.
Toxic fluids where no leakage past a safety valve is permitted.
Corrosive fluids that may cause progressive deterioration of a safety valve.
Fluids that may deposit solids or build ups that interfere with safety valve operation.
62
Require a larger allowance between the operating pressure and the set pressure.
If the operating pressure and the set pressures are too close, the disc can fail prematurely due
to pressure pulsations.
Loss of containment of reactor contents on operation i.e. valve does not reseat, unless used in
conjunction with a safety valve.
To mitigate against the loss of containment, due to operation of a bursting disc sized for the worst
case scenario, the installation of a smaller bursting disc/relief valve combination in parallel set at a
lower relief pressure can be considered. This smaller system operates in the event of nuisance
pressure build ups, due to maloperation, without total loss of containment. This technique has the
disadvantage, due to the poor set pressure tolerance of bursting discs typically, +10%, of
significantly increasing the set pressure of the main bursting disc to ensure the smaller disc
operates first. This may lead to an undesirable situation particularly when protecting for exothermic
runaway events.
The relief pressure, at which the relief device is fully open, should be set at the lowest pressure
practicable consistent with preventing nuisance operations for the following reasons :
For most exothermic runaway reactions, the reaction rate and heat release rate increases
exponentially with temperature. For a vapour pressure system, a low relief pressure means a
low relief temperature and hence a relatively low rate of heat release. The relief area required is
directly proportional to the rate of heat release by the reaction.
For a relief system venting a two-phase mixture, pressure relief acts to remove reactants from
the reactor. A low relief pressure allows a greater margin between the relief pressure and
maximum permitted pressure, and advantage is taken of this by the sizing methods to yield a
smaller relief area.
63
Section 6
Emergency Relief System Simulation Cases
Case/File Name
6.01
6.02
6.03
6.04
6.05
6.06
6.07
6.08
6.09
6.10
6.11
6.12
6.13
6.14
6.15
6.16
6.17
Description
Liquid Relief Sizing, API 520 Part 1, 3.8.2 page 53
Vapour Relief Sizing, API 520 Part 1, 3.6.2.2 page 44
Steam Relief Sizing, API 520 Part 1, 3.7.2 page 51
DIERS Example Appendix II-E (page 113)
DIERS Example Appendix II-E (page 113)
DIERS Final Report Appendix D (page 181)
DIERS Non-Tempered Relief Example 15 (page 438)
DIERS Non-Tempered Relief Example 15 (page 438)
DIERS Tempered Relief Example 11 (page 432)
DIERS Tempered Relief Example 11 (page 432)
DIERS TestT3A (Vessel page174 Test Data p.198)
DIERS Two Phase Rigorous Design Case
Relief Vent Piping Manifold
Flare Header Design
DIERS Dynamic Relief 2 DVSL
DIERS Dynamic Relief 3DVSL
Fire Induced H2O2 Runaway Reaction
64
Case 6.01
Liquid Relief Sizing (API 520 Part 1 section 3.8.2 page 53)
Set Pressure
(psig)
250
Specific
Gravity
0.9
Flow
gpm
1800
A =
Q
38 K d K w K c K v
G
p1 p 2
Where:
p1
p2
Kd
Kw
Kc
Kv
G
65
Case 6.14
Flare Header Design
Flare headers are specialized piping networks designed to convey relief device flows to a flare
where products are consumed before release to the atmosphere. The design and evaluation of flare
header networks, for many relieving scenarios, can be studied.
Figure 6-1 shows a simplified flowsheet for a relief header network. The flows and pressures from
relief devices RD0001-RD0008 are previously calculated. The piping network has been designed
but not certified or built. The pipes on this flow diagram represent continuous sections of pipe with
various fittings (elbows, etc).
RD0 001
E -1
2"
R D 0002
3"
4"
R D 0003
P -2
3"
E -7
6"
E -8
R D 0 004
2"
R D 0005
3"
E -9
4"
E-1 0
RD00 06
R D 0007
3"
6"
2"
4"
6"
8"
E - 11
To Fla re K / O p o t
E-1 2
R D0 008
3"
Figure 6 -1
E - 13
Problem Statement
Calculate pressures throughout the network when RD0001, RD0002, RD0007, and RD0008 are
in a relief scenario
Determine if an additional section of 8 inch pipe can be added to the discharge
Determine effects of a two phase flow scenario through RD0007
Calculate pressures throughout the network if only three segments of the network have a relief
event
These tasks may be solved using the NODE UnitOp and Piping Network. The problem will be
solved one section of piping at a time and the best approach is to divide into smaller problems.
Using results from smaller cases, as estimations, a complex network can be gradually built,
reducing calculation time.
66
One Branch
Begin by drawing one segment of the network as shown in Figure 6-2. The intermediate pressures
are unknown/variable and will be solved for. The flowrate and composition from RD0002 is known.
Specify either the discharge pressure to the system or the entrance pressure at the first node.
Arbitrarily specify the pressure at the first node, making the outlet pressure variable. When the
entire network is drawn the outlet pressure will be specified and the inlet pressure recalculated.
Figure 6 -2
Three Branch
Figure 6-3 shows second and third sections of pipe added to the network representing the flow from
RD0001. A T branch fitting is added to the pipe UnitOp after the node connecting the two sections.
The flowrate from RD0001 is known. The pressure at the first node for RD0001 is variable.
Consider RD0001 section as a separate section of the pipe, there is no degree of freedom to
specify. The pressure at the node where RD0001 flow combines with the RD0002 flow is a
dependent variable for the RD0001 flow; it is calculated by the RD0002 section.
67
Section 7
Heat Exchanger Design and Rating
Contents
1.0
2.0
2.1
2.2
3.0
Introduction
Fundamentals
Basic Theory
Heat Transfer Model Selection
Design Guidelines
Appendices
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
XII
XIII
XIV
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
68
1. 0
Introduction
Shell and tube heat exchangers are used extensively throughout the process industry and as such
a basic understanding of their design, construction and performance is important to the practising
engineer.
The objective of this section is to provide a concise review of the key issues involved in the thermal
design without having to refer to the extensive literature available on this topic.
The optimum thermal design of a shell and tube heat exchanger involves the consideration of many
interacting design parameters, which can be summarised as follows:
Process
Mechanical
(9)
There are several design and rating packages available, including Aspen BJAC, HTFS and
(6)
CCTHERM , which enable the designer to study the effects of the many interacting design
parameters and achieve an optimum thermal design.
It must be stressed that software convergence and optimisation routines will not necessarily
achieve a practical and economic design without the designer forcing parameters in an intuitive
way. It is recommended that the design be checked by an alternative short-cut method and by
running the model in the rating mode.
Detailed mechanical design and construction involving tube sheet layouts, thicknesses, clearances,
tube supports and thermal expansion are not considered but the thermal design must be consistent
with the practical requirements.
Source references are not indicated in the main text as this paper should be considered as a
general guidance note for common applications and is not intended to cover specialist or critical
applications. Sources for this paper have been acknowledged where possible.
The symbols, where appropriate, are defined in the main text. The equations presented require the
use of a consistent set of units unless stated otherwise.
69
2. 0
2. 1
Fundamentals
Basic Theory
The basic layout for a countercurrent shell and tube heat exchanger is shown:-
The fundamental equations for heat transfer across a surface are given by:
Q = U A T lm = w C p ( t ) (t 2 t 1 ) = W Cp ( s ) (T1 T 2 )
or
W L
Where:
Q
U
A
t lm
Cp(t)
Cp(s)
w
(kJ/h, Btu/h)
2 o
2
(kJ/h-m C, Btu/h-ft -F)
2
2
(m , ft )
o
( C, F)
(kJ/kg-K, Btu/lb-F)
(kJ/kg-K, Btu/lb-F)
(kg/h, lb/h)
The log mean temperature difference(LMTD) Tlm for countercurrent flow is given by:
Tlm =
Where:
T1
T2
t1
t2
In design, a correction factor is applied to the LMTD to determine the true temperature difference
which allows for the departure from true countercurrent flow.
Tcorrected = FtTlm
The correction factor is a function of the fluid temperatures and the number of tube and shell passes
and is correlated as a function of two dimensionless temperature ratios:
(t 2 t1)
S =(
T1 t1)
(T1 T2 )
R = ( )
t 2 t1
Temperature correction factor charts for all arrangements are available in TEMA
standards. The heat exchanger configuration correction factor should be >0.75. The correction
factor Ft for a 1-2 heat exchanger which has 1 shell pass and 2 or more even number of tube
passes can be determined from the chart in the Appendix VII and can be calculated from the
following equation:
70
Section 7
Heat Exchanger Simulation Cases
This section reviews the sizing methods for shell and tube heat exchanger design using CCTHERM. It must be emphasised that convergence and optimisation routines will not necessarily
achieve a practical and economic design without the designer forcing parameters in an intuitive
way. It is recommended that the design is supported by short-cut design methods and finally
checked/optimized by running the model in the rating mode. The optimum thermal design of a shell
and tube heat exchanger involves the consideration of many interacting design parameters which
can be summarised as follows:
Process
Process fluid assignments to shell side or tube side.
Selection of stream temperature specifications.
Setting shell side and tube side pressure drop design limits.
Setting shell side and tube side velocity limits.
Selection of heat transfer models and fouling coefficients for shell side and tube side.
Mechanical
Selection of heat exchanger TEMA layout and number of passes.
Specification of tube parameters - size, layout, pitch and material.
Setting upper and lower design limits on tube length.
Specification of shell side parameters materials, baffle cut, baffle spacing and clearances.
Setting upper and lower design limits on shell diameter, baffle cut and baffle spacing.
For process design using a simulation program the following preliminary conservative estimates are
given for pressure drops due to friction. An additional pressure change occurs if the exchanger is
placed vertically.
Initial Process Design Pressure Drop Estimates
Process Description
Pressure Drop (psi)
Liquid streams with no phase change
10
Vapor streams with no phase change
2
Condensing streams
2
Boiling streams
1
Pressure (kPa)
70
14
14
7
Description
Single Component Horizontal Condenser
Horizontal Vacuum Condenser
Multiple Component Horizontal Condenser
Liquid Liquid Heat Exchanger
Heat Exchanger Temperature Control
Heat Exchanger 3 way Valve Control
Shell and Plate Heat Exchanger
71
Section 8
Absorption and Stripping
1.0
2.0
2.1
2.2
2.3
2.4
3.0
3.1
3.2
3.3
4.0
4.1
4.2
4.3
4.4
4.5
5.0
5.1
I
II
III
IV
V
Contents
Introduction
Fundamentals
Molality, Molarity, Mole Fraction and Mole Ratio
3
ppm(w), ppm(v), mg/m , Mole Fraction and Weight Fraction
Mass Balance
Simplified Mass Transfer Theory
Thermodynamics
Physical Property Predictions
K Models
Vapour Liquid Equilibrium (VLE)
Gas Absorption
SCDS Column in CHEMCAD
Case 8.01 CO2 Absorption in 30% MEA-Water Solution Regular VLE Model
Case 8.01 CO2 Absorption in 30% MEA-Water Solution Mass Transfer Model
Case 8.02 CO2 Absorption in Selexol
Case 8.03 HCl Absorption in Water Mass Transfer Model
Liquid Stripping
Case 8.04 Air Stripping of Trichloroethylene
Appendices
Fast Chemical Reaction Systems
Design Note on Enhancement Factor
Column Simulation Convergence
Nomenclature
Guidance Notes Bromine, Chlorine, Ammonia, Carbon Dioxide, Hydrogen Sulphide,
Sulfur Dioxide, Sulfur Trioxide, Hydrogen Cyanide and Nitrogen Oxides
References
rd
1. R.E.Treybal, Mass-Transfer Operations, McGraw-Hill, 3 Edition,1984.
2. I. Smallwood, Solvent Recovery Handbook, Edward Arnold,1993.
3. J.R. Backhurst, J.H. Harker and J.E. Porter, Problems in Heat and Mass Transfer, Edward
Arnold,1985.
nd
4. R.M.Felder and R.W. Rousseau, Elementary Principles of Chemical Processes, Wiley, 2
Edition,1986.
th
5. R.H.Perry and D.Green, Chemical Engineers Handbook, 7 Edition, McGraw-Hill,1984.
6. J.E.Edwards, Process Modelling Selection of Thermodynamic Methods, P&I Design Ltd,
MNL031B, August 2008. (Now covered in Section 4).
7. G.A. Morris and J.Jackson, Absorption Towers, Butterworths Scientific Publications.
(The British Library Reference WQ6 7604)
8. W.S.Norman, Absorption, Distillation and Cooling Towers, Longmans, 1962.
(British Library Reference W36 7183)
nd
9. C.J.King, Separation Processes, McGraw-Hill, 2 Edition,1980.
10. R.Billet and M.Schultes, Pre-calculation of Mass Transfer Columns with Dumped and Arranged
nd
Packings, 2 European CHEMCAD Symposium, Berlin, May 1999.
11. R.Sander, Compilation of Henrys Law Constants for Inorganic and Organic Species of
Potential Importance in Environmental Chemistry, Max-Planck Institute, Version 3, April 1999.
th
12. A.Kohl and R.Nielsen, Gas Purification, Gulf Publishing, 5 Edition, 1997.
13. K.Y.Li and K.J.Hslao, VOC strippers: tray design, Hydrocarbon Processing, February 1990.
14 H.Kister, Distillation Design, McGraw-Hill, ISBN 0-07-034909-6.
Simulation Cases
In this section these cases are presented and are provided as downloads from electronic media.
Case/File Name
8.01
8.02
8.03
8.04
Description
CO2 Absorption VLE and Mass Transfer Models
CO2 Absorption in Selexol
HCl Absorption in Water
Air Stripping of TCE
72
1.0
Introduction
This section reviews the simulation of gas absorption and liquid stripping processes using
CHEMCAD software. It discusses thermodynamics and the basic chemical engineering
relationships for vapor liquid equilibrium(VLE), mass and energy balances.
Methods are discussed for interpreting vapor liquid equilibrium data, design procedures and
determining satisfactory operating condition. Simulations are developed for gas absorption, with
and without chemical reaction, and liquid stripping processes. Particular emphasis is given to the
absorption and desorption of CO2 in alkanolamines due to its importance in carbon capture
technology.
Gas absorption is used extensively throughout the process industries to achieve a variety of
process objectives:
Liquid stripping is used extensively throughout the process industries to achieve a variety of
process objectives:
Gas absorption and liquid stripping can be difficult processes to simulate because:
CHEMCAD has two basic vapor liquid equilibrium or mass transfer models:
SCDS
Simultaneous correction distillation with rigorous VLE calculations which is particularly suited to
activity coefficient models. Applicable for steady state or dynamic models and reactive distillation.
TOWER
Rigorous VLE calculations, which is particularly suited to equation of state models. Uses the InsideOut algorithm. Applicable for steady state or dynamic models.
In addition, special modules are available which are used to maintain mass balances and control
feed, recycle and make-up streams.
1.
2.
3.
4.
73
Section 9
Continuous Distillation
Contents
1.0
Introduction
2.0
2.1
2.2
2.3
2.4
2.5
2.6
Fundamentals
Mass Balance
Separation Criteria
Theoretical Stages
Reflux Ratio
Heat Duty
Pressure
3.0
3.1
3.2
3.3
Thermodynamics
Activity Coefficients
Thermodynamic Model Selection
Column Convergence
4.0
4.1
4.2
4.3
4.4
5.0
6.0
References
1. F.G. Shinskey, Process Control Systems, McGraw-Hill,1967.
2. Smallwood, Solvent Recovery Handbook, Edward Arnold,1993.
3. J.R. Backhurst, J.H. Harker and J.E. Porter, Problems in Heat and Mass Transfer, pages 190
197, Edward Arnold, 1985.
nd
4. R.M.Felder and R.W. Rousseau, Elementary Principles of Chemical Processes, Wiley, 2
Edition,1986.
th
5. R.H.Perry and D.Green, Chemical Engineers Handbook, 7 Edition,13-96 to13-103, McGrawHill,1984.
6. Scott Barnicki, Put Your Column on the Map, Chemical Processing, Article 2004/208.
7. R.E.Rooks, Draw Insights on Distillation, Chemical Processing, Article 2006/071, Dow.
8. E.C.Carlson, Dont Gamble with Physical Properties for Simulation, Chemical Engineering
Progress, October,1996.
th
9. R.H. Perry and D. Green, Chemical Engineers Handbook, 6 Edition,13-82 to13-90, McGrawHill,1984.
Simulation Cases
The following cases are presented in this section. For practice you can build the models or use the
models provided as downloads from electronic media available.
Case/File Name
9.01
9.02
9.03
9.04a
9.04b
Description
Minimum Boiling Point Azeotrope Finder
Maximum Boiling Point Azeotrope Finder
EtOH Water Distillation
Heterogeneous Azeotrope Distillation
Heterogeneous Azeotrope Distillation
74
1.0
Introduction
This section reviews the simulation of distillation processes. It discusses thermodynamic model
selection criteria and the basic chemical engineering relationships for vapor liquid equilibrium, mass
and energy balances.
Methods are discussed for identifying azeotropes and to determine the feasibility of a distillation for
a given application. Simulations are developed for simple distillation and the more complex
heterogeneous azeotrope distillation.
Distillation is used extensively throughout the process industries to achieve a variety of processing
objectives:
Distillation is probably the most difficult process to simulate and control because:
SHOR
SCDS
TOWER
Rigorous VLE calculations, particularly suited to equation of state models. Uses the
Inside-Out algorithm Applicable for steady state or dynamic models.
TOWER+
In addition, special modules are available which are used to maintain mass balances, satisfy
reaction stoichiometry and control feed, recycle and make-up streams. These include:
Component Separator enables component fractions to be split from a stream.
Controllers can force a Stream to copy / scale another Stream flow and composition.
Stream Reference can fix a recycle flow at a constant value.
RAMP controller can change stream parameters and switch streams on or off on a time basis.
75
2.0
2.1
Fundamentals
(1)
Mass Balance
At steady state the overall material balance must be satisfied for each component. Only one
product stream can be set independently making the other product stream a dependent variable. In
general, distillate flow is manipulated, directly or through varying reflux ratio, requiring the bottom
outlet flow to be adjusted to maintain the material balance, which is achieved by holding the reboiler
liquid level constant.
The material balance based on the more volatile component (MVC) yields:
D z x
=
F y x
This shows that the flow ratio D/F determines the relative composition of each product such that if
feed composition z changes then D/F must be manipulated to maintain constant x and y.
2.2
Separation Criteria
The separation factor S for a binary mixture based on the MVC is defined as:
y 1 x
S x
=
y
S =
x =
+ x S 1
1
y
x
1
y
y + S (1 y )
The relative volatility of an ideal binary system, consisting of components 1(MVC) and 2, is the
ratio of the pure component partial pressures at the same temperature.
0
p1
p2
can be obtained by using the Antoine equation or the simplified Cox equation.
Antoine Equation (mm Hg,K,ln)
Cox Equation (mmHg,K,log10)
ln p = A
C +T
Bc
log10 (p ) = Ac
+
T 230
Antoine Coefficients can be found in CHEMCAD: Thermodynamics > Other Data >Parametric Data.
To convert from lnlog10 for the Cox equation Ac=(A/2.303), Bc=(B/2.303), Cc=(273+C/2.303).
76
Section 10
Batch Distillation
Contents
1.0
Introduction
2.0
2.1
2.2
Process Characteristics
Thermodynamics
Mass and Energy Balances
3.0
3.1
3.2
3.3
3.4
Process Modelling
Parameter Specification
General Considerations
Column Testing Model
Simple Batch Distillation
Appendices
I
Nomenclature
References
1.
2.
3.
4.
5.
th
R.H. Perry and D. Green, Chemical Engineers Handbook, 6 Edition, 13-82 to 13-90,
McGraw-Hill,1984.
I. Smallwood, Solvent Recovery Handbook, Edward Arnold,1993.
F.G. Shinskey, Process Control Systems, pages 484 489, McGraw-Hill,1988.
U.M. Diwekar, Batch Distillation,Taylor & Francis,1996.
J.R. Backhurst, J.H. Harker and J.E. Porter, Problems in Heat and Mass Transfer,
pages 190-197, Edward Arnold,1985.
Simulation Cases
Case/File Name
10.01
10.02
10.03
Description
Batch Distillation No Heat Transfer
Batch Distillation with Dynamic Reactor and Column
Batch Distillation with Control System
77
1.0
Introduction
Batch distillation techniques are used extensively in the Fine Chemicals and Pharmaceutical
Industries to achieve a variety of processing objectives. These include:
The simplest form of batch distillation involves a single separation stage (represented by the act of
evaporation) which is used where a large difference in volatility exists between the components to
be separated. Such a distillation does not need a fractionating column between the batch still,
normally a stirred jacketed reactor, and the condenser. Simple batch distillation, sometimes refered
to as pot to pot, provides only one theoretical plate of separation.
When the difference in volatility between the components to be separated is small or when
operating over narrow composition ranges a rectification section is required between the still and
the condenser. Overhead facilities are also required to provide control of reflux ratio and layer
separation when handling heterogeneous azeotropes.
To establish the number of theoretical stages N required at total reflux to achieve a specified
separation of a binary mixture with near ideal behaviour the Fenske equation is used:
.
Nmin ln = ln F
Separation factor
XD 1 XS
F =
1 X D XS
Where X is the mole fraction of the more volatile component and D and S represent distillate and
still compositions respectively and is the relative volatility of the two components.
These relationships can be used to form the basis of establishing performance guarantee criteria or
to establish a performance benchmark based on a given test mixture.
78
In operation the system is brought to steady state under total reflux, which is taken as the start
condition for simulation. Overheads are continuously withdrawn in accordance with the reflux
control strategy. Cuts are taken by switching to different receivers, following a time or overhead
composition strategy, at which point simulation conditions may be changed whilst starting from the
current condition.
Batch distillation with rectification can be operated as follows:
The distillation is normally operated at atmospheric pressure. However reduced pressure operation
is sometimes required to achieve the desired separation, to reduce operating temperature for
economic reasons, or when processing temperature sensitive materials.
For multipurpose operation careful consideration is required when selecting column internals to
achieve acceptable column loadings and operational turndown.
The boil up rate achievable with stirred jacketed reactors is dependent upon many factors
associated with heat transfer and is discussed in Section 11.
There are several theoretical operating regimes for a distillation column:
Continuous non steady state distillation, where control system tries to maintain operating
parameters near to set point
Start up of a continuous column
Semi batch or batch distillation
Engineers can find theoretical solution for a continuous column using steady state simulation, which
is also the first step to determine initial state. The theoretical solution of semi-batch or batch
distillation can be investigated using CC-Batch.
To study the behaviour of real distillation columns, regardless of its regime, CC-Dynamic Column,
DCOL, is used.
79
Section 11
Batch Reactors
Contents
1.0
Introduction
2.0
2.1
2.2
2.3
3.0
3.1
3.2
3.3
3.4
4.0
4.1
4.2
4.3
4.3.1
4.3.2
4.3.3
Performance Characteristics
Jacket / Coil Services
Temperature Control
Jacket / Coil Services Configurations
Direct Jacket Heat / Direct Cool
Indirect Jacket Heat / Direct Heat
Indirect Jacket Heat / Indirect Cool
Appendices
I
II
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Description
Reactor Indirect Heating Direct Cooling
Reactor Indirect Heating Indirect Cooling
Reactor Jacket Circulation Hydraulics
Cryogenic Reactor System
Batch Reactor Kinetics
Storage Tank Heating
80
1.0
Introduction
Batch processes are used extensively in the manufacture of relatively small volume products with
relatively high value. These processes are frequently carried out in production facilities intended for
multi-purpose use.
The achievement of stable and reproducible operating conditions is important in order to achieve
the required product purity, yield and cycle times to satisfy the commercial requirements and
relevant regulatory authorities.
Batch processes are inherently transient in nature and the capability to demonstrate dynamically
the adequacy of the equipment design and performance provides a powerful design tool. Dynamic
modelling can prevent costly mistakes prior to start up. Once a process model signature has been
validated against real plant performance the dynamic model can be used as a diagnostic tool to
identify operating problems.
This Section reviews the basic characteristics of batch reactors and the techniques for dynamic
modelling the process and control systems.
81
2.0
2.1
Heat Transfer
A typical batch reactor system is shown which defines the key variables for determining the basic
thermal ratings.
The fundamental equation for heat transfer across a surface area A is given by:
Q = U A Tm = V j Cpj (t 2 t1 )
The mean temperature difference Tm for a batch reactor with contents temperature T and
services inlet and outlet temperatures t1 and t2 is approximated by:
(t1 + t 2)
Tm = T
2
The overall heat transfer coefficient U is the sum of the individual resistances as follows:
1
1 1
1
1 1
+
+
+
=
+
U hi hfi k w x ho hfo
When considering glass lined equipment the reactor wall thermal conductivity kw includes the glass
thickness xg and metal wall thickness xm with the thermal conductivities for glass kg and metal km to
give the following:
kw =
xm + xg
xg k g + xm k m
The inside fouling coefficient hfi and outside fouling coefficient hfo are determined by practical
experience and there is extensive literature available on this subject. The combined fouling
coefficient hf is given by:
hf =
hfi hfo
hfi + hfo
82
Refer 4.3.2
Reactor jacket temperature control systems require a constant recirculation flow rate at all times.
This requires that the simulation maintains a mass balance in the incompressible fluid circulation
loop by removing the same flow from the loop as is admitted. This is achieved by setting the Divider
7 Stream 11 output equal to Control Valve 3 calculated flow rate. This valve controls the -40C
Syltherm XLT coolant flow to the jacket. Jacket mixing nozzle pressure drop is not calculated
rigorously. To achieve this the DataMap facility is used to calculate the pressure drop and set the
jacket outlet pressure. See Case 11.03 for the correlations and further details.
The typical control system uses a reactor contents temperature on reactor jacket temperature
cascade control loop with the heating and cooling valves operating in split range. The heating heat
exchanger has a constant steam supply pressure of 10 bar to ensure that condensate return is
maintained under all load conditions. Temperature control uses a 3 way bypass valve to eliminate
the thermal lags associated with this type of system see Case 7.05a.
83
Refer 4.3.3
Section 12
Environmental Emission Control
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
Contents
Atmospheric Dispersion from Stack
Fast Chemical Reaction Scrubbing
Stack Height Determination
Atmospheric Dispersion from Pools
VOC Emissions from Process Vessels
Blowdown Drum Sizing
Carbon Adsorber Column
Fugitive Emission Auditing
References
84
1.0
This model allows the prediction of the short term pollutant concentrations directly downwind of an
elevated point source, at specified distances and as a function of atmospheric stability and wind
speed. The model calculates the downstream distance at which the maximum ground level
3
concentration (mg/m ) occurs for the different stability classes over a distance range from 80m to
10km. The model is based on the following assumptions:
1.
2.
3.
4.
Concentration profiles in the plume follow the Gaussian distribution in the y and z directions.
Constant mean wind speed and direction.
Dispersion in the x direction is negligible compared with bulk transport by the mean wind speed.
The pollutant is a stable gas or aerosol which does not chemically react or settle out.
metres
3
m /sec
kg/sec
C
C
10
12.5
12.5
12.5
15
15
m/s
m/s
m/s
m/s
m/s
m/s
The atmospheric stability class is divided into six classes and is a measure of vertical mixing in the
atmosphere by convection.
Class
Stability
Description
A
B
Most unstable
Unstable
Slightly unstable
Neutral
E
F
Stable
Most Stable
85
Section 13
Hydrocarbon Processing
Contents
1.0
2.0
2.1
2.2
Thermodynamics
Global K and H Models
True Boiling Point Curves
3.0
3.1
3.2
3.3
Crude Distillation
Process Description
Simulation Considerations
Crude Unit Simulation Case 13.01
4.0
5.0
5.1
5.2
5.3
Gas Sweetening
Process Description
Simulation Considerations
Sour Gas Amine Treatment Case 13.03
Appendices
I
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
th
The cases considered in this section are listed below. For practice, you can build the models or use
the models provided as downloads from electronic media available. It is strongly recommended that
you work with a copy of the jobs.
Case/File Name
13.01
13.02
13.03
Description
Crude Unit Feed
Vacuum Unit
Sour Gas Amine Treatment
86
1.0
(7)
Most refinery products are mixtures separated on the basis of boiling point ranges. The block
diagram, by API, shows overall relationship between the refining processes and refined products.
This section considers the simulation of crude distillation at atmospheric pressure, vacuum
distillation and sour gas amine treatment. A basic introduction to process modelling of distillation,
absorption and stripping processes has been covered in Sections 8 and 9.
The applicable thermodynamics are reviewed with special reference to the generation and
application of pseudocomponent curves.
Crude petroleum consists of thousands of chemical species. The main species are hydrocarbons
but there are significant amounts of compounds containing sulphur (0-6%), oxygen (0-3.5%) and
nitrogen (0-0.6%). The main groups are:
Aliphatics or open chain hydrocarbons
n-paraffin series or alkanes (CnH2n+2)
This series have the highest concentration of isomers in any carbon number range but only occupy
20-25% of that range and make low octane gasoline. Most straight run (distilled directly from the
crude) gasolines are predominately n-paraffins. The light ends primarily consist of propane(C3H8),
n-butane (C4H10) together with water which are defined as pure components.
iso-paraffin series or iso-alkanes (CnH2n+2)
i-butane(C4H10) is present in the light ends but these compounds are mainly formed by catalytic
reforming, alkylation or polymerization.
Olefin or alkene series (CnH2n)
This series is generally absent from crudes and are formed by cracking (making smaller molecules
from larger molecules). They tend to polymerize and oxidize making them useful in forming
ethylene, propylene and butylene.
87
Section 14
Carbon Capture and Storage
Contents
1.0
2.0
3.0
4.0
5.0
6.0
Introduction
Thermodynamics
I
II
III
IV
V
The author would like to acknowledge the support of Mott MacDonald, (Brighton, UK) in the
preparation of this section.
References
1. IEA, Improvement in Power Generation with Post Combustion Capture of CO2, Report
PH4/33, November 2004.
th
2. R.C.Reid, J.M.Prausnitz, and B.E.Poling The Properties of Gases and Liquids, 4 Edition, Mc
Graw Hill.
3. U.S. EPA, Air Pollution Control Technology Fact Sheet, Spray Tower Scrubber, EPA-452/F03-016
4. National Lime Association, March 2007, Sargent & Lundy, FGD Technology Evaluation Dry
Lime vs Wet Limestone, Project 11311-001.
5. W.D.Seider, J.D.Seader, and D.R.Lewin Process Design Principles - Appendix X, 1999, John
Wiley & Sons, Inc.
6. The Babcock and Wilcox Company, August 1997, Wet FGD Forced Oxidation: A Review of
Influencing Factors and a Comparison of Lance and Sparge Grid Air Introduction Methods,
Report BR-1645.
7. The Babcock and Wilcox Company, August 1999, Advances in Fine Grinding and Mill System
Application in the FGD Industry, Report BR-1679.
th
8. R.H.Perry and D.W.Green, Chemical Engineers Handbook, 7 Edition, Mc Graw Hill.
9. D.L.McCollum and J.M.Ogden, Techno-Economic Models for CO2 Compression, Transport,
and Storage, 2006, UCD-ITS-RR-06-14, Institute of Transportation Studies, University of
California.
10. U.S. EPA, Control of Gaseous Emissions, January 2000, ICES Ltd Contract No. 68D99022.
11. Praxair Inc., Advanced Technology for CO2 Capture from Flue Gases, First National
Conference on Carbon Sequestration, May, 2001.
12. U.S. Department of Energy, A Technical, Economic and Environmental Assessment of Aminebased CO2 Capture Technology for Power Plant Greenhouse Gas Control, Contract No. DEFC26-00NT40935, October, 2002.
th
13. A.L.Kohl and R.B.Nielsen, Gas Purification, 5 Edition, 1997, Gulf Publishing Company.
th
14. H.M.Spiers, Technical Data on Fuel, 6 Edition, 1962.
15. J.W.Sweny and J.P.Valentine, Physical Solvent Stars in Gas Treatment, Chem. Eng.,
September 1970.
16. A.Henni, P.Tontiwachwuthikul, A.Chakma, Solubilities of CO2 in PEGE and Selexol, Canadian
Journal Chem. Eng, Vol. 83, April 2005.
17. J.Yannick, Physical Solvents for Selective CO2 Capture at Elevated Pressure and
Temperature, Chem. Eng. Dept., Pittsburgh University.
88
References
18. D.J.Kubek, E.Polla and F.P.Wilcher, Purification and Recovery Options for Gasification,
27. W.Breckenridge, A.Holiday, J.O.Y.Ong and C.Sharp, Use of SELEXOL Process in Coke
Gasification to Ammonia, L.Reid Gas Conditioning Conference, Oklahoma, March, 2000.
Links
L1
http://www.co2captureandstorage.info/
IEA data
L2
http://www.netl.doe.gov/technologies/index.html
DOE reports
L3
http://www.amines.com/mdea_comp.htm
Amine data
L4
http://www.ieagreen.org.uk/reports.html
IEA reports
L5
http://www.glymes.com/businesses/glymes/internet.nsf/directname/home
Clariant data
L6
http://www.co2captureandstorage.info/networks/capture_workshops.htm
Networks
L7
http://www.netl.doe.gov/publications/
DOE conferences
L8http://www.gasification.org/Docs/Workshops/2007/Indianapolis/03Ciferno%20Pre%20and%20Po
st%20Combustion%20CO2%20Capture%20Summary.pdf
NETL Data
L9
http://www.dow.com/gastreating/solution/index.htm
Dow data source
L10
http://www.co2storage.org.uk/
Storage issues
L11
http://www.rite.or.jp/English/lab/chemical/090622results-e/csiro2009-e.pdf
CSRIO and RITE Symposium, May 2009, Amines for Post-combustion Carbon Capture
L12
http://www.dow.com/gastreating/solution/pa_ascor.htm
Simulation Cases
The following cases are presented in this section. For practice you can build the models or use the
models provided as downloads from electronic media available. It is strongly recommended that
you work with a copy of the jobs.
Case/File Name
14.01
14.02
14.03
14.04
14.05
14.06
Description
Pre-combustion Selexol Multiple Flash Regeneration
Pre-combustion Selexol Combined Flash and Thermal Regeneration
Pre-combustion Selexol Selective H2S and CO2 Removal
Post-combustion Flue Gas Desulfurization
Post-combustion Amine Absorption and Stripping
CO2 Compression and Transport
89
1.0
Introduction
(1, L 2)
The worldwide focus on clean power generation and carbon capture has increased the importance
of the associated technologies, which involve two distinct approaches, namely pre-combustion and
post-combustion carbon capture.
In pre-combustion CO2 capture, fuel is gasified by applying heat under pressure in the presence of
steam and air and/or oxygen to form synthetic gas(Syngas). CO2 is then captured from the Syngas,
before being mixed with air in a combustion turbine, resulting in the CO2 being relatively
concentrated and at a high pressure.
In post-combustion CO2 capture, mainly, pulverized coal is burnt in air to raise steam. CO2 is
exhausted in the flue gas at atmospheric pressure and concentrations of 10-15% v/v. This process
is more challenging due to the low pressure and dilute CO2 concentration resulting in a high volume
of gas having to be treated. Also trace impurities in the flue gas tend to reduce the effectiveness of
the CO2 absorbing processes and compressing the captured CO2 from atmospheric pressure to
pipeline pressure represents a large parasitic load.
Another post-combustion capture technology, oxy-combustion, involves combustion of the fuel with
near pure oxygen resulting in a flue gas stream of higher CO2 concentration. This technology
relates more to combustion, and is not discussed further in this paper.
When CO2 is captured, power station generating efficiency is significantly reduced. Therefore a
power cycle with a high thermal efficiency is essential to ensure an acceptable outcome. Power
cycle efficiency is continually being improved by increasing the steam temperature and pressure;
this development is limited by the availability of suitable alloys.
Historical Trend in Boiler Operation (Note: water PC = 221 bar, TC = 374C)
Pressure
Temperature
Efficiency
Operation
Year
bar
C
%
Subcritical
1960 - 1970
166
540
38 to 42
Supercritical
1970 - 1990
240
566
44
Ultra supercritical 1990 - present
310
600
49
Ultra supercritical (USC), once through utility (OTU) operation is considered to be the best option
for Clean Coal Technology. USC operation reduces the CO2 emissions and therefore improves
the economics of carbon capture and sequestration (CCS).
Pre-combustion and post-combustion CCS methods are similar, in that both require significant
power for blowers, pumps and compressors. Studies are indicating parasitic power in the range 15
to 25% with the technology, plant layout, pressure drop, compressor operation, transportation and
sequestration configurations each having a significant affect. Crucial to CCS economics is the
optimization of the heat integration circuits and the minimization of CO2 stripping steam to reduce
the impact on power plant parasitic power.
CO2 is present at much higher concentrations in Syngas than in post-combustion flue gas, so CO2
capture should be less expensive for pre-combustion than for post-combustion capture. However,
there are few gasification plants in full-scale operation, and capital costs are higher than for
conventional pulverized coal plants.
Process simulators are playing an increasingly important role in finding the optimum economic
solutions. CHEMCAD is being used in modeling pre-combustion and post-combustion processes
and has developed thermodynamics specifically for the pre-combustion and compression
technologies.
90
sequestered in a suitable reservoir. Proven technology includes the Selexol process that uses a
solvent, which is a mixture of dimethyl ethers of polyethylene glycol. Dow Chemical Company
retains the gas processing expertise, which is offered under license. Clariant GmbH offer an
Appendix II
equivalent solvent, available from their Genosorb range.
ATR H2 CO/H2
Generation
Natural
Gas
Shift
Reaction
Steam
Gas
Turbine
Steam
H2
CO2
CO2/H2
CO2
Compress
CO2
Stripper
Selexol
Absorbers
Steam
ATR reaction:
Shift reaction:
Power
Gasifier
& Heat
Recovery
Coal
Heavy
Oil
CO/H2
COS/H2
COS
Hydrolysis
& Shift
Steam
Gas
Turbine
Stea
CO2
/H /H S
Power
H2
CO2
CO2
Compress
CO2
Stripper
Steam
COS hydrolysis
91
Selexol
Absorbers
H2S
COS + H2O CO2 + H2S
Section 15
Emerging Markets
The process industry is continually evolving to encompass a diverse range of activities in what can
be referred to as emerging markets. Process simulation is proving to be an important element in the
development and optimization of the processes involved. The scope of this variety is shown in the
following chart.
Process simulators are being used extensively in all the above areas, with the exception of the
renewable energy sector.
The economic and scale up challenges in the Carbon Capture sector are already being studied
with, to date, only relatively small demonstration units being operated. However these units are
enabling simulations to be validated to provide greater confidence when scaling up to full scale
plants. Post-combustion carbon capture provides the biggest challenge because industrial
experience with CO2 absorption and desorption with amine based solvents is at elevated pressure
whereas this application is at around atmospheric pressure. The previous Section 14 has
considered some of these issues.
For biodiesel production there are three basic methods available, namely:
Each reaction has associated optimal operating parameters (T & P) and conversion, although much
of the available literature emphasizes the base catalyzed route because it is claimed to be the most
economical.
The reaction progresses in three reversible steps
Triglyceride reacts with the alcohol to form a diglyceride and a fatty acid ester,
Diglyceride reacts with the alcohol to form a monoglyceride and a fatty acid ester
Monoglyceride reacts with the alcohol to form glycerin and a fatty acid ester.
92
For example, if palm oil, with at least 9 different fatty acid groups, is used, there could potentially be
729 different triglycerides, 81 different diglycerides, and 9 different monoglycerides present.
Based on the assay of any particular starting oil/fat, there will be an associated assay of fatty acid
esters in the final biodiesel product. The type of alcohol used determines the type of esters formed.
A typical base-catalyzed process diagram is shown below. In this case, the preliminary processing
of Waste Oil is shown in dotted outline at the top left.
The initial biodiesel industry objective was to achieve the earliest on-line operation, using known
technologies, to meet market demand. However as competitive pressures have increased there has
been considerable incentive to achieve greater efficiency in yields and energy usage. Process
simulators are now being used extensively and are achieving real economic benefits.
Process simulators are being applied extensively in other areas such as energy from bio-mass, fuel
cells and hydrogen production by electrolysis of water. All of these technologies require significant
supporting process plant and services to be integrated with the particular application.
93
Section 15
Emerging Markets Simulation Cases
In this section cases are shown indicating the wide range of applications that can be simulated.
For practice you can build the models or use the models provided as downloads from electronic
media available. It is strongly recommended that you work with a copy of the jobs.
Case/File Name
15.01
15.02
15.03
15.04
15.05
Description
Base Catalyzed Biodiesel Process
Hexane Extraction Biodiesel Process
Hydrogen Oxygen Fuel Cell
Hydrogen Production by Electrolysis
Biodiesel Blender
94
Section 16
Utility Systems
Contents
1
2
3
4
5
5.1
6
Introduction
Steam Generation Case 16.01
Steam Turbine Case 16.02
Combined Heat and Power(CHP) and Steam Distribution Case 16.03
Psychrometry and Natural draft Cooling Tower Case 16.04a
Cooling Tower Case 16.04b
Refrigeration Plant Case 16.05
Appendices
I
II
II
The author would like to acknowledge the contribution of Dr R.Nicholson of Thermal Developments
Ltd.(UK) and P.B. Baines of Tekna Ltd.(UK) in the preparation of this section.
References
1
J.Broughton, Process Utility Systems, I.Chem.E., 1994.
2
Spirax Sarco, www.spiraxsarco.com/resources
3
Combined Heat and Power Association, www.chpa.co.uk
4
Dr R.Nicholson, Universal CombCalc-VI-B, Thermal Developments Ltd, 2010.
5
J.L.Ryans and D.L.Roper, Vacuum System Design and Operation, McGraw Hill, 1986.
6
W.D.Seider, J.D.Seader and D.R.Lewin, Process Design Principles, Wiley 1998.
7
I.Smallwood, Solvent Recovery Handbook, Edward Arnold, 1993.
8
W.S.Norman, Absorption, Distillation and Cooling Towers, Wiley, 1961.
9
R.M.Felder and R.W.Rousseau, Elementary Principles of Chemical Processes, Wiley, 1986.
10 F.G.Shinskey, Process Control Systems, McGraw Hill, 1967.
11 H.Mller-Steinhagen, Handbook Heat Exchanger Fouling, Publico, 2000.
12 SCT, Solid Chemicals Technology, www.solidchemicals.com
Utility Systems Simulation Cases
For practice you can build the models or use the models provided as downloads from electronic
media available. It is strongly recommended that you work with a copy of the jobs.
Case/File Name
16.01
16.02
16.03
16.04a
16.04b
16.05
Description
Steam Generation
Steam Turbine
CHP and Steam Distribution
Psychrometry
Natural Draft Cooling Tower
Refrigeration Plant
95
1.0
Introduction
(1)
Utility systems are a key part of the process plant operating infrastructure providing heating,
cooling, compressed air, vacuum, water, inert gases and ancillary systems. A high level of reliability
is required for consistent production and process safety.
Utility site energy usage can be as high as 70% with up to ~15% being lost when the efficiency of
equipment and operations are taken into consideration. The application of combined heat and
power(chp), to produce electricity, can significantly improve the efficiency. Process simulation can
provide a powerful tool in the optimisation of utility systems.
Total integration of utilities with the production facilities will result in significant economic benefits.
The sizing of individual utility units needs to consider the turndown capability in relation to the
maximum and minimum demands of the processes, redundancy issues for maintenance and future
expansion. Typical turndown ratios are in the 3 to 5 range, with the efficiency reducing significantly
at low turndowns. The economics of running large units, sized based on maximum demand, as
opposed to operating smaller units in parallel requires careful consideration; utility units are
frequently oversized. For example, if a single process unit has the requirement for a small duty at
very low temperature, consider a local refrigeration unit, sized accordingly, rather than make the
central refrigeration unit operate at this low temperature.
Utility distribution networks need to be designed to ensure the supply is at a satisfactory pressure,
flow and temperature, for all process units under all operating conditions. There are few process
plants that do not experience bottlenecks due to poorly designed networks. Header pressure control
using pressure control loops at the supply to return connections and flow conservation devices at
the individual process units will minimise distribution problems.
The following summary is not comprehensive but highlights some issues that should be given
consideration in the design and operation of utility systems.
Heating Systems
Steam
(2, 7)
Steam generated at about 10 barg(dry saturated temperature 184C) is suitable for most process
applications including tank heating, steam ejectors, distillation and stripping. Process temperatures
up to ~310C would require steam pressures of 100 barg making hot oil the preferred option due to
the lower operating pressure. Steam generated at superheated conditions can be used for driving
steam turbines with the exhaust steam then passing to the process. If superheated steam pressure
is not reduced through steam turbines the steam should be desuperheated before passing to
process heating duties.
When correctly applied, steam provides the most efficient heating medium. Typically steam heat
transfer film coefficients, as compared to those for heat transfer fluids, can be greater by factors of
5 to 10. This can have a significant impact when considering boil up rates on distillation processes.
The following summarises some general application points:
1. Generate and use steam at the lowest pressure consistent with providing adequate T for
heat transfer.
2. Match steam generation to process usage to minimise or prevent venting to atmosphere.
Provide back pressure control on steam generators to prevent carry over of condensate into
the steam mains to avoid severe corrosion and stress corrosion cracking.
3. Pressure let down stations, using control valves or orifice plates, waste energy and can
create superheated steam which is inefficient for heat transfer. Note that the heat given up
by superheated steam down to dry saturation temperature, is small compared to the
enthalpy of evaporation released when at saturation.
96
4. Reducing pressure through steam turbines releases work which can be used for electricity
generation or pumping applications. Superheated steam is good for this.
5. Saturated steam is best for heat transfer provided that air is not allowed to accumulate.
Correctly located steam traps with air removal capability are required.
6. Typical steam line sizing velocities for saturated steam are 20-30 m/s(0-1barg), 30-40 m/s
(>3.5 barg), superheated steam(>13.5 barg) 45-60 m/s and hot condensate 0.5-1 m/s.
7. Condensate is highly corrosive and should be removed efficiently using appropriate steam
(3)
traps . Condensate build up in heat transfer equipment will reduce the area for heat
transfer. If steam flow to a unit is being manipulated to control temperature, condensate
removal should be by means of a float trap, essentially a level controller, to prevent control
hunting. Alternatively, steam traps based on the venturi orifice principle, have no moving
parts, provide smooth control, air removal capability and improved efficiency.
8. Condensate can become contaminated, as a result of leaks from process equipment,
requiring the location of the hot well to be considered carefully and the provision of
condition monitoring equipment.
Heat Transfer Fluids
Heat transfer fluids can be used in the temperature range -100C to 400C and provide the benefit,
when correctly applied of being single phase liquid systems. Applications are discussed in more
detail in Section 11.
1.
2.
3.
4.
5.
6.
Single phase systems avoid the requirement for automatic routing valves to connect
different services and provides seamless heat-cool changeover for temperature control
systems.
These systems eliminate the requirement for water treatment, reduces the corrosion
associated with water based systems and reduce the maintenance costs.
Heat transfer fluid specific heats are much lower than those for water systems. This
requires significantly higher flow rates for the same heat removal capability.
Heat transfer fluid viscosities can increase dramatically at low temperatures affecting heat
transfer capabilities.
The presence of water can result in steam formation at high temperatures and freezing at
low temperatures. Methods to remove water include running at high temperature and
venting off the steam at high points which can result in pump damage due to cavitation,
passing through molecular sieves and the use of an additive, such as a glycol, to capture
the water.
Contamination with water due to heat exchanger failure can result in serious outages. The
facility to isolate individual systems will allow the leakage source to be identified more
readily.
Cooling systems
1.
2.
3.
4.
5.
6.
7.
8.
Use the highest temperature coolant consistent with providing adequate T for heat
transfer; do not use refrigerated coolants when cooling water can achieve adequate cooling.
Optimise batch process cycle times by switching coolants when appropriate.
The variation in ambient conditions can affect cooling tower efficiencies resulting in issues
associated with vacuum plant.
On low temperature systems the heat gain from ambient and energy from pumps is
significant. Use high efficiency insulation and consider variable speed drives.
Protect refrigeration equipment from peak loads by using hot-cold side tanks for thermal
storage.
To conserve energy consider the use of variable speed drives on main circulation pumps to
match process demand.
Monitor the freezing points of water based refrigerants using ethylene glycol or propylene
glycol (preferred in the food and pharma industries) to check for dilution from other water
services.
Cooling water systems impose an onerous requirement for water treatment, monitoring for
microbial contamination and avoidance of legionella.
Careful consideration should be given to the sanitisation of water based systems.
97
Section 17
General Engineering Data
Contents
1
2
3
98
ULTRA-SONIC
POSITIVE
PHASE
LOW
3
1 21
DISPLACE
CORIOLIS
TARGET
350C,
0.1%
0.05%
(Saturated)
0.2-5cp
200
C,r1.0%
r
(Clean Liquid)
1
f (z ) =
(2)
Acknowledgements
To all colleagues and associates, who have provided support and advice, throughout my career in
the process control and chemical engineering fields. In particular I would like to give thanks and
acknowledgement to the following:
L.Lowe, R.Menzies, and D.Campbell when at UKAEA, Sellafield
Professor E.B.Chain, R.Falini and G.Banks when at Imperial College, London
R.Guimond, R.Johnson and G.Shinskey when at The Foxboro Company, Massachusetts
R.Fuhs, R.Frey, R.Halter and R. Hughes when at Rohm and Haas, Philadelphia and Teesside
Dr R.Nicholson of Thermal Developments Ltd., Sedgefield, UK for his comments on Section 7 and
his support on Section 16.
D.Ransome, P.Parkin, M.Morgan, D.Regan, M.Edwards and N.Al Qazzaz of P&I Design Ltd
P.Baines of Tekna Ltd.,UK for his contribution on Section 16
Special thanks to D.Harburn of DRH Associates for all the support in developing our environmental
software and for preparing the Scrubbing Guidance Notes in Section 8.
A special thanks to my colleagues at Chemstations, Houston, who have always given support in my
process simulation work and the preparation of the articles that make up this book:
N.Massey, Ming der Lu, S.Brown, D.Hill and A.Herrick and W.Schmidt, Germany.
Finally my wife Mary who has always supported me at the major cross roads of my career.
About the Author
John E.Edwards is the Process Simulation Specialist at P&I Design Ltd based in Teesside, UK.
He started working for the UKAEA, Sellafield in 1959 as a technician in the Process Equipment
Development Section. The main activity was in the development of mixer settlers for fission product
separation processes. He obtained a UKAEA sponsorship to study, on a sandwich course basis, for
a Chemical Engineering Degree at Glasgow University and graduated in 1963.
On graduation he worked at Imperial College, London in the Biochemistry Department under
Professor Chain. He was involved in the design, build, commissioning and operation of the
Fermentation Pilot Plant to produce metabolic products in sufficient quantity to allow for further
study. This involved submerged culture aerobic fermentation, solvent extraction, evaporation and
distillation processes.
In 1967, an interest in Instrumentation and Control resulted in a move to The Foxboro Company in
Massachusetts in their projects group. This involved applying the new technologies, evolving in the
1960s, namely the changeover from pneumatics to electronics, advanced process control, such as
feedforward developed by Shinskey at Foxboro, and Programmable Logic Controllers. He obtained
an MSc in Engineering Management, on a part time basis, at Northeastern University, Boston in
1971.
On returning to the UK in 1972 he worked for Rohm and Haas Ltd, initially as the Senior Instrument
Engineer and subsequently as Plant Manager of the Acrylates Production Unit. This was an
innovative process involving the catalytic oxidation of propylene to form acrylic acid and
subsequently acrylate esters using centrifugal extraction and vacuum distillation. Major production
problems were resolved by applying new measurement techniques, namely nucleonic level and
remote sealed transmitters.
In 1978 he formed P&I Design Ltd to provide a service in the Process and Instrumentation fields,
primarily to fill the skills gap between Chemical Engineers and Instrument Engineers. The main
activities have been process and control system design in the speciality chemical and
pharmaceutical fields leading to specialisation in process simulation, in 1998, using the
Chemstations software suite CHEMCAD and now presents courses worldwide on their behalf.
99