Concentration Changes in A CSTR (Continuous Stirred Tank Reactor)

The University of Birmingham
School of Chemical Engineering

Process Systems (2PS)
Assessed Tutorial 2010/2011
(To be handed in by Thursday 20th January 2011, 15:30)
Concentration changes in a CSTR (continuous stirred

tank reactor)
Please create a folder in your personal area (U: drive) called 2PS_assess_tut_1011
and save any files you produce as part of the assessed tutorial in this area.
1 Objectives of the Tutorial

i.
ii.
iii.
iv.
v.
To develop a deviation variable model from a dynamic model of the

concentration within a CSTR, in which the concentration is measured
by a conductivity probe.
To solve the resulting differential equation analytically for two common
types of input change
To use these analytic solutions to generate data showing measured
concentration-vs-time response plots, and then to compare these plots
with simulated response plots generated using Simulink (this is the
system response without control the open loop response).
To investigate the behaviour of a feedback loop simulating the
controlled response of the concentration to both set point changes and
disturbances.
Based on this behaviour, to comment on the suitability of different
concentration controller designs for use in industrial CSTRs.
Within this tutorial there are a set of tasks that need to be completed. We
recommend that you use the worksheet provided on WebCT.
2 The system
The system to be investigated is a CSTR with two inlet streams and one outlet
stream.
F1, CA01
F2, CA02
F, CA
1
The reactor is well mixed and has a fluid volume of V (m3).

F1 is the volumetric flowrate of stream 1 (m3 s-1)
CA01 is the concentration of substance A in stream 1 (kmol m-3)
F2 is the volumetric flowrate of stream 2 (m3 s-1)
CA02 is the concentration of substance A in stream 2 (kmol m-3)
F is the volumetric flowrate of the outlet stream (m3 s-1)
CA is the concentration of A in the outlet stream (kmol m-3)
The CSTR is set up so that the concentration of A in stream 2 can be
controlled, the concentration in stream 1 can be viewed as a disturbance (see
Appendix A).
The reaction within the CSTR can be represented by the following 1st order
reaction equation:
rA = kC A
Eqn (2.1)
Task 2.1 (8% of the marks)

Using an unsteady state material balance show that the CSTR described
above can be represented by Eqn(2.2) below. What do p, Kp1 and Kp2
represent and how are they related to the reactor parameters listed above?
Make a note of any assumptions used to derive the equation.
dC A
+ C A = K p1C A 01 + K p 2C A02
dt
Eqn (2.2)
end of task 2.1

To measure the concentration a conductivity probe is used. The probe can be
modelled using a first order differential equation of the form:
dC Am
+ C Am = K m C A
dt
Eqn (2.3)
Where m (s) is the probe time constant, CAm (mA) is the probe signal (a
representation of the concentration), Km (mA (kmol m-3)-1) is the probe gain
and CA (kmol m-3) is the actual concentration.
The concentration CA needs to be controlled at the set point. In practice the
control system would look to manipulate both p and CA02 to achieve this. For
this tutorial however, p is considered constant and control is achieved
through manipulation of CA02.

Show that by combining Eqns (2.2) and (2.3) that the following 2nd order
differential equation can be derived.
d 2C Am
dC Am
+ 2
+ C Am = K m K p1C A01 + K m K p 2C A02
2
dt
dt
2
Eqn (2.4)
Identify the parameters and .

end of task 2.2
3 System data and further information

Initially the system is at a steady-state with the following parameters.
V = 5 m3
F1 = 0.04 m3 s-1
F2 = 0.01 m3 s-1
F = 0.05 m3 s-1
CA01 = 4 kmol m-3
CA02 = 1 kmol m-3
CA = 1.36 kmol m-3
p = 40 s
Kp1 = 0.32
Kp2 = 0.08
k = 0.015 s-1
The probe (sensor) has the following parameters
Km = 2 mA (kmol m-3)-1
m = 10 s
As part of the CSTR / probe system there is a feedback control system which
is able to adjust the value of CA02 by manipulating a valve, depending on what
the value of the set point CA,set is (note: without changing the value of F2). The
feedback loop contains:
An electronic controller with proportional (P-only) action, which has a
gain, KC (mA mA-1). Initially KC = 1 mA mA-1.
An electronic-pneumatic transducer (gain, Kep = 0.5 p.s.i mA-1) to
convert the controller signal into an equivalent pneumatic signal for the
control valve.
A pneumatic control valve (which manipulates CA02) with first order
dynamics. Model parameters KV = 2 (kmol m-3) (p.s.i)-1 and V = 25 s.
(see Appendix A for further explanation of this)
4 Further Tasks (You are advised to use the

worksheet provided)
4.1 Derivation of a deviation variable mathematical model for
the uncontrolled CSTR + probe system
Task 4.1 (a) (2% of the marks)
By using Eqn (2.4) and the following deviation variables
C A* = C A C As ;
C A* 01 = C A01 C A01s ; C A* 02 = C A02 C A02 s ;
*
C Am
= C Am C Ams ;
Show that
*
*
d 2C Am
dC Am
*
+ 2
+ C Am
= K m K p1C A* 01 + K m K p 2C A* 02
2
dt
dt
2
Eqn (4.1)
Task 4.1 (b) (23% of the marks)

Solve Eqn (4.1) analytically to obtain the response of C*Am with time, t for:
i.
a step change in the concentration of A in stream 2,
C*A02 = 2 kmol m-3.
ii.
a sinusoidal change in the concentration of A in stream 1,
C*A01 = 0.1sin t
30
In case (i) C*A01 = 0 and in case (ii) C*A02 = 0. In both cases all flowrates are
maintained at their steady state values.
end of task 4.1
4.2 Testing of the Mathematical Model

Construct the following open loop models of the CSTR and sensor (probe).
The values for the CSTR TF (transfer function) and Sensor TF are correct for
the values already given in 3. The values for the input depend on Task 4.1 b
(i) and (ii).
Model (i) parameters in 3
Compare the numeric solution for C*Am generated by Simulink to the

analytic solution you derived in Task 4.1 b (i). Run the simulation till
an end ('stop') time of 250 seconds. Comment on your results.
Comment on the differences between the values output to the
workspace for CA_star (C*A) and CAM_star (C*Am).
Model (ii) parameters in 3
Compare the numeric solution of C*Am generated by simulink to the

analytic solution you derived in Task 4.1 b (ii). Run the simulation till
an end time of 500 seconds. Comment on your results.
Comment on the differences between the values output to the
workspace for CA_star (C*A) and CAm_star (C*Am).
You may wish to force Simulink to use more time points for its numeric
solution (so you get smooth curves). To do this select <simulation>
then <Configuration Parameters>. Under Solver Options change the
max step size from auto to 1.
end of task 4.2
4.3 Further observations of the simulated open loop

behaviour
Task 4.3 (a) (12% of the marks) Investigating Model (i) from Task 4.2
For all of your simulations in this task use an end time of 1000 s, you may
wish to add additional scopes or 'To worksheet' blocks to your simulations.
Also from the Simulation menu choose Configuration Parameters and
change the Max Step Size in the Solver Options section from auto to 1.
Investigate the following step changes, and sensor time constants m.
Step changes:
Time constant:
C*A02 = 0.1, 2, 10 (kmol m-3)

m = 10, 60 (s)
(This represents 6 simulations in total).
Observe and record

(i)
The time taken to reach the final, settled value of C*Am and
C*A.
(ii)
Compare the shapes of plots of C*Am and C*A against time.
(iii)
Comment on the effect of m.
Task 4.3 (b) (13% of the marks) Investigating Model (ii) from Task 4.2
For the simulations in this task use an end time of 1000 secs. You may wish
to add additional scopes or 'To worksheet' blocks to your simulations.
( b ) ), and sensor time
Investigate the following sinusoids (C*A01 = a sin t

constants m.
Sinusoid changes: (i) a = 0.1, b = 30; (ii) a = 1, b = 30; (iii) a = 1, b = 60

Time constant:
m = 10, 60 (s)
(This represents 6 simulations in total).
Observe and record
(i)
The amplitude of C*Am and C*A after any initial transients
have decayed.
(ii)
(iii)
end of task 4.3
4.4 Simulations of the controlled behaviour

Task 4.4 (a) (12% of the marks)
Construct the following Simulink block diagram based on the control diagram
in Appendix A. You may wish to add additional scopes or 'To worksheet'
blocks to your simulations.
At a simulation time, t = 0, there is a step change in C*A,set = 2 kmol m-3 and a

disturbance in C*A01 = 0.1sin t
30
Run the simulation with KC = 1, 10, 50 till an end time, t = 1000 s:

(i)
M = 10 s
(ii)
M = 60 s
(again 6 simulations in total)
For all the simulations, observe, record and compare:
Has a steady-state been reached after 1000 s?
If a steady-state is reached how long did it take?
What are the values of CAm and CA at t = 1000 s?
Task 4.4 (b) (8% of the marks)
i.
Select two industrial CSTR applications and briefly state what
quality of control would be acceptable for them (e.g. must control
be at the set point? is some variation acceptable?). Give your
reasons.
ii.
Did any of the controller and sensor parameters in Task 4.4 (a)
give adequate control? If yes, state which ones and why, if no,
state which ones and why.
iii.
Describe briefly how you would look to improve the performance
of the control system.
end of task 4.4
P T Robbins November 2010
Appendix A
CA01
CA02
AC
CA
Figure A1: A possible way to control CA02
The following block diagram can be developed (which the Simulink model for
Task 4.4 is based on), it is assumed that the control of CA02 can be done
without changing the flowrate F2 and also that the flowrate F1 does not
change, hence if the outflow stays at F (= F1+F2) then the volume in the CSTR
remains at V. First it is noted that Eqns (2.2) and (2.3) can be re-written in
terms of deviation variables (given in Task 4.1 (a)) and then Laplace
transforms taken to give:
K p1 *
K p2 *
C (s) =
C
s +
C
s
s + 1 A 01 ( ) s + 1 A02 ( )
p
K
*
C Am
( s ) = m C A* ( s )
ms +1
*
A
Eqn (A1)
Eqn (A2)
The first task is to create the open-loop response (Eqn (A1))
C*A01
C*
A02
K p1
ps +1
K p2
C *A
ps +1
The value of C*A is then measured via the probe (Eqn (A2)) and this signal
compared to the set point, to give:
C*A01
C*A,set
Km
C*A02
K p1
ps +1
K p2
+
+
C*A
ps +1
C*Am
Km
ms +1
The error signal is then processed by the controller for this tutorial
proportional only control is considered (controller gain = KC), then converted
to a signal for the valve (Kep) and finally acted upon by the valve. The valve
response is not instantaneous so its response is modelled via a first order
ODE, this has the same form of transfer function as the probe. The final
control diagram looks like:
C*A01
C*A,set
Km
KC
Kep
KV
V s + 1
C*Am
C*A02
K p1
ps +1
K p2
+
+
C* A
ps +1
Km
ms +1
10

Concentration Changes in A CSTR (Continuous Stirred Tank Reactor)

Uploaded by

Copyright:

Available Formats

Concentration Changes in A CSTR (Continuous Stirred Tank Reactor)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Concentration Changes in A CSTR (Continuous Stirred Tank Reactor)

Uploaded by

Copyright:

Available Formats

What is the system being modeled?

What is the system being modeled?

What are the objectives of the tutorial?

What are the objectives of the tutorial?

The University of Birmingham

School of Chemical Engineering

Assessed Tutorial 2010/2011

(To be handed in by Thursday 20th January 2011, 15:30)

Concentration changes in a CSTR (continuous stirred

1 Objectives of the Tutorial

To develop a deviation variable model from a dynamic model of the

The reactor is well mixed and has a fluid volume of V (m3).

Task 2.1 (8% of the marks)

end of task 2.1

Task 2.2 (2% of the marks)

Identify the parameters and .

3 System data and further information

4 Further Tasks (You are advised to use the

C A* 01 = C A01 C A01s ; C A* 02 = C A02 C A02 s ;

Task 4.1 (b) (23% of the marks)

end of task 4.1

4.2 Testing of the Mathematical Model

Compare the numeric solution for C*Am generated by Simulink to the

Model (ii) parameters in 3

Compare the numeric solution of C*Am generated by simulink to the

end of task 4.2

4.3 Further observations of the simulated open loop

C*A02 = 0.1, 2, 10 (kmol m-3)

(This represents 6 simulations in total).

Observe and record

( b ) ), and sensor time

Investigate the following sinusoids (C*A01 = a sin t

Sinusoid changes: (i) a = 0.1, b = 30; (ii) a = 1, b = 30; (iii) a = 1, b = 60

4.4 Simulations of the controlled behaviour

At a simulation time, t = 0, there is a step change in C*A,set = 2 kmol m-3 and a

Run the simulation with KC = 1, 10, 50 till an end time, t = 1000 s:

P T Robbins November 2010

The first task is to create the open-loop response (Eqn (A1))

You might also like