UNIVERSITY OF THE WEST INDIES

DEPARTMENT OF CHEMICAL ENGINEERING

CH32B Process Dynamics and Control II
Midterm Examination: Monday, April 4/05 (Chem. Eng. Design Office)
Answer all questions. Copy the files created in each question onto a floppy disk using
the suggested filenames. Label your disk with your name, student number and the
number of the computer you used. All written work you wish to have marked must be
written on the question paper.
The concentration of dilute liquors by evaporating solvent from the feed stream is an
important industrial process used in sugar mills, alumina production and petroleum
refining. Figure 1 is a conceptual representation of the first effect of the pilot-scale
evaporator system modelled by Newell and Fisher (1972). The solvent is water and the
solute is triethylene glycol.
PC1

Overhead
Vapour

V2

LC1
FC1

Feed
V1
FC2

Bottoms
Product

V3

V4

Steam

Figure 1 - Evaporator system.

Page 1 of 8

CC1

The liquid feedstream contains 3.2% triethylene glycol (TEG) by weight and is charged
to the evaporator at a rate of 5 ft3/hr. Its temperature and pressure are 190 F and 38.8
psia, respectively. The volume of the evaporator is 1 ft 3. At the initial steady-state, the
evaporator is half-full and operating at a temperature of 225 F.
Flow control valves V1-V3 are linear-trim and designed to be 50% open at the nominal
flowrates. The design pressure drop across V1 and V3 is 20 psi and that of the overhead
vapour valve V2 is 11 psi. The steam control valve V4 is manipulated by concentration
controller CC1 in order to regulate the mass fraction of TEG in the concentrated product.
When V4 is fully-open, the steam coil is capable of transferring 2 105 Btu/hr of heat
into the evaporator.
The bottoms product pump delivers a pressure rise of 40 psi with an efficiency of 75%.
The flow transmitters are ranged for 0 10 ft 3/hr and the pressure transmitter for 10 30
psia. The on-stream concentration analyzer is calibrated to detect TEG mass fractions in
the range 0 0.1 and exhibits first-order measurement dynamics with a time constant of 1
minute.
(a) State the action (reverse or direct) required for each of the controllers in Fig. 1.
Explain the reasoning behind your choice of action for FC2, LC1 and CC1.
[10 marks]

Page 2 of 8

(b) Construct a steady-state simulation of this process using HYSYS.Plant. Name the
file test1 (ss).hsc. When building this model,

Choose the UNIQUAC property package.

Specify the proportional gain K c and integral time I of the flow controllers as 0.7
%/% and 0.1 minute, respectively. Set the tuning constants of the pressure controller
to 2 %/% and 5 min and the parameters of concentration controller CC1 to 2 %/% and
10 min.

Use a value of K c in the LC which will ensure that the controller output equals
100% when the level rises to 55% and is zero when the level falls to 45%. (For the
purpose of computing this controller gain, the setpoint may be assumed constant at
the initial liquid level of 50%.) Show your work in the space provided below.

Page 3 of 8

Select the Signal Processing option in the FC1 and FC2 PID blocks and install PV
filters with time constant 5 seconds.

Model the analyzer dynamics using a Lag 1 transfer function with unity gain.

Configure a strip chart for each controller which will sample the PV, SP and OP every
six seconds and trend the preceding six hours of operation. Change the span of the
PV axes on the flow controller charts to 2 7 ft 3/hr. Change the span of the PV axis
on the concentration controller chart to 0.04 0.07.

Uncheck the Active box for Pressure Rise on the Dynamics tab of the pump and select
Power instead. Activate the dynamic pressure specifications for each boundary
stream on the flowsheet.

Left-click the Dynamics Assistant button on the toolbar to ensure that no problems
will be encountered when transferring your model to Dynamics Mode.
[45 marks]

(c) Switch to Dynamics Mode. Place FC1, PC1 and LC1 in Automatic and secondary
flow controller FC2 into Cascade mode. (If the level-to-flow cascade system has been
properly built, the letter R will appear on yellow background on the FC2 controller
faceplate.) Run the simulation for one minute and then save your file as test1 (dyn).hsc.
[5 marks]

Page 4 of 8

(d) Open test1 (dyn).hsc and save it as test1d.hsc. Simulate 2 hours of open-loop
operation (i.e. with CC1 on manual), introducing a 20% step increase in the controller
output at time t = 10 minutes. Fit a first-order-plus-deadtime model to the reaction curve
and describe the method used to estimate each parameter. The steady-state gain should
be expressed in dimensionless form, i.e. with engineering units %/%. Save your case file,
then close it.
[15 marks]

Page 5 of 8

(e) Open test1 (dyn).hsc and save it as test1e.hsc. Simulate 2 hours of open-loop
operation, introducing a 0.5 ft3/hr step increase in the setpoint of feed flow controller
FC1 at t = 10 minutes. Fit a first-order-plus-deadtime model to this reaction curve and
describe the method used to estimate each parameter. Express the steady-state gain in
dimensionless form. Save your case file, then close it.
[10 marks]

Page 6 of 8

(f)
Open test1 (dyn).hsc and save it as test1f.hsc.
Simulate 2 hours of
feedforward/feedback control of the bottoms product composition, introducing a 0.5 ft3/hr
step increase in the feed flowrate at time t = 10 minutes. The setpoint of the
concentration controller is to remain fixed at 0.0485 (mass fraction TEG). Explain the
method used to design your feedforward compensator. When the simulation is complete,
save your case file and close it.
[15 marks]

Page 7 of 8

END OF QUESTION PAPER

Page 8 of 8