Process Control Assignment Case Study - Reactor-Flash unit

In this assignment you will apply the principles of control to a specific case study-a reactor/flash unit polymerisation plant with recycle.  In the first assignment, we will look at dynamic responses and process instrumentation. In the second assignment, we will design and tune control loops for this process.

You have been hired as an external process control consultant to advise on a range of process control issues for a new reactor flash unit that has been installed at the local SMUCE Ltd plant. You need to prepare a written report with recommendations to the Chief Engineer of SMUCE Limited. 

Part A - Report 

Present your results as a brief report addressed to the SMUCE Ltd CEO. The SMUCE CEO likes easy to read reports where everything is clearly laid out and explained.  10 marks will be towards the report clarity and professionalism. The report should be standalone - that means that someone who has never worked on the SMUCE Ltd process should be able to read your report and understand what you are recommending for the control system. This means you should include your individual responses and transfer functions, including printouts of the responses, as well as answers to all the specific questions,

Section 1: Process response analysis:

Q1: Using your unique plant parameters, estimate the transfer function for each of the responses, due to the step changes as indicated. Clearly list gain, time constant etc for each response and include working. 

Q2: Draw a block diagram of the combined reactor+flash unit (see Figure 3) showing the transfer functions which describe how the flowrate of w4 and concentration of stream 4 (x4A) is affected by the flowrate of stream 1, stream 2 and stream 8. Show as many transfer functions as possible in your diagram.    

Section 2: Instrumentation recommendations:

Q3: Referring to Figure 2, what kind of process instrumentation (temperature, flow, pressure transmitters etc.) would you recommend for each of the following streams. Specify type of instrumentation (eg thermocouple or RTD) and range(s) as much as possible, stating any assumptions.

a) Reactor A feed w1

b) Reactor B feed w2

c) Reactor effluent w3

d) Flash unit liquid product w4

Q4: What kind of instrumentation would you recommend for the reactor and flash unit? Specify as much as possible.

Part B -

Simulink analysis-

You will now begin to simulate your process in SIMULINK, which is part of MATLAB, which allows you to draw block diagrams and simulate control loops without using MATLAB code. MATLAB and SIMULINK are available on the Engineering computers. Basic introductory guides are also on Moodle. Each student will have a unique set of values for Gp and Gd, and therefore each loop will have unique parameters.  

Get your loop parameters-

Draw a loop including a PID controller similar to Figure 4, which controls the production flowrate w₄. Using the information from part A, determine the transfer function for Gp and Gd for your plant.

Loop pairing-

QB1: Produce a 4x4 RGA analysis to determine the best pairing of manipulated (w₁, w₂, w₆, w₈) and controlled variables (w4, x_4A, HT, x_8D) to minimise interactions.

QB2: Sketch this control strategy you recommend above on the partial P&ID shown in Figure 2.    

QB3: Add the remaining loops on the P&ID for temperature(s), level(s) & pressure control as needed.  

Single Loop tuning-

QB4: Add a P controller to the loop, and determine (by trial and error) the controller gain that gives a stable response in w4 with the least amount of offset for a SP change.   Use the loop pairing from the previous section to determine what input to pair with w4.

QB5: Using the same controller setting as part (a), determine also the offset for a disturbance change.

QB6: Tune the P-only controller using the Ziegler Nichols tuning method (you will need to find ultimate gain Ku and ultimate period P_u). Check the response of the loop to a SP change.

QB7: Remove the P controller and replace it with a PI controller. Calculate the tuning settings based on the Ziegler Nichols method.  Compare the P and PI controller responses and make a recommendation to SMUCE Ltd on which you recommend based on a discussion of the responses (eg overshoot, decay ratio, offset, etc). 

QB8: Make a set point change in each loop, and prove that there is a disturbance in the opposite loop. 

QB9: Design an appropriate, realisable decouplers for the loops. 

QB10: Perform a Factory Acceptance test of the decouplers, by adding the decouplers you have designed to the system, and showing whether the interactions are reduced when the decouplers are in place, and discuss the response and your recommendations (single or double decouplers, static or dynamic?)

QB11: How close can you get to the control specifications for w4 that SMUCE Ltd requested in part 2.0? What other recommendations do you have to achieve this? Include this in your report and recommendations.

Attachment:- Assignment.rar