Specialized course Process Control at Norwegian University of Life Sciences (NMBU), Spring 2018


Compulsory exercise to lessons in Lecture 1

For the problems given below, send a short report presenting the results (plots and calculations), and the Matlab scripts and the Simulink model to the teacher in email no later than the deadline given in the lecture plan on the course homepage.

 

1.      Some feedback control basics: Open the simulator Temperature Control of Liquid Tank.

a.       Use the PI(D) controller. What happens to the stability of the control system if the process time-delay is relatively large?

b.      Use the PI(D) controller. Include random temperature measurement noise with max amplitude of 0.2 deg C. The D-term of the controller can be activated by increasing Td from 0 to one quarter of the integral time (Ti), which is the ratio between Td and Ti according to Ziegler and Nichols. Compare the behaviour of the control signal using PI controller and PID controller. With PID controller, select a filter time-constant so that you become content with the noise level in the control signal.

c.       Use the On/off controller. Is the mean value of the control error zero or non-zero?

2.      Simulation of transfer function: In Chapter 2 of the textbook Basic Dynamics and Control, a mathematical model of a mass-spring-damper system is presented.

a.       Derive the transfer function, H(s), from force F to position y.

b.      Try to replicate the responses shown in Figure 2.8 in the textbook by using the lsim function in Matlab.

c.       Implement a simulator of H(s) in Simulink, and try again to replicate the responses shown in Figure 2.8 in the textbook. The simulation should be run using the sim function in a Matlab script. Also in that script, the model parameters should be defined (the Simulink block diagram should contain no numerical values), and the responses should be plotted using the plot function (but you may also include Scopes in the block diagram).

3.      Simulation of state-space model represented as a block diagram: See Exercise 2.3 in the Basic Dynamics and Control exercise book.

a.       Implement a simulator of the system in Simulink using a fixed-step solver (e.g. the ode1 solver which implements Euler forward numerical integration). Use the MATLAB Function block to calculate the time-derivatives of the state variables, and use Integrator Limited blocks to integrate these time-derivatives (i.e. to calculate the levels). Set appropriate max and min levels (on the integrators). Select proper parameter values yourself. You can assume zero initial state (levels).The simulation should be run using the sim function in a Matlab script. Also in that script, the model parameters should be defined (the Simulink block diagram should contain no numerical values), and the responses should be plotted using the plot function (but you may also include Scopes in the block diagram).

b.      Run a simulation where u2 is kept constant, while u1 is changed as a step from zero to a proper nonzero value at some point of time larger than zero. Plot the responses in levels h1 and h2 in one plot and u1 in another plot (using the subplot command).

c.       Verify that the simulated levels are equal to the analytically calculated levels under steady-state (static) conditions.


Updated 22 December 2017 by Finn Aakre Haugen, course teacher.