Closed Loop Systems Analysis
Closed Loop Systems Analysis
S
GO
A
R
E
SI L
TY OF
UNIVERSITY OF LAGOS
Chemical Engineering Department
2018/2019 SESSION, 1st SEMESTER
Tutorial III:
Closed Loop Systems Analysis
[Block diagram representation, Closed loop transfer functions, Characteristic equations, Stability
Analysis – Routh test, and root locus diagrams]
Yd(s)
Question 1.2
Normoglycemia refers to the normal blood condition of glucose, which is within the range of 70
mg/100 ml blood to 110 mg/100 ml blood. When food is consumed by the human body, and broken
down by the digestive system, the blood glucose level rises. The hypothalamus (in the brain)
contains cells that are glucose sensors. These cells sense the increase in glucose levels and pass
the signal to the brain’s control center which in turn tells the pancreas to release insulin - the
hormone that maintains a steady level of glucose in the blood thus preventing diabetes.
Draw a simple block diagram for glucose control in the human body
Question2.1
For the block diagram shown below
d
gd
g c1 g1 + +
yd + +
y
g c2 g2 +
-
Page 2 of 7
Question 2.2
Obtain the closed loop transfer function for the block diagram below.
Yd Y
Question 3.1
A process has two poles and one zero. The poles are located at -1 ± 0.5i and the zero is
located at 0.5. Sketch the type of response that you expect to a step change in input. Explain.
Find the transfer function and verify these results assuming a gain of one.
Page 3 of 7
Question 3.2
Three tanks of different sizes are arranged in a non-interacting manner as shown in Figure 3.2,
where the liquid level in tank 3 is controlled by manipulating the inlet flow rate to the 1st tank,
Fo.
Assuming negligible dynamics for the measuring device and the control valve, that is g m = gv = 1 and
that the process transfer function is given by
1
g ( s) =
(3s + 1)(2s + 1)( s + 1)
Using SIMUKLINK, plot the closed loop response to a unit step change for the following conditions:
i The liquid level controller is a Proportional controller, with kc = 5 . Plot the closed loop system
response. On the same plot, show the effect of increasing the proportional gain from 2 to 12 in
increments of 2. Pls note that all plots should be on the same graph for comparison purposes.
Comment on the plots
ii The controller in (i) is replaced with a Proportional-Integral controller, with kc = 4.5
andτ I = 5.2 , plot the system response on a different graph. Show the effect of increasing
proportional gain from 2 to 12 in increments of 2. On another graph, show the effects of
increasing the integral time constant from 2 to 12 in increments of 2. Comment on the plots.
iii Now using a PID controller with the following parameters, kc = 6, τ I = 3.1, τ D = 0.775 ,
show on another plot the system response. Show the effect of increasing the derivative time
constant from 0.4 to 1.2 in increments of 0.2 on the same plot. Comment on the plots
iv Give a general comment on the use of P, PI and PID controllers on the liquid level system in
figure 3.2, and comment on the effect of increasing kc , τ I , and τ D .
Question 3.3
You are an engineer at an ice cream factory. There is a storage vat with a cooling system that
has a PI controller scheme. It has the following characteristic equation
10s 4 + 3s3 − kc s2 + 3s + (τ I + 6) = 0
You are to determine the constraints on the controller parameter values, kc and τI that will
ensure the closed loop system is stable.
Page 4 of 7
Question 3.4
For the system whose open loop transfer function is given by:
kc
gOL =
s(s + 2)(s + 4)
(i) Find the kc value that causes the system to be at the verge of instability.
(ii) Using the rlocus command in MATLAB, plot the root locus diagram.
(iii) Confirm (i) from (ii)
Question 3.5
The characteristic equation of a closed loop system is given by
2 0.5
1 + gc gv g = 0 , where g = ' gv =
(20s + 1) (s + 1)
The root locus plot of the system has loci that originate from four points, 0, -0.05, -1, and -5,
on the real axis, and two of the loci terminate at points -0.1 and -0.5.
(i) Sketch the probable root locus diagram of the system.
(ii) What is the specific controller being used?
(iii) What are the values, where possible, of the controller parameters?
Question 3.6
The figure below is the root locus diagram for a process,
(i) Obtain the open loop transfer function, gOL for the process
(ii) What are the values of the 2nd order parameters, τ and ζ that best characterizes the transient
response when kc = 10? Sketch this response, and label the response with the following
features: rise time, overshoot, decay ratio, settling time and period.
(iii) Can this system ever be unstable, if yes, at what kc value is the system on the verge of
instability?
(iv) Sketch the transient responses to a unit step input for the following kc values: 0.5, 1.25, 30,
55
Page 5 of 7
Question 4
For the system in figure 4 below,
yd + e u y
kc gOL
-
s+7
gOL =
s ( s + 5)( s + 15)( s + 20)
(i) Using the Routh criterion at what kc values will the system remain stable, be at the verge
of instability (marginal stability)., and be unstable?
(ii) Using the method of direction substitution or the Routh array, find the roots at the verge of
instability
(iii) using the guidelines/rules for plotting a fairly accurate root-locus diagram, plot the root
locus diagram.
(iv) Create a MATLAB file to do the following:
a. plot the root locus diagram
s = tf(‘s’);
gol=(s+7)/(s*(s+5)8(s+15)*(s+20));
rlocus(gol)
This is just a suggestion, there are so many ways to go about specifying the transfer function
(v) using the rlocfind ( ) command in MATLAB, select the intersection between the root locus and
the imaginary axis to confirm your answers in (i) and (ii).
(vi) use the sgrid command to plot lines of constant damping ratio, ζ , and natural frequency, ωn
1
where ωn =
τ
(vii) It is desired that the closed loop system response have an overshoot less than 5%, show
that this corresponds to ζ > 0.7, and a rise of 1 second, show also that this corresponds to ωn
>1.8. Locate on the root locus diagram the possible regions for a kc value?
zeta = 0.7;
wn=1.8;
sgrid(zeta,wn)
(viii) use the rlocfind command in MATLAB to select desired poles on the locus
[kc,poles]=rlocfind(gol)
Just click on the plot where you want the closed loop poles to be.
Page 6 of 7
(ix) For the K value, corresponding to the point clicked in (vi) above, obtain the closed loop
transfer function. Confirm this using the feedback command in MATLAB
gol_cl=feedbacl(kc*gol,1)
step(gol_cl)
(xi) Repeat steps (vi) –(viii) for points on the root locus diagram that will give a critically
damped response, overdamped response and an unstable response.
Page 7 of 7