Lab 9 Report: Process Control Trainer (II)

31 October, 2022

Group 5A

T Pratham − 2020EE10559
Afreen Haider − 2020EE10463
Pushp Srivastava − 2020EE30954
Anmol S − 2020EE10471
Arshia Dhiman − 2020EE11002

1
1 Objective
To familiarize with industrial processes and the control systems applied in them, and study
characteristics of distance/velocity lag, transfer lag, two step control and proportional control.

2 Equipment Required
• 37-100 Process Control Trainer

• Digital Oscilloscope

• Function Generator

3 Control Application
The basic circuit of a closed loop process control system is shown as below

Figure 1: Closed Loop Process Control System

There are two basic control operations

1. Open Loop Operation: In this, control input is a function of reference input, i.e.,
u = f (r)

2. Closed Loop Operation: Control input is a function of both reference and output. One
of the basic closed loop control approach is proportional control, and in which control input
is proportional to the error signal, i.e., u = K · e

3.1 Disturbance and System Response

Disturbances are unwanted signals entering into a feedback control system. A disturbance may
act at the input or output of the plant. Here we consider the effect of input disturbance.
To characterize the effect of a disturbance input on the feedback control system, we consider
the modified block diagram (Figure 2) that includes a disturbance input.
The basic circuit of a closed loop process control system is shown as below

2
 Figure 2: A feedback control system with reference and disturbance inputs.

Let r(t) denote a reference, and d(t) a disturbance input; then the system output is expressed
in the Laplace domain as:
KG(s) G(s)
y(s) = 1+KGH(s) r(s) + 1+KGH(s) d(s)

Assuming unity-gain feedback configuration (H(s)=1) , the tracking error, e(s), is computed
as:
1 G(s)
e(s) = 1+KG(s) r(s) + 1+KGH(s) d(s)

By using the FVT, the steady-state error is expressed as:

1 G(0)
e(∞) = 1+Kp r(∞) − 1+Kp d(∞)

where Kp is the position error constant.

A large loop gain (large Kp ) reduces steady-state error in the presence of both reference and
disturbance inputs. A large controller gain, K , can be used to increase Kp , however, a large
K would generate a large magnitude input signal to the plant, which may cause saturation in
the actuator devices (amplifiers, mechanical actuators, etc.).

3.2 Frequency Response

One way to conceptualize a process is as a series of transfer stages, each with a time lag and a
distance/velocity lag. When a control loop is opened and a sine wave is applied to the input,
measurements of the output signal’s amplitude and phase are useful for figuring out the system’s
closed-loop stability.

Figure 3: Control system with transfer blocks

A Bode diagram or a Nyquist diagram might be used to plot the outcome. Finding the
frequency at which the phase lag is 180◦ and measuring the gain at this frequency are the
primary goals in each situation. Gain is defined as:
∆Vout
Gain = (1)
∆Vin
In some controllers the input and output signals may be in different units, and to avoid conver-
sion factors the output/input relationship may be expressed as ’Proportional Band’.The relation
between gain and percentage proportional is:
100
Gain = (2)
%proportional band

3
Phase is the angle in degrees by which the output signal lags the input signal, and is defined as:

time lag, input to output signal × specs × 360

ϕ= (3)
Period of signal, Specs

3.3 Compound Controller Action

This is used to analyse the effect of proportional, integral and derivative control on a system.

1. Proportional: Increases gain and oscillations of the system

2. Integral: Reduces steady state error

3. Derivative: Reduces overshoot response

4 Critical Analysis
4.1 Disturbance and System Response
The set value is adjusted to 50% and proportional band to 100%. The blower throttle control
is adjusted to 4. The set value disturbance is provided as a square wave output of peak to peak
voltage of 2 V at 1.5 kHz.
A step change in the set value disturbance produces an immediate change in control output and
a delayed response in the measured value due to the distance/velocity and transfer lags.
At 100% response is damped enough to give little oscillation or overshoot as shown in the image
below

Figure 4: Proportional band to 100%

At proportional band of 50% the response is damped oscillatory with some overshoot

Figure 5: Proportional band to 50%

4
 When proportional band is set to 50% and set value at 50◦ , the graph obtained for measured
value of control voltage in open and closed loop is as shown below. Series 1 corresponds to closed
loop while series 2 corresponds to open loop.

Figure 6: Measured value against blower throttle control

4.2 Frequency Response

4.2.1 Open Loop Response
Set value is adjusted to 35◦ and throttle control set to 4. The input is varied over a range of
frequencies from 0.1 Hz to 3 Hz and the phase and amplitude relation of the input and output
are measured.

fin (Hz) Vin,p−p (V ) Vout,p−p (V ) ϕ( in ◦ ) Gain(in dB)

0.13 2.16 2.24 34.3 0.315885344
0.29 2.56 2.4 69.5 -0.560574472
0.495 2.24 1.6 120.2 -2.922560714
0.79 1.96 1.12 158.2 -4.860760974
1 2.24 1.04 -1.9 -6.664293581
1.2 2.28 0.72 -61.5 -10.01204701
1.4 1.92 0.56 -48.8 -10.70226403
1.6 1.88 0.56 -85.5 -10.51939645
1.8 1.84 0.88 -147.3 -6.406703017
2 1.88 0.4 -63.3 -13.44195716
3 1.92 0.32 -5.6 -15.56302501
0.89 2.16 1 172 -6.689075023

Table 1: Open loop response

The bode plots of the open loop response are given below

5
 Figure 7: Magnitude Plot

Figure 8: Phase Plot

When phase lag = 180◦ , gain ≈ 0.463

Frequency of input signal = 0.89 Hz
For sustained oscillations, Proportional band X can be calculated from
100
×G=1 (4)
X
Proportional band = 46.3 %

4.2.2 Closed Loop Response

Set value is adjusted to 35◦ and throttle control set to 4. The input is varied over a range of
frequencies from 0.1 Hz to 3 Hz and the phase and amplitude relation of the input and output
are measured.

6
 fin (Hz) Vin,p−p (V ) Vout,p−p (V ) ϕ( in ◦ ) Gain(in dB)
0.2 4.96 4 -47.8 -1.868433703
0.594 4.64 2.56 -103 -5.165560305
0.839 4.48 1.76 -179.1 -8.115306924
1 4.48 1.36 -263 -10.35478211
1.2 4.4 0.64 -304 -16.74545405

Table 2: Closed loop response

When phase lag = 180◦ , gain ≈ 0.393

Frequency of input signal = 0.89 Hz
For sustained oscillations, Proportional band X can be calculated from
100
×G=1 (5)
X
Proportional band = 39.3 %
After adjusting proportional control to the above obtained value, oscillations are observed.
Double the proportional gain gives a stable system

Figure 9: Oscillations with proportional control set to 39.3%

4.3 Compound Controller Action

The throttle control is set to 4 and the system is allowed to stabilize. The proportional control
is varied and the set and measured values are noted.

Kp Set value Measured value

2 40 32
3 40 35
4 40 37
5 40 38

Table 3: Measured voltages at various control voltages

Oscillations are observed to start at proportional gain Kp = 5

The gain is then reduced to 2.5. The measured value is obtained to be 34 and the set value is
at 40.
Error = 6◦
The integral controller is then switched in and adjusted to make difference between set and
measured values to be zero.

Ki = 0.4

7
The X-Y curve of the above P-I controller is obtained as shown below

Figure 10: Proportional and integral controller

The derivative control is then switched in and adjusted to reduce overshoot as Kd = 12 The
X-Y plot of the PID controller is shown below.

Figure 11: Proportional integral and derivative controller

5 Practical Examples
The industrial applications of such systems are generally termed as temperature control systems,
and they are used in a variety of manufacturing and operations processes. A block diagram of
the process is given below:-

8
 Figure 12: Temperature Control System

A temperature control system is mainly used to control the temperature of an area or an

object, and it’s uses range from air conditioners to heat exchangers. We can divide its uses in
the following brackets:-

1. Heat Treat/Oven: We use such type of control systems in heat-treating application

such as furnaces, ceramic kilns, boilers, heat exchangers etc.

2. Packaging: All machines that are equipped with seal bars, glue applicators, hot melt
functions, shrink wrap tunnels or label applicators, need to operate at specified tempera-
tures so as to function properly. Herein, temperature control monitors are used to regulate
these operations and ensure high quality products.

3. Plastic Industry: In the plastic industry, we see applications of temperature control

in portable chillers, hoppers, and drying, molding and extruding equipment. We use
temperature control system to precisely monitor and control temperatures at different
critical points in the production of plastic.

4. Healthcare: To increase the accuracy and precision of temperature control, temperature

controllers are used. Common equipment are laboratory and test equipment, autoclaves,
incubators, refrigeration equipment, crystallisation growing and test chambers, as these
require specific temperature parameters.

6 Key Takeaways
6.1 Disturbance and System Response
We observe that as proportional band is decreased, the oscillations in the response increase when
a disturbance is applied to a system. We also see that when the blower throttle is increased the
control voltage decreases. The decrease (slope) is less for open loop than for close loop.

6.2 Frequency Response

In open loop configuration, the gain as well as phase decrease on increase in frequency. A phase
lag of 180◦ is observed at a frequency of around 0.9Hz. The proportional band in open loop is
calculated as 46.3%.
In closed loop, the proportional band is calculated as 39.3%.

9
6.3 Compound Controller Action
We have seen that a PID controller can be designed to get a error-driven control system. A
proportional controller only gave an offset value initially and then, introduced oscillations in
the system.
The value of Kp for which oscillations began = 5. Hence, the optimal choice of Kp is
= 2.5 and the difference in set value and measured value at this Kp = 2.5 is = 6.
Now, when integral controller was added, the value of Ki for which the error reduces
to zero is = 0.4
Now, the value of Kd found for which overshoots are smoothened is = 120.

7 Beyond the Manual

The degree of control of the process depends on the values need to be chosen for the proportional,
integral and differential constants. The value of proportional, integral and derivative can be
estimated using various tuning methods. In the experiment that we performed, we have only
considered proportional gain. We have not made use of the other 2, i.e. integral and derivative
constants. By adding integral gain of suitable value (Ki), we can improve the steady state
characteristics of the process. By adding differential gain of suitable value (Kd ), we can improve
the transient characteristics – such as reducing the overshoot.
Ideally, we should be able to consider all (Kp , Ki , Kd ) combinations and look at the responses
and choose the best triplet, in terms of generating the output with lowest rise time, lowest
steady state error and lowest overshoot characteristics. But that would be very tedious to do.
Thus, we should try to make use of the classical PID tuning methods such as Ziegler-Nichols or
Cohen-Coon method.

Ziegler-Nichols method: It is a heuristic method. We set Ki and Kd to zero, and increase

Kp from zero until it reaches the ultimate gain, at which output is marginally stable, i.e. has
stable oscillations. The Ku and time period Td values are then used to set the P, I, and D gains.

Figure 13: Ziegler Nichols Method

Cohen Coon method: Alternatively, the Cohen Coon method can be used, which is suited
for a wider variety of processes than Ziegler Nichols.

10
 Figure 14: Cohen Coon Method

Also, these tuning methods only give a starting point for (Kp , Ki , Kd ) values. We can
adjust the values around that point and observe the output response, and find the most optimal
(Kp , Ki , Kd ) triplet.

8 Possible Reasons for Error

• Difficulty in measurement of transport lag due to absence of a sharp peak in amplitude.

• Noise (from atmosphere, equipment, etc.)

• Incorrect measurement due to faulty equipment

11