100% found this document useful (1 vote)
216 views13 pages

Experiment - 5: Temperature Control in A Mixing-Heating Process

1. An experiment was conducted to control the temperature in a mixing-heating tank process using an on-off controller and PID controller. 2. When using an on-off controller, the temperature response oscillated with a constant amplitude around the setpoint. When using a PID controller, it was able to track a change in setpoint and reject a disturbance in inlet temperature. 3. The Ziegler-Nichols method was used to tune the PID controller parameters, which resulted in improved setpoint tracking and disturbance rejection compared to the default PID settings. Controller performance was also evaluated under various process parameter changes.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
100% found this document useful (1 vote)
216 views13 pages

Experiment - 5: Temperature Control in A Mixing-Heating Process

1. An experiment was conducted to control the temperature in a mixing-heating tank process using an on-off controller and PID controller. 2. When using an on-off controller, the temperature response oscillated with a constant amplitude around the setpoint. When using a PID controller, it was able to track a change in setpoint and reject a disturbance in inlet temperature. 3. The Ziegler-Nichols method was used to tune the PID controller parameters, which resulted in improved setpoint tracking and disturbance rejection compared to the default PID settings. Controller performance was also evaluated under various process parameter changes.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 13

Experiment – 5

Temperature Control
in a Mixing-Heating
Process

Name: K. Navya Manisha


Roll No.: 186122

1 | Page
Aim:

To control temperature in a mixing-heating tank process using On-Off Controller, PID


controller.

Description:

A temperature control system for a water tank with continuous inflow and outflow is
simulated. The water is heated by a heating element which is controlled by the controller. The
temperature is measured by a temperature sensor which in practice may be a Pt100 element
or a thermocouple.

The process model on which the simulator is based is a first order model based on energy
balance under the assumption of homogenous conditions in the liquid and in the tank. The
process model also contains a time delay. In addition the simulator contains a first order
transfer function representing a time constant in the heating element.

The parameter values are shown on the front panel of the GUI. The process model used in the
simulator is based on energy balance:

(1) d(c ρ V T 1)/dt = Ke u + c w (Tin - T1) + U (Tenv - T1)

where Keu = P is the power delivered by the heating element. T1 is the temperature in the tank
assuming homogeneous conditions. In practice there is a time delay between an excitation in
the heating element and the response in the temperature sensor. P is the manipulated variable.

(2) The measured value of temperature is T(t) = T 1(t-θ)

We assume that this time delay is inversely proportional to the mass flow w:

(3) θ = Kt/w

By taking the Laplace transform of the model above, one can get the following transfer
function from the control signal to the temperature T:

(4) Gp(s) = T(s)/u(s) = [Kp e-θs /(τs+1)]

The corresponding expressions for the model parameters are:

Gain Kp = Ke / (c w + U)

Time constant τ = ρV / (w + U/c)

Time delay θ = Kt / w

In addition, the heating element also has dynamics and is represented with a first order
transfer function model with unity gain and time constant.

2 | Page
The nominal operating conditions in the process are:

● The tank temperature T is 50 degrees C.


● The inlet temperature Tin is 20 degrees C.

The default process parameter values are shown on the GUI.

Procedure:

Unless otherwise stated, all the default values of the parameters are given in the GUI.

1. On/Off Control: Set the controller in On/Off mode. Observe the oscillations and the
magnitudes. Draw the corresponding graphs.
 
2. PID Control: Set the controller in PID mode. Use the default PID controller values
given and for a specific set-point of tank temperature, observe the closed loop
performance. (a) Change the set point as a step, say from 40 to 50 degrees C. (b)
Change the disturbance Tin from 20 to 15 degrees C, and analyze the disturbance
rejection ability of the controller. Are you able to obtain zero steady-state control
error? Draw the graphs.
 
3. Controller Design using Ziegler-Nichols Method: Tune the controller using the
Ziegler-Nichols closed loop method. For this, ultimate gain and ultimate period values
need to be obtained for the closed loop. Based on these values, design a PI controller.
(ex. Kp = 2.6, Ti = 6.3 min, Td = 0s). Implement the PI controller and then evaluate the
closed loop performance for set point tracking and disturbance rejection. Draw the
graphs.
 

Additional Understanding: Reset the settings of the simulator to its default values
(menu: Operate / Reinitialize All to Default). Check what happens with the stability
of the control system at the below parameter changes. In each of the
subtasks/experiments you can excite the control system using a small step in the set
point. The experiments shall be performed independent of each other, that is, the
parameters should be reset to the default values for each study.

a. Change the Controller gain Kp .


b. Change the Integral time Ti 
c. Change the Derivative time  Td 
d. Decrease the Mass flow ‘w’

3 | Page
e. Change the Measurement element gain Km. (If you change the measurement
range, you must also change the set point range accordingly, otherwise there
will be a steady-state error.)

Observations:

1. The response of an on/off controller results in sustained oscillations with a constant


amplitude and the setpoint was 40°C. Remaining parameters are at default values and
constant throughout the experiment.
2. a) The PI Controller response is recorded for setpoint change and disturbance change ,
The controller gain is maintained at 2 and integral time constant Ti at 300sec. The
initial set point was kept at 40°C and the system is allowed to reach the steady state
value. At 70 min, a set point change is given from 40°C to 50°C and because of the PI
controller, the output reaches the steady state in 20 min.

b) Now the system is initialized and allowed to reach a steady state value equal to the
set point of 40°C. After this a disturbance change (Tin) is given. Tin is changed from
20°C to 15°C and because of the controller, the system reaches the steady state i.e, to
the set point. We can observe that the response is first decreased and then increased to
reach the set point.

3. For designing a controller using Ziegler - Nicholas method, we have to find ultimate
gain and ultimate period.

a) The value of Kc is changed from 2 to 30, by keeping Ti and Td to be 0. Now Kc


has increased from 2 to 30 by using the values: Kc =2, 5, 10, 15, 20, 25, 30. As Kc is
changed from 2 to 25, the number of oscillations increased and the offset decreased.
We can observe that the oscillations had a decrease in amplitude between successive
peaks and eventually reached a steady state. But as Kc is changed from 25 to 30, the
amplitude of successive peaks starts increasing, this shows the unstable region.

b) From the above part, the ultimate gain of the system is to be Ku = 26.08, Ti=Td=0.
There are sustained oscillations with constant amplitude. Ultimate period (Pu) is the
time between two successive peaks and is equal to 5 mins.

c) The values of Kc and Ti are found by the respective formulae of the Ziegler -
Nicholas method and from those values, response of the system is observed. Kc is
found to be 12.06 and Ti to be 250sec.

d) Now for the PI controller with K= 12.06 and Ti = 250 sec, set point tracking is
done and observed the changes. The set point has been increased from 40°C to 50°C.
The system was able to reach the steady state/ setpoint with some oscillations.

4 | Page
e) In the disturbance(Ti) is changed from 20°C to 10°C and the system was able to
achieve the setpoint of 40°C with some oscillations.

f)The values of Kc,Ti, Td are found by the respective formulae of the Ziegler -
Nicholas method and from those values, response of the system is observed. Kc is
found to be 16.08, Ti to be 150sec, Td to 37.5sec.

g) Now for the PID controller with K= 16.08, Ti = 150 sec, Td = 37.5s, the set point
tracking is done and observed the changes. The set point has been increased from
40°C to 50°C. The system was able to reach the steady state/ setpoint with some
oscillations.

h) In the disturbance(Ti) is changed from 20°C to 10°C and the system was able to
achieve the setpoint of 40°C with some oscillations.

4. a) The controller gain (kp) is changed from 2 to 20 and the effect is observed on the
closed loop response. As the Kc is increased, the number of oscillations increases and
so does the response time. Setpoint was at 40°C for Kc=2 and was changed to 50°C
when Kc was changed to 20. Ti = 300 and Td=0.

b) Integral time is changed from 300s to 100s and the effect is observed on the closed
loop response. Kc is kept at 2 and Td = 0. Remaining parameters were set to be
constant at their default values. For Ti= 300s, the set point is 40°C and for Ti= 100s,
the set point was changed to 50°C. As Ti decreased, the number of oscillations
increased.

c) Derivative time(Td) is changed from 10 to 50s and the effect is observed on the
closed loop response. Kc is kept at 2 and Ti = 300s. Remaining parameters were set to
be constant at their default values. For Td= 10s, the set point is 40°C and for Td= 50s,
the set point was changed to 50°C. As Td increased, the number of oscillations
decreased.

d) Mass Flow(w) is changed from 0.33kg/s to 0.033kg/s by changing the volumetric


flow from 20L/min to 2L/min. Kc is kept at 2, Ti = 300s, Td = 0s. Remaining
parameters were set to be constant at their default values. The set point was at 40°C
for w = 0.33kg/s and was changed to 50°C when w= 0.033kg/s. The peak value
attained also decreased and the response time was the same. For constant signal there
is decrease in the final value with decrease in the mass flow(w)

5 | Page
Graphs:

Temperature in tank (◦C) vs Time (min) Control Signal (kW) vs Time (min)
(Temp_sp vs t) ( u vs t )

Response of an
On-Off controller

Input Parameters:
1.T_sp = 50◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C

2)PID Control

a)Set Point Tracking

Input Parameters:
1.T_sp = 40 to 50◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C
5.Kc= 2Kw/c
6.Ti=300s
7.Td=0s

6 | Page
b) Disturbance
Rejection

Input Parameters:
1.T_sp = 50◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C to 15◦C
5.Kc= 2Kw/c
6.Ti=300s
7.Td=0s

3)Controller design
using Ziegler-Nichols
Method

a)Finding Ultimate
gain and Ultimate
period.
Input Parameters:
1.T_sp = 40◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C
5.Kc=
2,5,10,15,20,25,30Kw/c
6.Ti=0s
7.Td=0s

b)Response of the
system of P controller
with Ultimate gain
Input Parameters:
1.T_sp = 40◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C
5.Ku=26.08Kw/c
6.Ti=0s
7.Td=0s

7 | Page
c) For PI controller

Input Parameters:
1.T_sp = 40◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C
5.Kc=12.06Kw/c
6.Ti=250s
7.Td=0s

d)For PI controller
and set point tracking

Input Parameters:
1.T_sp = 40◦C to 50◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C
5.Kc=12.06Kw/c
6.Ti=250s
7.Td=0s

e)For PI controller
and disturbance
rejection

Input Parameters:
1.T_sp = 40◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C to 10◦C
5.Kc=12.06Kw/c
6.Ti=250s
7.Td=0s

8 | Page
f) For PID controller

Input Parameters:
1.T_sp = 40◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C
5.Kc=16.08Kw/c
6.Ti=150s
7.Td=37.5s

g)For PID controller


and set point tracking

Input Parameters:
1.T_sp = 40◦C to 50◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C
5.Kc=16.08Kw/c
6.Ti=150s
7.Td=37.5s

h)For PID controller


and disturbance
rejection

Input Parameters:
1.T_sp = 40◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C to 10◦C
5.Kc=16.08Kw/c
6.Ti=150s
7.Td=37.5s

9 | Page
4) Additional
Understanding

a) Effect of controller
gain Kc

Input Parameters:
1.T_sp = 40◦C to 50◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C
5.Kc=2 to 20 Kw/c
6. Ti = 300s
7. Td=0s

b) Effect of integral
time Ti

Input Parameters:
1.T_sp = 40◦C to 50 ◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C
5.Kc=2Kw/c
6.Ti=300 to 100s
7.Td=0s

c) Effect of derivative
time Td

Input Parameters:
1.T_sp = 40◦C to 50 ◦C
2. F = 20 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C
5.Kc=2Kw/c
6.Ti=300s
7.Td=10- to 50 sec

10 | Page
d) Effect of mass flow
w
Input Parameters:
1.T_sp = 40◦C to 50 ◦C
2. F = 20 to 2 L/min
3.Temp_env = 20◦C
4.T_in = 20◦C
5.Kc=2Kw/c
6.Ti=300s
7.Td = 0s

Inference and Conclusions:

1) On/Off controller

On/Off controller sets the input to the higher or the lower value present. As we can
see from the graphs, when there is a power input of 80kw, the power is given and the
temperature increases. Once it crosses the setpoint the controller shuts the power. Due
to inertial nature the temperature decreases and once it reduces below the setpoint the
controller toggles into ON mode and the power is given. Hence we can see the
oscillations as the measured temperature is continuously decreasing and increasing
with the power input.

2) PID controller

a)For the first case as the setpoint is increased from 40 to 50°c, we can see that
measured temperature increases and reaches the steady state value with no offset as it
is a PID controller. We can also observe that to maintain the setpoint at 50°c, the
controller signal power input increases to 75 kW

b)For the second case of disturbance rejection as the input temperature is decreased
from 20 to 15°c,, there will be a decrease in measured temperature. In order to reach
the setpoint value of 50°c, the power input is increased and steady state value of 50°c
is achieved and the disturbance is totally rejected. With the PID controller, the offset
can be removed and response speed is improved.

11 | Page
3) Controller Design

Controller Design using Ziegler Nicholas method :

Ziegler Nicholas is a closed loop tuning method for finding controller settings.

Ultimate gain (Ku): It is the largest value of the controller gain Kc that maintains
closed loop stability when a proportional controller is used. From the graph, the
ultimate gain is at 26.08 and is found by testing the simulator at various values of Kc.
The period of oscillation is the time period between two peaks and at Ku it is the
ultimate period(Pu) and is 5 mins for the system.

Ziegler Nicholos tuning relations:

Controller Kc Ti Td

P 0.5Ku - -

PI 0.45Ku Pu/1.2 -

PID 0.6Ku Pu/2 Pu/8

PI Controller settings => Kc = 0.45Ku = 0.45*26.08 = 12.06

=> Ti = Pu/1.2 = 300/1.2 = 250s

PID Controller settings => Kc = 0.6Ku = 0.6*26.8= 16.08

=> Ti = Pu/2 = 300/2 = 150s

=> Td = Pu/8 = 300/8 = 37.5s

Setpoint Tracking:

We can see that in both PI and PID controllers, the steady state value of 50°c is reached when
there is a step change from 40 to 50°c. But the time taken to reach steady state is 27 min with
PI and 17min with PID controller. This infers that the PID controller provides faster response
when compared to PI controller. This is due to the derivative action.

12 | Page
Disturbance Rejection:

As the environment temperature decreases from 20 to 10°C, more power is required to


achieve the setpoint value. From graphs we see that both PI and PID controllers nearly take
the same time i.e. 12 min to reach the steady state value. From the graphs, clearly we can say
that the effect of the PID controller is less due to disturbances.

4) Additional understanding:

a) Change in controller gain:

As the value of the controller gain is decreased, the number of oscillations also
increases and so does the response time also increases. We can also infer that as Kc
value increases the settling time decreases and the system reaches a steady state.

b) Change in Integral time(Ti):

As the value of integral time decreased from 300 to 100s we can see that, as the
integral time decreased, the number of oscillations increased.

c) Change of derivative time(Td):

As the value of derivative time is increased, the number of oscillations decreases.

d) Decrease the mass flow(w)

The change in volumetric flow will provide the same results as mass flow and they
are directly proportional to each other. So from the graph we can see that as the flow
rate decreases, settling time also decreases. Also, as the mass flow is decreasing,
power required to raise the temperature to setpoint is also decreasing.

13 | Page

You might also like