CONTROL ENGINEERING

EXPERIMENT NO:-

Aim: Study of On OFF, P, PI, PD and PID controller for liquid level control
system.

Apparatus: A liquid level control trainer set-up, compressed air supply, P.C set up,
software.

Theory:

1) Proportional controllers: There are many situations in industrial as well as

commercial instillations, in which the controller output is continuously required
and hence, on/off controllers are not suitable at such locations. Here, we have to
use such a controller in which the output is an analogue value and it is
continuously adjusted. Most of the continuous controllers presently use some
combination of proportional, integral and derivative action. The control algorithm
that generates a linear control output proportional to deviation is called
proportional action. The amount of change in the measured value (or deviation) is
expressed in percent of span that is required to cause the control output to change
from 0 to 100% is called the proportional band. In proportional controllers, the
output is proportional to the error signal and hence, we call them proportional
controllers. The output is large for large error signal and vice versa. There are two
possibilities. One, the output might increase when the error is positive. Second, the
output may decrease when the error is positive. In first case, we call the controller
as direct acting and in the other case; the controller is called reverse acting. For
example; in the case of a tank level control, it depends on the location of control
valve. If valve is controlling flow out of the tank, we would like that a positive
error should increase the control output by opening valve and letting more fluid out
of the tank. It will require that we should use a direct acting controller. On the other
hand, if a valve is controlling in flow to the tank, a reverse acting controller would
be used. This controller would reduce the flow in to the tank whenever a high level
is observed. The controller output is given by,

DEPT. OF MECHANICAL ENGG. P.V.P.I.T BUDHGAON 1

CONTROL ENGINEERING

P = b + Kp.e

Where, P = output, b = bias, e = error signal

Kp = proportional gain (100 % proportional band)

DEPT. OF MECHANICAL ENGG. P.V.P.I.T BUDHGAON 2

CONTROL ENGINEERING

DEPT. OF MECHANICAL ENGG. P.V.P.I.T BUDHGAON 3

CONTROL ENGINEERING

If there is no biasing, output P will become zero when error is zero. Hence biasing
is essential for a proportional controller to have some output when error is zero.
Biasing decides the value of output when error is zero. When error has non-zero
value, the output will increase. The controller may have fixed bias value in some
cases, but in others it may be mutually adjustable. The more is the proportional
gain, the more will be the output for certain change in the extent of control action
that will generate by a given error signal. The problem with proportional
controllers is that there is always some difference in the set points and process
variables. This is also called offset. Hence, it requires that some non-zero signal
should be available to generate the control output necessary to stabilize the process
set point. The control output require to achieve the set point would have to be
exactly equal to the bias value for the control to be accurate at the same point. In
other situations, we will require a non-zero signal to generate the appropriate
control output. This will result in an offset as shown in Figure 3.1. We can reduce
the offset by increasing the proportional gain but it cannot be eliminated. If we try
to eliminate offset by increasing gain indefinitely, the output will become
oscillatory.

2) Proportional + Integral (PI) controller: As mentioned above the main

problem with proportional controllers is the offset. To overcome this problem,
integral action is added to control. The control algorithm that applies changes in
output as long as deviation exists, so as to bring the deviation to zero, is called
integral action. When integral action is used, the parameter that determines how
fast the output will change in correspondence to some amount of deviation is
referred to as the integral time, Shorter the integral time, stronger the integral
action (the greater the output rate of change). Output of proportional integral
controllers is given by, P = Kp + Ki ∫ e dt

Where, P = output

Kp = proportional gain (100%proportional band)

DEPT. OF MECHANICAL ENGG. P.V.P.I.T BUDHGAON 4

CONTROL ENGINEERING

e = error signal

Ki = integral gain

DEPT. OF MECHANICAL ENGG. P.V.P.I.T BUDHGAON 5

CONTROL ENGINEERING

DEPT. OF MECHANICAL ENGG. P.V.P.I.T BUDHGAON 6

CONTROL ENGINEERING

Integral action goes on increasing the output till the error becomes zero in case of
positive error and goes on decreasing in case of negative error. The above equation
applies to non-interactive controllers, in which proportional and integral gains can
be adjusted independently. There are interactive controllers, in which the output is
given,

P = Kp (e + Ki ∫ e dt)

In these controllers, proportional as well as integral gain decides the integral action
and such controllers are interactive. Integral gain is generally given in repeats per
minute. It decides how many times the proportional action is “repeated” each
minute. In some controllers, this gain is given in minutes per repeat i.e. the minutes
required to repeat the proportional action. The function of integral action is to
eliminate the need of bias value found in proportional controllers. It also eliminates
the problem of offset. It happens because error is continuously added up over time.
The control output of a controller with integral action will continue to change as
long as the error is non zero. It ceases to change when the error becomes zero.
Some people call integral action as reset because it resets the offset value to zero
over time. Integral action provides an automatically adjusting floating bias that
continuous to change as long control as the error is non zero. In stable process,
integral action will guarantee that the steady state error eventually becomes zero as
shown in the Figure 3.2
3) Proportional + Derivative (PD) controller:

This mode of control is decided by the relationship,

P = Kp e + Kd

Where, P = output

Kp = proportional gain (100%proportional band)

DEPT. OF MECHANICAL ENGG. P.V.P.I.T BUDHGAON 7

CONTROL ENGINEERING

e = error signal

Kp = derivative gain

The derivative action is generally used for the slow processes. Larger the derivative
time larger is the action. Smaller is the proportional band the larger is the
derivative action. If the control object has a large time constant or dead time, with
P or PI action alone there will be causes where the response will slow, overshoot
will occur, and the control system will be unstable.

In order to achieve faster response and more stable operation in these cases one
uses derivative action to apply an output component proportional to the input
(deviation) rate of change. D action must be used with P action or PI action as PD
or PID action.

4) Proportional + Integral + Derivative (PID) controllers:

The function of the derivative action in continuous controllers is to help them in

dealing with sluggish processes, including large inertia, lot of time even with
integral action. Even overshooting is not ruled out. Derivative action can overcome
this problem. The output of a PID controller is given by,

P = Kp e + Kp e dt

Where, P = output

Kp = proportional gain (100% Proportional band)

e = error signal

Ki = integral gain

Kd = derivative gain

DEPT. OF MECHANICAL ENGG. P.V.P.I.T BUDHGAON 8

CONTROL ENGINEERING

As in proportional integral controllers, PID controllers are also sometimes

implemented with interactive gains. In interactive controllers, the derivative action
is function of proportional as well as the derivative gain. To accelerate the sluggish
process, derivative action responds to the rate of change of error. Derivative action
sometimes called “rate” in some controllers because of this action.

The derivative action basically reduces the overshoot, anticipating that the process
is about to reach the set point.

Take the example of controlling the temperature of a large room. An increase in the
set point may require the control output to be 100% for a longer period of time to
heat up the area rapidly. Once the temperature rises, the heat input needs to be
sharply decreased to avoid over shooting the set point. The derivative action
achieves this efficiency. In some controllers, the derivative action is also called as
anticipation because it said to predict the future value of the error. Derivative
action can cause some problems in actual practice and should be used wisely, as
the derivative action responds to the rate of change of error, there may be problems
with the rapidly changing error signals caused by noisy process variable
measurement or set point changes. The derivative will behave erratically in such
cases and may result in unstable controller performance. Therefore, the practice is
to address these problems by smoothing the process signals to eliminate noise.
Applying the derivative action to the process variable rather than the error solves
the problem of fast changing set point.

The integral and derivative controllers are always used in conjunction with the
proportional controller. These are never used independently. Earlier, various modes
of a controller used to be realized mechanically, pneumatically etc. now we can
easily realize them electronically. We can select the proportional, integral and
derivative gains to get suitable response. Selection of these gains is called tuning of
controller. Tuning of controller is time consuming and requires patience. Today,
controllers come equipped with auto tuning features and hence can be easily tuned.
Analogue controller suffers from ageing and drift in various parameters. After
certain time, these controllers require calibration. It is must in any industry to
maintain the controller to assure quality and consistency of the product. To

DEPT. OF MECHANICAL ENGG. P.V.P.I.T BUDHGAON 9

CONTROL ENGINEERING

overcome these difficulties, some manufactures have introduced digital controllers.

PID controllers is used for controlling almost all variables like Temperature, flow,
level, pressure experienced in a continuous or batch process.

DEPT. OF MECHANICAL ENGG. P.V.P.I.T BUDHGAON 10