Water Tank Level Control System

using Self-Adaptive Fuzzy-PID Control

International Journal of Engineering Research & Technology
(IJERT)
ISSN : 2278-0181

Presentation by

: Noviyanti Sagala

Fuzzy Logic
Controller

Introductio
n
Metodology

Result

Conclusion

 Basic problem in the industries is water level controlling system

Control system consists of many constraints like non linearity, inertial lag,

time delay, time varying parameters which are very difficult to develop
the mathematical model of such systems.
Fuzzy Logic Controller is applicable to non-linear system and doesnt
depend on the precise mathematical model.
Fuzzy Logic can be understood as computation using linguistic variables
instead of numbers, whereas fuzzy control uses IF-THEN statements
instead of equations.
PID controller requires the exact model of the system while Fuzzy control is
an intelligent control method and this feature makes it self-adaptive.

Automatic water tank level control system

Water level measurement and precise water level control is an important process in improving
the manufacturing quality of products.

Mathematical Model

Assume R=2

Valve Lag (Measurement Lag (

 Transfer function of measuring lag

Transfer function for control valve lag
The complete transfer function of the given system by using ZieglerNicholas method tuning of PID controller

By substituting Calculation for balancing criterion:

)+
Real part :
Imaginary part:

= 0.6 = 7.62
= = 14.27
= = 3.55

PID Controller Response

X-axis: Time
(sec)
Y-axis: Level
(mm)

Adaptive Fuzzy Logic Controller

Database

Fuzzy Rule
Base

3
input

Fuzzifier

Inference
engine

defuzzifier

output

Block Diagram of Fuzzy PID

The main system is PID controller while the function of fuzzy logic is to repair
the response and recovery time of disturbance.

1. Input and Output

a. Deviation the difference between the desired height of water and the actual height
of water
Positive error the actual height of water below the reference height of water
Negative error the actual height of water above the reference height of water
b. Deviation change the difference between delayed-error value and new error
value

Positive Delta error error tends to going down

Negative delta error error tends to going up
c. Valve Position
2. Fuzzification
The process of transforming crisp values into grades of membership corresponding to
fuzzy sets using linguistic terms
Membership functions which used are Gaussian and triangular membership as the limit
value range.

Membership Function of input variables

1. Deviation
Fuzzy Variable

Crisp input range

High

(0.3, 1)

Zero

(0.3, 0)

Low

(-0.3, -1)

2. Rate of deviation change

Fuzzy Variable

Crisp input range

Positive

(0.03, 0.1)

Zero

(0.03, 0)

Negative

(-0.03, -0.1)

Membership function of Output variable

3. Valve Position
Fuzzy Variable

Crisp Output
Range

Close fast (CF)

(-1.0, -0.9, -0.8)

Close Slow (CS)

(-0.6, -0.5, -0.4)

No change (NC)

(-0.1, 0, 0.1)

Open Slow (OS)

(0.2, 0.3, 0.4)

Open Fast (OF)

(0.8, 0.9, 1.0)

3. Fuzzy Rule Base

Database
The database is a component for defining the fuzzy set of input and output

Fuzzy Rule Base

Fuzzy rule base is a collection of statements of rules ' IF- THEN ' which is based on expert
knowledge .
The method used in determining the fuzzy rule base is using MIN.
This method has the form of rules such as the following equation : IF x is A and y is B
then z = C

Rules Component

The goal is to get the output.

Fuzzy set is obtained by taking the maximum value of rule, then use it to modify fuzzy area and
implement it to the output.

Rule Editor

4. Fuzzy Inference Engine

a. FIS also called fuzzy inference engine is a system that can perform reasoning with
similar principles as humans do reasoning with the expert instincts .
b. FIS used is mamdani which works based on linguistic rules and has a fuzzy
algorithm which provides a aproximator to input mathematical analysis .

Defuzification

can be defined as the process of changing the amount of fuzzy presented in the form of
associations with the output fuzzy membership function to regain the form he asserted ( crisp
)

The result of defuzification which is used to adjust the value of Kp, Ki, and Kd

For example:
We assume that Error = 0.3, error=0.06
1. Determining the membership degree of error
From the membership function, we get
that
Min
(, 0.3
) is belonging to low and zero, so:
i.

Fuzzy Low

ii.

Fuzzy Zero

2. Determining the membership degree of error

From the membership function, we get that 0.06 is belonging to low and zero, so:

Fuzzy Zero
i.
ii.

Min (, )

Fuzzy Positive

3. Determining the membership of output

Min (, )
4. Determining the fuzzy variable of the output
0.2 is belonging to Open Slow (OS)
So, the result will be : IF error is Low and error is Zero then Valve position is Open Slow

Simulation Diagram & Result

Fuzzy inference is the process of formulating map from a given input to an

output using fuzzy logic. The mapping then provides a basis from which
decisions can be made, or patterns discerned.
Mamdani-type inference, expects the output membership functions to be
fuzzy sets. After the aggregation process, there is a fuzzy set for each
output variable that needs defuzzificationIn the controller, the
proportional factor Kp=7.71, the integral factor Ki=0.074, the differential
factor Kd=3.55, the proportion factor of fuzzy output Ku= 0.833,
simulation time is 500sec.

Conclusion

The control method adapted in this paper infers that the self-adaptive
fuzzy-PID controller has small settling time, minimum peak overshoot and
has high disturbance rejection capability compared to conventional PID
controller.

The future scope is that the adapted strategy can be applied for higher
order systems with more number of physical constraints and increased
rules.
X-axis: Time
(sec)
Y-axis: Level
(mm)