Fuzzy-PID based Liquid Level Control forCoupled Tank (MIMO) Interacting System

Abstract – Process Liquid level control in
a tank has always been an important issue in
process industries. For MIMO nonlinear system
the multivariable control strategy must be
properly designed. Various control techniques
based on particular performance index are
applied for improving the dynamics of level
control. PID controller is often used for the
control of liquid level. Control signals of PID
can be tuned by changing the PID parameters
(Kp, Ki and Kd) for obtaining better oscillation,
steady state response and reduced error. PID
parameters are tuned using fuzzy logic controller
and this gives promising results. A self-tuning
fuzzy PID controller is designed based on the
error approximation and by deciding the crisp
values in the range. Two input and three output
fuzzy controller is used instead of two input and
single output. The control signal from the Self-
Tuning Fuzzy-PID controller is then given to
the coupled interacting tanks within the extreme
points. As compared with classical PID the
self-tuning fuzzy-PID controller shows better
performance and can be implemented with
positive approach.
Keywords – MIMO Theory, Coupled Tank
System; Proportional-Integral-Derivative (PID)
Controller; Fuzzy system; Self-tuning Fuzzy-
PID Controller
I. INTRODUCTION
Many industrial processes having problems
related with turbulence and smooth inflow,
level control is necessary. 25% of emergency
shutdowns are caused due to improper control
water level in steam generator. Such shutdowns
greatly affect the plant availability and must be
reduced. Water level control of coupled tank
interacting system is a very difficult, becauseof the system nonlinearities and uncertainties.
At present, constant gain PI controllers
consisting problems related to peak overshoot
are implemented in various small scale and
large scale industries where process control is
carried out. Therefor PI controller is avoided
due to the overshoot and oscillatory problems.
Performance improvement for process level
control is therefore in great need.