A. GANESH RAM et al: REAL TIME IMPLEMENTATION OF FUZZY BASED ADAPTIVE PI. . . . DOI 10.5013/IJSSST.a.14.06.01 ISSN: 1473-804x online, 1473-8031 print 1 Real Time Implementation of Fuzzy Based Adaptive PI Controller for a Spherical Tank System A Ganesh Ram Department of E&I, FEAT, Annamalai University, Annamalainagar, Tamil Nadu, India E-Mail: agram72@gmail.com S. Abraham Lincoln Department of E&I, FEAT, Annamalai University, Annamalainagar, Tamil Nadu, India E-Mail: linsun_2k5@yahoo.co.in Abstract — This paper proposes a new fuzzy adaptive variable digital PI controller for a single input single output non-linear spherical tank level process system. The open loop transfer function models are carried out at three different operating regions and those models are formulated based on the real laboratory scale system. The proposed FAPI controller is a combination of two input two output Fuzzy logic controller and a Variable digital PI controller. The input to the fuzzy controller is error and change in error and its outputs are K P and K I . The PI controller’s parameters are estimated on-line based on error and change in error. The real time implementation and control of the process plant is done in MATLAB using VMAT-01 Data Acquisition Module. The objective is to make the output to settle fast with minimum overshoot and the disturbances do not affect the performances of the system. Keywords - Fuzzy Adaptive PI (FAPI), Spherical tank, SISO, VMAT-01 I. INTRODUCTION Most of the process control industrial systems present challenging control problems due to the dynamic behavior, uncertain and time varying parameters. Controlling the level of the spherical tank is a challenging problem due to its change of shape which gives rise to the non-linearity. The spherical tanks find wide application in gas plants and petrochemical industries. The proportional Integral and Derivative (PID) controllers are commonly used in industries because of its simplicity in tuning the parameters and achieve satisfactory performance. Thus traditional PID algorithm doesn’t hold good for such systems which has disturbances by nature, so the conventional controller is not enough to control the highly non-linear process like spherical tank because the change in shape gives rise to the non-linearity. It requires some intelligent controller or adaptive controller to achieve the optimum performance when facing different operating condition and various type of disturbances. In this work, the proposed control algorithm that can overcome these limitations. The on-line estimation of the PI controller output u(k) is based on K P and K I , these parameters are estimated with current error (e) and change in error (ce) using fuzzy logic algorithm. The fuzzy controller is a non-linear controller and the fuzzy control algorithm is based on the intuition and experience about the plant to be controlled. The Fuzzy adaptive PI control technology has been widely used in many fields because of its simplicity, high precision and high robustness. It has many advantages in the form of fast response, minimal overshoot and good anti-interference ability. The real time liquid level spherical tank process is interfaced with MATLAB using simple cost effective VMAT-01 module. It consist of one ADC, one DAC and two PWMs. II. FUZZY ADAPTIVE PI CONTROL ALGORITHM The self-tuning fuzzy adaptive PI controller is an auto adaptive controller that is designed by combining fuzzy and digital PI controller. The structure of fuzzy adaptive PI controller is shown in Fig. 1. The fuzzy controller uses the error and rate of change in error as its input and meet desire of self-tuning parameters K P and K I . The objective is to find the fuzzy relations among K P , K I , error, and rate of change in error. With continual testing, the two output parameters are adjusted on-line so as to meet different requirements and achieve good stability. The digital PI controller tune on-line by using those fuzzy parameters and find the new controller output by the following equation.           k i i i P e T T k e K k u 0 ) ( ) ( (1) In the above equation, u (k) is the output of adjuster in the k th sampling. K is the sampling number (k=0, 1, 2…; K P ), and K P is the adjuster scale factor; e (k) is the error value in the k th sampling; T i is the integral time; T is the sampling period. K I =K P ×T/T i , write the above equation into the incremental control   ) ( ) 1 ( ) ( ) 1 ( ) ( k e K k e k e K k u k u I P       (2)