Modelling and Control of Distillation Column's Steam Temperature using MPRS and PRBS Signals

Abstract

Many studies have been conducted to demonstrate that a distillation column works best with a nonlinear model. This study proposes a nonlinear modelling and control using conventional Proportional Integral Derivative (PID) and Fuzzy-PID controllers for steam temperature of Steam Distillation Pilot Plant (SDPP). Two perturbation input signals, Multi-level Pseudo Random Sequence (MPRS) and Pseudo Random Binary Signal (PRBS) were employed. The Nonlinear Auto-Regressive with eXogenous input (NARX) model was selected based on the high percentage of fitness and low Mean Square Error (MSE). Then, the controllers were designed using the estimated transfer function. Fuzzy Logic Toolbox in MATLAB was used to construct Mamdani Fuzzy Inference Systems (FIS) based on 5 membership functions and 25 fuzzy rules. Both inputs and outputs used triangular and trapezoidal membership functions. Tyreus-Luyben technique was used to tune the controllers. Step change test and load disturbance test have been simulated for both controllers to determine its stability and performance. The performance of conventional PID and Fuzzy-Pidcontroller was evaluated based on the rise time (Tr), settling time (Ts), percentage of overshoot (%OS), and Integral Squared Error (ISE). The results indicated that the Fuzzy-PID controller outperforms the conventional PID controller in terms of performance and stability through setpoint change and disturbance tests.