Nonlinear Modelling and Fuzzy Control of Steam Temperature for Distillation Column Pilot Plant

Temperature is the most critical characteristic in the steam distillation process since it directly affects the amount of oil produced and its quality. This paper presents the nonlinear modelling and Fuzzy Logic Control (FLC) for steam temperature control of the distillation process. Pseudo Random Binary Sequence (PRBS) and Multi-level Pseudo Random Sequence (MPRS) were used for nonlinear modelling of the steam temperature. The suitable transfer function for Nonlinear AutoRegressive with eXogenous input (NARX) modelling was selected with a high percentage of fitness and low value of mean square error. Proportional Integral Derivative (PID) and FLC control tuning method was design based on the estimated transfer function. Then, triangular and trapezoidal type of Membership Function (MF) is used in an FLC system that consists of 2 inputs and 1 output, which are error, derivative error, and voltage, respectively. 7MF with 49 fuzzy rules was used to perform the FLC. The simulation result reported that FLC with MPRS input signal presented better performance with 2364 s rise time, 2914 s peak time without overshoot and concluded as faster response to control the steam temperature compared to FLC and PID with PRBS input signal.