Nonlinear Modelling for Time-Varying Water Temperature PID Control

This paper is proposed to model and control the water temperature of Steam Distillation Pilot Plant (SDPP) using nonlinear modelling and PID control, respectively. Pseudo Random Binary Sequence (PRBS) and Multi-Sine (M-Sine) have been implemented to this system to perturb the input of the process and excite the nonlinearity behaviour of system dynamic. The Nonlinear Auto Regressive with Exogenous Input (NARX) model structure was used to estimate and validate the temperature output model of SDPP. The best transfer function model was chosen from the simulated result of different NARX model structures based on high percentage of fitness, low value of final prediction error and the mean square error. Later, a PID controller was designed using Ziegler-Nichols and Cohen Coon tuning methods. The results presented that the Ziegler-Nichols tuning method demonstrated better performance than Cohen Coon for PRBS input in terms of rise time (T
_r
), settling time (Ts) and integral square error (ISE) values while Cohen Coon tuning method showed better performance compared with Ziegler-Nichols for M-Sine input based on T
_r
, percent overshoot (%OS) and ISE values.