A non-minimum phase robust nonlinear neuro-wavelet predictive control strategy for a quadruple tank process

In process industries model-plant mismatch is a significant problem. Quadruple tank process (QTP) can be configured both in minimum phase and non-minimum phase (NMP). However, in NMP, the control of QTP poses a challenge. This paper addresses that and presents a novel robust wavelet based non-minimum phase control (NMPC) strategy for the challenging QTP using genetic algorithm to find the optimised value of the manipulated variables in NMPC at every sampling time. The QTP is modelled based on wavelet neural network. The simulation results indicate that significant improvements have been achieved both in modelling and control strategies for a QTP system compare to conventional approaches such as the Levenberg-Marquardt.