A Novel Self-tuning Control Method with Application to Nonlinear Processes

Abstract

Hammerstein models have been considered as a class of well-known nonlinear systems, which have been prove to be attractive for system modeling and controller design tasks. In this brief, we introduce a new control strategy for such kind of systems. Interestingly, the system uncertainties, including the input block description error, the linear subsystem's unstable zero property and the colored added noise issues, have all been considered. According to the modified cost function, the parameter adaptation law has been online implemented throughout the use of a robust estimator. Meanwhile, based on the parameter estimates, the control law has been designed for the compensation of the modeling mismatch which is caused by unmodeled dynamics estimation. A simple but rigorous proof has been given to illustrate that the nonlinear model based control system stability can be properly achieved based on some reasonable and practical conditions. Finally, the proposed controller has been used for a representative nonlinear system, that is, a continuous stirred tank reactor (CSTR) system. Comparison studies have been presented to show the wider applicability of the novel method than some existing ones.