An Approach to Observability and Controllability Analysis of Nonlinear Plants on the Basis of TSK Models

摘要

Most industrial plants are nonlinear, multivariable, inertial and with model uncertainty. They are difficult to model using classical approaches and thus their observability and controllability necessary for the design of the controller are hard to analyze. The aim of the present research is to derive conditions for the analysis of the observability and the controllability of nonlinear plants, represented by state space Takagi-Sugeno-Kang (TSK) models. The main results are a simple and general approach to observability and controllability study of nonlinear plants, which is based on equivalent linear systems and illustrated on a two-variable nonlinear plant - a laboratory two-tank system. The TSK plant model needed can be derived from an existing nonlinear plant model or applying a suggested procedure for development of modified transfer-functions-based TSK models from expert and experimentation data.