Self-tuning of predictive controller used in automatic controller - uses simulation of controller to train fuzzy logic expert system to set parameters of adaptive controller

摘要

The controller has an expert system using fuzzy logic making use of a database, rule base and an inference engine. The predictive controller is modelled as an equivalent closed loop regulator. The output of the process (Y) is compared with a set value (W) generated by the predictor to alter control parameters under a quadratic error criterion. The operation of the system is simulated and the results used to determine the stability margins and the robustness for different values of the parameters. The results are then used in the database and the rule base. USE/ADVANTAGE - Allows setting up of controller without aid of specialist knowledge.