The Defuzzification Methods Influence on Fuzzy Control of Nuclear Reactor

Abstract

This paper deals with the influence of the defuzzification methods on the quality of fuzzy control. There exist many defuzzification methods in theory. Each of them has its own benefits and application areas. The scope of this paper is focused only to the defuzzification methods available in Matlab Simulink environment. These are compared and evaluated using the integral quality criteria. The fuzzy PI controller was designed with normalized universes to simplify the tuning of controller parameters. The nuclear reactor model has been used as a reference system during experiments. The nuclear reactor model has been constructed using the Point kinetics model and the Thermal model. Step response characteristics of the closed loop control system were evaluated using integral criteria to compare available defuzzification methods. Finally the achieved results were briefly summarized.