Evolutionary optimization of fuzzy systems for water level control in the steam generator of nuclear power plant

Stabilizing water-level of the steam generator in nuclear power plant is a very important problem since dynamics of the system has nonlinear characteristics. Nonminimum phase phenomenon known as ´swell and shrink´ effect in the steam generator is more salient in low operating power. The swell and shrink effect is one of main factors for frequent shutdown of the nuclear power plant. In this paper, a new framework for building an intelligent fuzzy controller for stabilizing water-level of steam generator is proposed. In the proposed framework, two fuzzy systems are used and those fuzzy systems are optimized using evolutionary algorithm. The first fuzzy system is a fuzzy model, which is used to approximate the input-output relationship of the real steam generator. After the fuzzy model is obtained, a group of fuzzy controllers is optimized by evolutionary algorithm while performing control test on the obtained fuzzy model.