Fuzzy neural network water-mixed control system based on hybrid algorithm

A fuzzy neural network algorithm was proposed for the water-mixed system, for the system usually has the property of high nonlinearity, large time-delay, time-variant and it´s difficult to establish mathematical models. The genetic algorithm and the BP algorithm were combined together. The fuzzy membership function was optimized by the improved genetic algorithm which has ability of global searching. At the same time, the fuzzy neural network´s weights were optimized by BP algorithm which has ability of local searching. At the end, the systems will have better adaptability as well as better convergent speed. To make the floor heating system safer and steady a new method was proposed to change the frequency of the booster pump. The simulation results show that the proposed controller which can rapidly control the temperature is better than fuzzy controller.