Application of Fuzzy Neural Networks to Vacuum Level Measuring

As the accidents are frequent in the process of vacuum level measuring of vacuum interrupters and the measuring results are with bad consistency. The measuring accuracy can´t be assured by only one measuring method. More than two measuring method are used in this paper, such as the method of magnetism-control discharges and the method of shield electric-potential. Fuzzy neural networks controller is designed for each method, to realize ¿vertical¿ fuzzy judge. Then data fusion method based on fuzzy neural networks is used to realize ¿horizontal¿ fuzzy control and to gain vacuum value of highly reliability. In the end, the ineffective data is eliminated based on fusion matrix theory and whether the transducer is invalid is judged. That is utilized as the fuzzy control´s assistant method.