Application of T-S Fuzzy Neural Network to Pattern Recognition of Corona Discharge

Based on the T-S fuzzy neural network, the types of corona discharge are recognized. In this paper，three types of corona discharge test model are designed, then characteristic signals of corona discharge, the maximum, minimum and mean value and including the fractal dimension, are extracted, and then used as the input vectors of neural network, furthermore, accordance with the numbers of input vector and speciality of membership function, the T-S FNN topological structure is investigated and constructed, at the same time, the 4 neurons are selected used as input neuron and the 3 Gaussian functions are choose membership function, then the T-S FNN topological, which is trained and tested for this network, the structure is confirmed as the pattern recognition network in the types of corona discharge, and results show that the network is great effective for pattern recognition of corona discharge.