Application of Quantum Neural Network Based on Rough Set in Transformer Fault Diagnosis
Xianwen Ren, Feng Zhang, Lingfeng Zheng, Xianwen Men · 2010
With the conception of quantum mechanics, quantum neural network has a fuzzy character, and the fuzzy and uncertain datas can be distributed to different patterns, through which the uncertainty of pattern recognition is decreased. In this paper, the classified effect of quantum neural network has been used in the fault diagnosis of transformer. Firstly, the parameter space is mapped to the fault state space rationally by updating the connection weights, and macroscopic information is collected by a classifier. Secondly, the renewed quantum intervals are smoothed, and the uncertain data can be related to different types with proper ratios, through which the accuracy of pattern recognition is improved. Through combing with rough set effectively, quantum neural network can improve the speed by reducing the redundant samples at the same time. Finally, the quantum neural network is compared with BP neural network in dealing with the actual examples, and the validity and feasibility has been proved.