부산대학교 도서관

In this paper, a hierarchical multi-source information fusion method for power grid fault information diagnosis is proposed. Firstly, this paper carries out feature fusion and decision fusion for wavelet packet transform features and power spectrum features of power grid fault information. At the same time, an unsupervised feature reduction method based on entropy measure is used to reduce the power spectrum features. The system realizes the rapid diagnosis of single fault based on the switching quantity of SCADA system. At the same time, the switching quantity information is integrated with the electrical volume information of fault recorder and WAMS. The simulation results show that this method can get accurate fault diagnosis results by fusing the information of electric volume and switching quantity. This method has a good application prospect in power grid fault diagnosis.