Power Grid Fault Diagnosis based on Support Vector Machine

Abstract

Power grid dispatching is developing in intelligent directions, and data-driven fault diagnosis has become a development trend for power grids. However, the current power grid fault diagnosis method is lack of data-driven model for power grid fault diagnosis. Based on the quantitative characteristics and time series distribution characteristics of alarm information, a fault diagnosis method of power grid based on Support Vector Machine is proposed. Considering the duration of a single alarm information, the time sequence characteristics of a single fault device and the information amount of the whole fault event, the fault type decision model and the faulty device decision model are constructed based on the Support Vector Machine algorithm. The alarm information generated by the simulation system is used for fault diagnosis. The experimental results show that this method can deal with different forms of alarm information and adapt to the complex working conditions of the engineering site.