Neural Network-Based Precise Location of Short-Circuit Faults in Medium- and High-Voltage Cables

With the upgrading of China's power grid scale and voltage level, improving power transmission capacity has become a key issue to be solved for power grid development. In this study, a simulation model of medium- and high-voltage cable is constructed in MATLAB/Simulink. The short-circuit faults under cable-overhead line, copper-core-aluminum-core hybrid form, and different grounding systems are analyzed, and the fault current waveform is transformed into wavelet energy spectrum by wavelet algorithm to extract the fault eigenvalues, so as to construct the training and testing sample data sets, and train the fault accurate location model by using GA-BP neural network. Simulation results show that this method can realize the precise location of short-circuit faults and the error of ranging accuracy is small, which has important practical application value.