Design of Fault Location Algorithm for Distribution Network Based on FTU

Distribution network is mainly composed of power points, feeders, sectional switches, contact switches, feeder terminals and other parts, and its wiring structure is mainly determined by the closed state of each switch. Radial network is the main network structure of distribution network in China. FTU (Feeder Terminal Unit) is generally installed on the above switches in the distribution network to detect and control the situation in each area of the line. This paper presents a fault location algorithm for distribution network based on FTU. Firstly, FTU is used to collect fault current data and direction of distribution network, and then AI algorithm is used to realize fault location of distribution network. By applying GA (genetic algorithm) to BPNN (BP neural network), the initial weights and threshold settings in the original BPNN are improved to achieve the optimization goal. The research results show that the average distance error is $0.3467 \%$, the maximum error is $0.892 \%$ and the minimum error is $0.063 \%$ in seven fault location tests at different locations, which can meet the needs of practical engineering.