Research on Fault Localization Method of Flexible Distribution System Based on Deep Learning

This paper proposes a fault localization method for flexible distribution systems based on SSA-LSTM to address the problems of excessive line branches, random disturbances caused by distributed energy access, and significant fault localization errors in multi terminal flexible distribution systems. By analyzing the traveling wave attenuation theory of line propagation, find the nonlinear correspondence between the fault electrical quantity and the fault location. Then, build an LSTM network and use the SSA sparrow algorithm for iterative optimization of network parameters to achieve fault localization of flexible distribution lines. Finally, a multi terminal flexible distribution system model was constructed in PSCAD/EMTDC and fault location research was conducted. The experiment proved that this scheme also has advantages such as small positioning error and strong anti-interference ability in scenarios with multiple line branches.