A Short Circuit Fault Diagnosis Method for Power Systems Based on Data Physical Models

Cascade faults that cause power system collapse are usually caused by the further expansion of the harm caused by short-circuit fault events. To avoid cascading faults in the power system and ensure the safety of the power grid, it is necessary to effectively identify key information of simple faults in the early stages of the fault. This article proposes a method for diagnosing power system faults using a Data-Physical model, with the power system as the physical model and the BP neural network model as the data model. By utilizing the electrical parameters of non-faulty nodes to diagnose fault information in power systems, a fault set for diagnosing faults using non-faulty nodes is established. Then, a method using Zebra Optimization Algorithm to optimize the parameters of BP neural network is proposed to improve the matching degree between BP neural network and fault set. Finally, taking the 'IEEE30 node' model as an example for fault detection. The results indicate that compared with the unoptimized BP neural network, the method proposed in this paper has a more accurate ability to diagnose faults in power systems.