Research on the HHO improved BP neural network fault prediction algorithm for distribution networks based on weather factors

• Published in: Journal of physics. Conference series Vol. 2831; no. 1; pp. 12014 - 12019
• Main Authors: Xu, Yan, Dai, Zikuo, Zhao, Yi, Zhao, Chenxing, Li, Ming
• Format: Journal Article
• Language: English
• Published: Bristol IOP Publishing 01.08.2024
• Subjects: Algorithms; Back propagation networks; Fault diagnosis; Hierarchies; Network reliability; Neural networks; Parameters; Predictions; Reliability analysis; Weather
• ISSN: 1742-6588; 1742-6596; 1742-6596
• DOI: 10.1088/1742-6596/2831/1/012014

Weather factors have a substantial impact on the operation and maintenance of the distribution network, but the current system fault prediction platform mainly focuses on the operating characteristics of the system, that is, the internal parameters, and rarely considers the external parameters. Firstly, by studying the fault data of the distribution network in a city in northern China caused by weather factors, this paper obtains the main weather factors affecting the reliability of power supply in this area. Then, the Harris Hawks optimization (HHO) improved BP neural network algorithm is used to establish the main reliability evaluation models suitable for this region, and based on this, the fault probability of this region is predicted under the target weather. Finally, the prediction results are compared with the actual fault data caused by weather factors in this area. The results show that both the PSO-BP algorithm and HHO-BP neural network algorithm achieve better prediction accuracy than the BP neural network algorithm, and the HHO-BP algorithm has better convergence speed.