Multi-dimensional Reliability Evaluation Method of Distribution Network Based on Improved Neural

The in-depth construction of new power systems has raised higher requirements for the management of line losses in urban power distribution areas. Traditional methods for perceiving the effectiveness of distribution area management largely rely on the subjective judgment of staff members, making it difficult to achieve refined control over the management process. Therefore, to better adapt to the needs of line loss management under the background of new power system construction, an improved neural network-based method for perceiving the effectiveness of distribution area line loss management is proposed. Firstly, by comprehensively considering the multi-dimensional dynamic factors in the process of distribution area line loss management, an evaluation system for the effectiveness of distribution area line loss management is constructed from four dimensions: economic, safety, management, and low-carbon. Relevant indicators strongly correlated with management effectiveness are selected through correlation analysis. Secondly, an improved neural network with an adaptive learning rate is used to analyze historical data of distribution area line losses, obtaining a comprehensive perception model that reflects the effectiveness of distribution area line loss management. Finally, taking multiple distribution areas in a certain location in Shanghai as examples, the effectiveness of the indicator system and evaluation method is verified through simulation analysis, confirming the validity of the proposed method.