Risk Identification Method of Power Purchase Agent for Grid Enterprise Based on PSO-BP Neural Network

• Published in: 2024 6th International Conference on Energy Systems and Electrical Power (ICESEP) pp. 902 - 908
• Main Authors: Wang, Xiaotian, He, Guixiong, Wu, Binbin, Chen, Ping, Zhang, Feixia, An, Qi
• Format: Conference Proceeding
• Language: English
• Published: IEEE 21.06.2024
• Subjects: Accuracy; Analytical models; Costs; Fluctuations; Interpretative structural model; Power purchase agent; PSO-BP neural network; Risk identification; Solid modeling; Training; Uncertainty
• DOI: 10.1109/ICESEP62218.2024.10651819

China promotes all industrial and commercial users to enter the power market, and grid enterprises become agent for those users who cannot enter the electricity market to participate in transactions for the time being. A study of the risk identification method of power purchase agent for grid enterprise is of great significance for the promotion of power market reform. First, the impact of uncertainty factors on the agency cost of power purchase is analyzed from the multi-dimensional aspects of load demand, power source structure, and market price, and the key influencing factors are revealed based on the explanatory structural model. Second, the BP neural network optimized by particle swarm optimization (PSO) algorithm is used to develop the agent power purchase risk identification method, and the extracted key influencing factors and agent power purchase cost risk categories are used as inputs and outputs to train and validate the model, respectively. In the final, the effectiveness of the risk identification method proposed in this paper is verified through Monte Carlo simulation for scene sampling. The results show that the accuracy rate of the PSO-BP neural network-based agent power purchase risk identification model is 94.3%. In conclusion, the research provides a reference for grid enterprises to reduce transaction risk and formulate a multi-market portfolio trading strategy.