Research On Line Loss Rate Forecasting Using Echo State Network And Optimizing By PSO Algorithm

The line loss rate of power grid is a comprehensive evaluation index.It can not only evaluate the rationality of the physical structure and operation mode of the existing power grid,but also reflect the technical strength and management level of the power enterprises.The "13th Five-Year Plan" required the upgrading and transforming distribution network,and put forward specific requirements of line loss rate of distribution network.The research on line loss in distribution network has become a hot topic in the power grid enterprises.At present,the power loss calculation and prediction have made a lot of research and achieved remarkable results,but mostly concentrated in the transmission network.Due to the complexity of its own structure and the relatively backward objective conditions of the distribution network in China,the research is still in the stage of development.In this paper,the traditional line loss calculation methods are investigated and the advantages and disadvantages of each calculation method are analyzed.This paper also introduces the definition of distribution network,and analyzes the structural characteristics of distribution network.Combined with the development of new information technology and the characteristics of distribution network,three feasible schemes for line loss prediction of distribution network are proposed.The work flow,data requirement and data collection channel of the three schemes are described.This paper also analyzes the characteristics and application scenarios of the three schemes.According to the current development of distribution network,this paper focuses on the research of line loss forecasting based on electric quantity prediction.By comparing ESN and BP neural networks through the experiment,ESN is selected as prediction algorithm in the prediction scheme.By analyzing the key parameters affecting the performance of ESN,a PSO-ESN prediction model is established,which forecasts line loss rate use ESN and optimizing by PSO algorithm.The test results show that PSO-ESN line loss prediction model performed well,and is suitable for the line loss forecasting work of distribution network.According to the prediction model,a line loss forecasting system is developed to be convienient the user,which includes predicting the quantity of electricity and predicting line loss.