Research On Short-term Forecasting Method Of Power Load In Liaoyang Region Based On Improved BP Neural Network

Short-term power load forecasting is very important for local power system operation and real-time regulation.With the development of market economy,on the premise of pursuing operation efficiency and improving economic benefits,the relationship between supply and demand in the market is closely related to the accuracy of power generation planning,so it is particularly important to forecast the power demand of users.Accurate power load forecasting has become an indispensable part of power bargaining and balance between supply and demand.In order to improve the accuracy of short-term power load forecasting,this paper studies the short-term power load forecasting model based on BP neural network and particle swarm optimization algorithm and its typical application in Liaoyang area.Firstly,the classification of power load and the characteristics of short-term power load in Liaoyang area are analyzed,and the periodicity,randomness and similarity of load are studied,as well as the temporal and nonlinear characteristics of short-term power load.This paper analyzes the short-term load forecasting of liaoyang region based on "datum date" power system,and clarifies the method and error of the current short-term load forecasting of power system.Pearson correlation coefficient analysis of load and influencing characteristic factors in Liaoyang area is carried out,which lays a foundation for load forecasting modeling in the later stage.Next,the algorithm of BP neural network and process were analyzed,and in liaoyang area as the background,on the basis of historical load data collection,analysis and study of the regional load characteristics,considering the factors such as weather,to quantify the selection of the initial input data and normalized processing,based on the BP neural network short-term load forecasting model of liaoyang region,And the simulation prediction experiment is carried out to predict the load value of each hour of 24 h in the summer of 2019,autumn and winter of 2020 in Liaoyang area.Compared with the current short-term load prediction results,the prediction accuracy has been improved,and the relative error of prediction is within 10%.Finally,aiming at the shortcomings of BP neural network algorithm,such as long iteration time,easy to fall into local extremum and slow convergence rate,particle swarm optimization algorithm is introduced to optimize BP neural network architecture.Application based on particle swarm optimization neural network algorithm for modeling and simulation,and with the application of BP neural network model of short-term load forecasting results compare and analysis,the prediction results show that the model based on particle swarm optimization neural network in electric power system short-term load forecasting has good nonlinear mapping ability,can promote learning precision and learning rate,The relative error of the forecast is within 3%.After analysis,it is concluded that the prediction result of the neural network model optimized by particle swarm optimization has higher accuracy,which can meet the prediction needs of the dispatching department of electric power enterprises and provide strong support for the real-time regulation of power system.