Research On Short-term Load Forecasting Based On BP Neural Network

With the development of modern information technology,smart grid has become one of the main components of smart city.In order to make full use of smart grid,the ability of intelligent dispatching and planning power transmission is very important.In practice,many factors will affect power consumption,which requires the use of information fusion technology to thoroughly understand power consumption.Therefore,researchers have studied the methods of collecting information related to power consumption and various multi factor power consumption prediction models.In addition,comprehensive analysis and accurate power consumption evaluation are th e premise and basis for more robust and efficient power grid design and transformation.Therefore,it is of great significance to explore a prediction model that can effectively reflect the power consumption change and internal relationship in the fusion i nformation.Power consumption prediction based on neural network is a hot research topic in recent years,and BP neural network algorithm has been recognized as a mature and effective method.Firstly,BP neural network is used to predict power consumption,and many factors such as temperature,humidity,air pressure,wind speed and holidays are considered.The impact of each factor was evaluated for a deeper understanding.Furthermore,a combined model of ultra short-term power demand forecasting based on Improved BP neural network,chaotic optimization genetic algorithm and simulated annealing algorithm is proposed.In the development process of the model,many characteristic variables affecting load demand are deeply studied,and modified many times in sta ndard BP and GA to achieve high prediction accuracy.The model helps to achieve the global minimum in the search space and avoids the improper convergence that leads to the significant improvement of prediction accuracy.In addition,this study also focuse s on selecting the best input variables based on correlation analysis,which is of great significance to improve the prediction accuracy of power load demand.The trained model is successfully applied to the actual data,and higher prediction accuracy and faster convergence speed are obtained.This accurate short-term load demand forecasting can realize the economic dispatching of power through improved demand response,which is helpful to effectively determine the dynamic price in the deregulated power mar ket.The simulation results show that the model can be effectively applied to the actual power market,especially the smart grid.