Research On Accuracy Of Power Load Forecasting Method Based On Fusion

Power load forecasting plays a key role in power system planning and efficient and safe operation,which is conducive to the realization of scientific power planning,improving the economic benefits of the power grid,and promoting the development of the smart grid.Although the performance of power load forecasting has been improved,the accuracy,real-time,and visualization of power load forecasting in smart grids need to be further improved.The development of artificial intelligence has expanded new research ideas and provided new technical support for power load forecasting.How to achieve more accurate forecasting based on limited historical power consumption data and new technologies has become a difficult and hot spot in power load forecasting research.This paper proposes a new multi-model fusion load forecasting model based on Residual Network and Attention mechanism optimization,which further improves the accuracy of load forecasting.The main innovations of this paper are as follows:(1)In this paper,a new prediction model based on the multi-model fusion of Residual Network and bi-directional long-term and short-term memory network optimized by Attention mechanism is proposed.The improved Residual Network is used as encoder for feature extraction,and the bi-directional long-term and short-term memory network optimized by the Attention mechanism is used as decoder for prediction output,Compared with other prediction models,this model has better prediction performance.(2)The data fusion model is constructed to improve the forecasting results of power load forecasting.The practical application of data fusion technology in power load forecasting is realized from three aspects of data fusion layer,feature fusion layer,and multi-model fusion,which improves the accuracy of load forecasting.(3)The method of feature extraction based on the Residual Network model is improved,and the Residual Network model of two-dimensional image classification prediction model is used in power load forecasting.Based on the advantages of Residual Network,the optimized residual model realizes the mining of deep features of power load data,enhances the fusion of artificial features and deep features extracted by the neural network model,and improves the accuracy of prediction.