Research On Power Load Forecasting Of Intelligent Residential District Based On Diagonal Recurrent Neural Network

Load analysis is the premise and foundation of local power planning.Under the background of the development of the National Smart Grid,intelligent residential district will be the main form of the residential area in the future.Therefore,the research on the trend and characteristics of the power load will be a magnificent project.At present,the electric power enterprise carried on the correlation analysis and forecast to the electricity load of intelligent residential district,but the accuracy and the reliability are not high in the actual engineering application.Based on the study of the present main load forecasting methods,in this paper,we study the influence factors of the electric load in the intelligent residential district and deal with the data dimension,and then propose a data processing method based on diagonal recurrent neural network.First of all,we analyzed the characteristics of some variables in intelligent residential district,such as population,load.We analyze the influence of main factors on the power load of intelligent residential area,and establish the grey relational analysis model.We obtain the grey correlation degree between the main parameters and the load.The results show that there is a strong correlation between the power consumption and some internal factors and external factors,and there are some differences in the influence of the parameters on the seasonal,regional and temporal conditions.Secondly,In order to improve the processing speed of diagonal recurrent neural network,we propose a new method based on principal component analysis.By controlling the dimension of input data in 10 dimension,it not only reduces the amount of data and eliminates redundant data,but also improves the computing speed,and avoids the small amount of data and guarantees the accuracy of operation.On the basis of data reduction processing and influencing factors analysis before load forecasting,in this paper,we propose a method to deal with the electric load data of intelligent residential district based on diagonal recurrent neural network.In order to improve the accuracy and precision of diagonal recurrent neural network,We propose a model parameter optimization method based on particle swarm optimization algorithm,and the validity of the method is verified by a cell of an intelligent residential district.Through historical data,we train and build models which with the seasonal,regional,population and other factors as the input and with the power consumption as the output,in order to predictthe actual power consumption of the intelligent residential district in a certain period of time.Then compared with the actual power consumption statistics,We judge the prediction effect of the method.The simulation results of MATLAB show that the proposed method can predict the load accurately,and the validity has been proved.