Research On Spatial Load Forecasting Methods Based On Load Regularity Analysis

Spatial load forecasting(SLF)is a prerequisite for power system planning and design.Therefore,the high precision results of space load forecasting are of great significance to urban power grid planning and design.Compared with the system load,the space power load has obvious temporal and spatial distribution characteristics.Historical data is the foundation of SLF,the authenticity and accuracy of the data used in the forecast will affect the accuracy of prediction.When there are human disturbances in the power system due to communication and other causes,the historical load data will be mixed with many suspicious data.The emergence of these suspicious data may lead to difference between predictive model,the results and the actual level of the system,so that the prediction work is of practical significance.Therefore,to improve the credibility of the forecast results,the basic data for the analysis and processing is very important.It is of great practical significance to select the appropriate historical load to analyze the rationality,carry out power grid substation distribution and line corridor planning accurately,economically and rationally,so that SLF in the urban power grid planning to play a greater practical significance.This paper introduces in detail the research status of spatial load forecasting method,and analyzes the classification of various SLF data preprocessing methods about the advantages and disadvantages respectively.Load density index method based on load regularity analysis is proposed to introduce the load density coordination coefficient for the uneven distribution of classified loads.The relationship between the classification load density and the cell load was established by using the trend method to predict the target annual load density index,the reasonable maximum value of the cell load without intrinsic error is determined by obtaining the historical annual load density index,and the more accurate spatial load forecasting is realized.A method of determining the reasonable maximum value of cell in space load forecasting based on Ensemble Empirical Mode Decomposition(EEMD)decomposition is proposed to avoid the measurement data of cell load In the measurement,communication and other errors into the forecast results,resulting in lower accuracy prediction problems.It decomposes the cell load into a series of modular functions,and establishes a filtering mechanism,reconstructs the modular function which can represent the regular part of the cell load as the main component,extracts the maximum of the principal component as the maximum load of the cell value.Based on the correlation between electricity consumption and electric power load,a bidirectional prediction method of total load of urban power grid is proposed,which exploits theintrinsic relationship between the historical data and the power load data,and improves the accuracy and stability of the data.The method of determining the maximum value of cell load is introduced into the space load forecasting.On the basis of the grid load density index method,by analyzing the error trend of the prediction results and the calculation speed under different resolutions,the best resolution can be obtained,the optimal spatial resolution is selected to improve the accuracy of spatial load forecasting.