Application Of Improved Gray Model In Mid And Long Term Load Forecasting

The continuous innovation and improvement of load forecasting technology is the guarantee for the long-term development of power system in the contemporary society.The scientific data provided by medium and long-term load forecasting is an important basis for formulating the power development plan.Based on the "small sample" "poor information" "uncertainty" characteristics of gray model,this paper studies the application of this model in the medium and long-term load forecasting and makes improvement,to explore the best method to improve load forecasting accuracy.This paper elaborates the background and significance of the research and analyzes the current situation of power load forecasting and gray theory,comparing the shortages of load forecasting methods.Because characteristics of gray system theory fit the medium and long-term load forecasting research object,and then to study its application in medium and long-term load forecasting.Due to the existence of many defects in the traditional gray model,this article aims at improving these defects and studying the reasonable improvement algorithm,building a gray prediction model with residual error correction and dynamic equal dimension prediction model.The gray prediction model with residual error correction gets the residual of the model value and the actual value superimposed correspondingly on the traditional gray model,improving the original model fitting ability.Dynamic equal dimension prediction model supplements dynamically models predict new information by eliminating outdated information,reducing computational complexity due to random changes in the forecasting process to improve model accuracy.On the basis of this,the idea of residual error correction technology is integrated into the dynamic forecasting model,and the new improved model-residual error correction model based on dynamic equal dimension is reshaped.The model expands the applicability of the original model and gets more efficient prediction model parameters,greatly reducing the impact of random interference factors,which is another manifestation of the practical value of the gray model.In order to compare and verify the predictive effects of the three models,this paper selects the total electricity consumption data of A region in 2008-2016 to make example analysis for three kinds of improved gray model and prior to the traditional gray model prediction.In this paper,using VC++ program programming achieves medium and long-term load forecasting process by the gray model.The prediction results are simulated by MATLAB,and the improvement of the predictive effect of the improved grey model is verified by comparison.