Application Of Combination Forecast Model In The Medium And Long Term Power Load Forecast

This paper is the study of how to make medium-and long-term power loadforecasting more accurate. So far, many scholars have proposed a variety of loadforecasting method. The results produced by different forecasting methods are not thesame. However, the accurate prediction of the load has a profound impact on the safeand economic operation of power systems and the development of the local economy.Therefore, the choice for the region’s forecasting model is essential.Medium and long term load forecasting is mainly the annual, quarterly units’forecast in the power system. This paper uses Jiangxi Province Gao’an supply electricitysales of quarters of2004-2011, utilizes combination of prediction methods in the regionof medium-and long-term load forecast. The combination forecasts are based onindividual prediction model. Select three single prediction models to be combinedforecast, which have high prediction accuracy and can better play the characteristics ofthe combination prediction. Giving each single model prediction the right weight is thekey technology of the combine forecasts. In this paper, according to the characteristicsof the data obtained, firstly choose the gray prediction model and linear regressionmodels as the base models, respectively apply the variance-covariance predictionmethod and the deviation prediction method to combine forecasting. Then equallyweighted average combine the two sets of predictions, establish the final combinationforecasting model. The combined model can integrate the results of a variety forecastingmethods, take full advantage of the information of a sample of each model, andeffectively reduce the risk of interference of each single model in the forecastingprocess by external objective factors. Finally compare the predictive value to the realvalue. The example shows that the combination forecasting method has a highprediction accuracy, and the error is very small. It has scientific and engineeringpractical significance that apply the combination forecasting method to middle and longterm load forecasting of Gao’an areas.