The Research Of Electricity Demand Forecasting Model——Based On The Statistical Data In Anhui Province

Electricity demand forecasting is the basis of optimizing and dispatching power system, so improving the prediction accuracy is significant for power industry and even economic development. Anhui Province is now in the period of economic transformation when the power structure is better than before, but the electricity demand is affected by other external factors, such as economic structure and weather factors, besides economic growth. And the weather factors have the maximum impact on the short-term electricity demand. In the past, weather factors cannot performance well in the long-term electricity demand forecasting. Now, by studying the mechanism weather factors impact on the electricity demand, we can establish appropriate forecasting model to improve the prediction accuracy.Firstly, this paper analyze the power structure and the economic structure of Anhui province since1990to qualitative their relationship. The conclusion is that Anhui is in the period of heavy industrialization but the power efficiency of the secondary industry is higher than the third industry which is the irrationality of our industry structure, so we need to continue internal upgrading of the economic structure to reach the cost-efficient economic growth. Secondly, the paper identifies the main factors affecting the electricity demand in Anhui by the principal component analysis method:Macroeconomic situation; Living standards; weather factors. Then, it make a case study in temperature to analyze the mechanism weather factors impact on the electricity demand. The paper conclude that influence of temperature on electricity consumption have three grades:The temperature base higher, unit temperature brings more electricity.Based on the conclusion of descriptive analysis, it separate the weather power and economy-related power from total electricity consumption by the method of trend decomposition which is applied in the study of macroeconomic fluctuations. By the comparison of three methods:separating the time trend from time series, seasonal decomposition, H-P filtering method, the paper choose seasonal decomposition as the weather electricity isolation method of the monthly electricity consumption data in Anhui Province.The weather power describes short-term fluctuations in consumption and the weather factor which affects electricity consumption changes every month, so we establish every single weather power forecasting model for each month. The economy-related power has steady growth trend, so we forecast it by fitting the trend equation. The prediction accuracy of this merge model is so high that we can reasonably believe that the separation of weather power is effective in the short-term electricity demand forecasting.At the end of this paper, we forecast the electricity demand of Anhui in the next three years by the long-term equilibrium models which is the traditional method of long-term electricity demand forecasting. The traditional model predicts very well. According to the forecast results, the electricity demand will keep the upward tendency and the elasticity ratio of electric power consumption will surpass1since2013because Anhui province is in the later stage of the process of industrialization. Besides, the macro-economic policy of "steady growth, transformation of export growth mode, adjusting economic structure" at present will lead to the phenomenon that power consumption growth pulls ahead of economic growth and this is a positive signal for Anhui province to speed up the industrial restructuring.