Research On Forecasting Model Of Power Demand Fluctuation Trend And Turning Points

Influenced by the economic cycle and some other factors, power demand in China presents the obvious cyclical fluctuations during the growth process. Correctly grasping the rule of power demand cyclical fluctuations, and forecasting future power demand fluctuation trend and turning points are significant to instruct the decision makers to make the corresponding strategy complied with the cycle, so that the power industry could realize the continual, stable and healthy development.Now the research on the forecasting of mid-long term power demand rarely take the periodic fluctuations into account, which makes the prediction results can not reflect the fluctuation rule of power demand growth rate; while the research on the power demand cycle mostly focus on the qualitative analysis of fluctuant features and fluctuation factors, which does not involve the forecasts of power demand fluctuation trend. As a result, the growth of installed power generation capacity is always mismatching with the growth of power demand in our country, which leads to phenomenon of the supply falls short of demand or demand exceeds supply. Aiming at solving this problem, this paper firstly analyzes the variations of power demand growth rate to get a whole understanding of the power demand fluctuations features and reasons in our country. In addition, based on the spectrum estimation can discover the frequency-domain feature of time series, while the maximum entropy method is suitable for the spectrum estimation of the short series with low frequency, and ARMA model is perfectly for the prediction of the random time series, the paper proposes the forecasting model of power demand trend and turning points based on the maximum entropy and ARMA model. The example results tests the feasibility and effectiveness of model in the prediction of the power demand fluctuation trend and the turning point in China. This paper finally predicts the power demand fluctuation trend and turning point of our country during the twelfth five years by using the model. Combined with the specific influence factor analysis, the paper still predicts the possible deviations between actual power demand growth and the forecasting results, which could provide a certain reference for decision-making of mid-long term electric power planning work.