Reseach On The Mid-Long Term Load Forecasting In Power System

Mid-long term electric load forecasting is the foundation and premise of power system security dispatch, economic operation, optimization planning and scientific management, its precision is quite important to each department of national economy, so research on mid-long term load forecasting method is valuable both on theory and practice. There have been a great number of efficient forecasting methods in this field of load forecasting. But, the preconditions of some classical methods can hardly be applicated in practice. And sometimes the parameters of some models have few definite physical meanings.This paper presents the features of electric load and the effects of electric power demand forecasting. In this study, regression model, principal component analysis, clustering analysis reveal the inner relationship between electric load growth and several factors such as economic factors and climate impact.The author proposes a method for power system load forecasting modeling based on variable structure co-integration theory. The established model reflects the equilibrium relation of long-term static state and short-term dynamic fluctuation between electricity consumption and economic parameters. To determine the point of discontinuity in time series, an approach based on Group Method of Data Handling (GMDH) modeling GMDH auto-searching structural change is put forward to avoid the possibility that the traditional endogenous methods maybe falsely decide the most outstanding break-point. By use of the information of discontinuity point the forecasting accuracy of the model in improved, meanwhile the pseudo-regression that ever-presents in traditional load forecasting and affects the modeling can be avoided. Using the proposed method the forecasting and analysis of annual electricity consumption in Shanghai are carried out, and the results show that the proposed method is available.Because of the complexity of the energy system, it is difficult to solve the problems comprehensively and accurately with only one single forecasting method, the advantage of the various methods can be integrated by the combination forecasting method, which can make the forecasting results more precise and stable. This paper proposes a combination forecasting method based on co-integration and GMDH. It is proved that this method can improve the precision in the load forecasting in Shanghai.