Study And Application Of Mid-long Term Load Forecasting In Power System

In recent years, power system has become increasingly complex with the rapid increasing of power industry, separating of plants and grids, steadily booming of power market trading and smart grid in China. The change of power load become more diverse, it increased the difficulty of mid-long term load forecasting in power system work that all kinds of unexpected events on the political and economic. More accurate power load forecasting has become urgent and important task of an electricity worker. Accurate system load forecasting is not only related to grid's planning and design, but also to economic operation, fair trade and many other things. So it has become an indispensable research on 21 st century operation control system. At the same time, load forecasting is an important basis for planning and development of the power industry. The results of power load forecasting determine new capacity, distribution and time of plant generating units' installation in the future. It also provide a scientific basis to new construction, renovation and adjustments in transmission and distribution network. Power supply enterprise's mission is to provide energy which is safe, reliable, highquality, economical and eco-friendly. Supply and demand of load should always maintain homeostasis that it meet the requirements what is consumption and quality. We shall ensure a scientific load forecasting when planning and operation management. For which many experts and scholars tirelessly research on the power system load forecasting.This paper summarize the development of power system load forecasting as well as narrate future trends and achievements in this field in recent decades. The scholars have made comprehensive comparison of the merits of its application which each mature method in the last few years. This paper selects improved gray model theory, one variable number non-linear regression model and improved the elasticity coefficient method model that forecast mid-long term load in a county power grid, and evaluate the predictive accuracy of three models. The prediction results show that the composite model combines the advantages of three basic models, it greatly improve the prediction accuracy. Finally, this paper summarizes the problems of the research and application in mid-long term load forecasting, and makes the prospects for further research which combine its Shortcoming.