Research Of Load Forecasting Method Based On Regression Analysis In Distribution Network Planning

Load forecasting is all important research content of power system planning and running, and is a premise for reliable suppling and economic nmning, and also is the basis ofpower system planning and construction. Exact degree of load forecasting shall directly affect rationality ofinvestment, network layout and running. This paper introduces some load forecasting models and its application in urban power network. Based on the new characteristic of modern load forecasting theory, the importance and its development actuality are given briefly, then the modern load forecasting theory is discussed systematically.Second, on account of medium to long-term load forecasting,this paper introduces the commonly used model of the multiple linear regression, also is known as the least squared regression method. The paper discusses specially the harms of the multiple linear regressions in the modeling analysis and its corresponding solution. This raises the partial least-squares regression method (PLS).After that, this paper presents an GM-PLS method. It can take full advantage of the grey model, which requires few data, doesn't consider the distribution rule and changing trend, has convenient operation, checks up easily, and so on. Then, the auther uses the GM-PLS method to forecast the loads of the Hengyang city. The utility proves its feasibility.Finally, this paper discusses the merits and shortages of the method and the improvement of medium-long term load forecasting methods for the new situation under distribution network..