Applications Of Improved Analytic Hierarchy Process In Mid-Long Term Forecasting

Power load forecasting is one of the important parts in power system planning. An accurate power load forecasting is beneficial socially as well as economically in power system. This thesis aims to express the result decision on the research of load forecasting. After study thoroughly on the decision theory of analytic hierarchy process (AHP), the improved AHP is employed to the decision of mid-long term load forecasting result.First, this thesis introduces the fundamental background and characteristic of AHP. Next, summaries of the application in generation side technology, transmission and distribution network planning, power load forecasting, etc are described in details. Qualitative information can be analyzed quantitatively through the scaling method and judgment matrix of AHP. Prior to this, the thesis emphasizes that AHP has an excellent application foreground in terms of power system planning, operation and the evaluation of power companies in power system market.Due to the shortcoming of traditional AHP, the thesis shows the efforts of research and study of the improved AHP, including its characteristic, fundamental background, consistent judgment matrix and weight solution method. Various qualitative and quantitative factors which can affect the development of power load are analyzed according to the interrelated prescript and some collected information.Focusing on a certain point, a recursion order hierarchy structure model of load forecasting has been built in this thesis. The collected results and fact show that forecasting precision can be improved by employing AHP on decision of mid-long term load forecasting result.The research and study on load curve forecasting methods are not carried out further; however, those are still very important in power system planning and power market. The author propose a new theory on selection of typical day based on the analysis of minimum distance cluster and interconnect degree, this method analyzed the daily load curve shape and load characteristic, found and wiped off the affection of aberration day by cluster analysis based on minimum distance, found out the typical day by analyzing and calculating the interconnect degree between every day and benchmark day. According to the available data in China, a new method named improved AHP of typical day daily load curve forecasting is proposed. By using generating operation of Grey System method, Mathematical model that is used to solute quadratic programming is set up, and this mathematical model has a convenient algorithm. Case study shows its efficiency.