Medium-and Long-Term Power Load Forecasting Based On Cloud Model And Load Characteristic Analysis

The contents of this thesis is part of the science and technology project,"Shuozhou Grid Load Characteristic Research and Analysis", hosted by Shanxi Electric Power Company. With in-depth study of the electricity market in Shuozhou, this thesis elaborates the status quo of the economic development of this region, as well as analyzing its load distribution and supply capacity of critical loads Through the analysis of the load characteristics of the grid, this study aims to grasp the regional load variation, and to give actual parameters for the grid operation. It may have a great significance on the development planning, stability and economic operation of the grid by using the actual data and a variety of methods to predict the Shuozhou grid load.This thesis firstly analyzes the variation of the load characteristics in this region in terms of the year, month, day and a typical industry. Besides, the study also calculates and analyzes the annual and seasonal load characteristics index, focusing on the electricity law and influencing factors in Shuozhou, in order to effectively improve the grid stability and loading rate in the region. Secondly, based on the analysis of load characteristics, this thesis introduces the medium-and long-term load-forecasting model, such as improved gray prediction model, the cloud theory prediction model and the combined forecasting model, focusing on the latter two models. In view of the fact that the current long-term load forecasting is influenced by many factors, this study uses a new theory—cloud theory to establish the uncertainty-predicting model, effectively converting the concept qualitative value and the digital quantitative value. With the combination of the model and actual situation, it can apply to the medium-and long-term load forecasting. Taking the prediction of the load in Shuozhou in2010for example, its error rate is4.8%. In order to reduce the incompleteness of single prediction and the forecast defects caused by itself, the study proposes a combined forecasting model, and uses two methods to predict, making sure the accuracy is higher than96%.Finally, based on the analysis of load characteristics and the forecasting model, the researchers developed the load characteristic analyzing system and medium-and long-term load forecasting software using Visual Basic6.0researching platform in order to meet the needs of the science and technology project of Shuozhou Power Company. This software successfully connects VB and the Access database, as well as VB and the MATLAB, and packages all of the method libraries.The software has passed the inspection and evaluation, and has been applied into practice in the power company. This software is practical, easy to install, simple to operate, and can be promoted in a certain extent.