The Research and Realization of Short Term Load Forecasting Based on Analysis of Load Characteristic

Short-term load forecasting is one important part of Energy Management System. And it is the premise and basic of safe, economic and reliable operation of power system. The forecasting precision will affect the economic benefit of power system. To study the load characteristic, the analysis theory, forecasting arithmetic and develop an efficient and practicable load forecasting system is becoming a very important issue currently. The paper mainly studies load-analysis, forecasting arithmetic and the design and development of the load forecasting system.The paper introduces the character of short-term load forecasting, and deeply analyzes the effect of temperature and rainfall to load. The paper creatively raises load K-line graph according to the stock K-line, and raises a way of dealing abnormal data using wavelet singularity detection theory and getting rid of burr employing curve smooth technology. This paper puts forward a preference combined model based on virtual forecasting and similar day which is selected by joint mining with fuzzy classifier and gray correlation degree. According to the comparability of special festival and weekend load, the paper raises a special holidays forecasting model based on similar weekend load, and amended by a property relation of the same special festival load to the corresponding similar weekend load in past years. The paper raises a forecast model for very similar day, and a standardization load model. Above arithmetic is tested by actual data of Hunan Province, which shows that they have a good practicability and feasibility.Finally, the paper introduces an integrative short-term load forecasting system which applies a popular structure of browser/server. And the system consists of five function models, namely load analysis, load forecasting, test management, etc. The application in Hunan grid has shown that the system is stable, convenient and accurate.