The Research And Implementation Of Short-Term Load Forecasting For Local Electric Power Network

As an important component of Energy Management System (EMS), short-term load forecasting (STLF) is a hot research field of power system. Efficiently and accuratly short-term load forecasting help to economically arrange for the power generating unit commitment, relats power transmission overhaul plans and appropriate measures can also be taken for peak shifting and averting. Thus, we can effectively reduce the cost of electricity and maintain the power grid operation secure, economical and stable. With the gradual formation of the electricity market, STLF will become more importance.Since the regional power grid construction is different, each regional load has their characteristics. This paper studies the load characteristics and future trends of an appointed region and analyzes the multiple factors which affect load changes. By comparing the relationship between electric load and meteorological factors, basis is provided for the forecasting model establishment, and the foundation was laid for achieving forecast.In the process of building models, the evidence theory was applied to choose similar days in short-term load forecasting. By using the speciality of the evidence theory that judge and reason after formal uncertain knowledge of objective problems from the uncertain reasoning perspective, similar days are quantitativly studied and selected. At the same time the paper proposed a combinational forecasting model with alterable value based on similar days. In the model, weight distribution laies particular stress on the forecasting effect of virtual forecasting days which has higher-degree similarity with the characteristics of days to be forecasted. Thus the advantage of the model was integrated.The article finally introduced the practicality and feasibility of the models through a practical example and the design process of the short-term load forecasting system.