Research Of Short-term Load Forecasting Method For District Power Grid

Load forecasting especially short-term load (STLF)forecasting is important for the Power Running & Dispatching System, it's also a important sign of the power system running management modernization, the importance of it went without saying. There are many influencing factors for STLF, besides the inherent law of load, the external factors playing a important role. Be one of them, the meteorological factor is seen as the most important.In this paper, the characteristic of short-term load and it's influencing factor are analyzed firstly, considering the number of influencing factor is large, and the short-term load as a completed nonlinear system. The artificial neural network (ANN) is a excellent modeling tool for nonlinear system, so the ANN is chosen as the modeling tool for the STLF.As the most important influencing factor, the knowability of meteorological factor is lags behind the load, especially in the season in which the load is sensitive to the weather. In addition, the load is influenced by the weather previous days, it can be seen as an embodiment of the accumulation effect of temperature. The study on the accumulation effect of temperature is not comprehensive at present, the accumulation effect of temperature is not considered in much theory of short-term load forecasting, and the influence degree of accumulation effect on load, the number of days that accumulation effect works are not clear even if the accumulation effect of temperature is considered. As the diversity of load and grid structure, the degree of accumulation effect on load is changing. This is the research direction in this paper. Based on this point, this paper puts forward a short-time load forecasting methodology considering the accumulation effect of temperature in summer on this basis. Not only consider the day-type, precipitation, temperature, and other related factors, but also account into the effect of temperature of the other day in the case of continuous high temperatures.The STLF for holiday is the difficulty of the STLF. Because the particularity of holiday load, the number of holiday load sample data is small, and the regularity of holiday load is more difficult to grasp than the work days. Aiming at this problem, a method for holiday short-term load forecasting based on improved similar historical day data is presented by this paper.Based on the methodology and model in this paper, considering the demand of power market, simple STLF software is designed.Although the STLF is facing many problems, because the prospect for the development of STLF is great, in the end of paper, the development direction of STLF is discussed simply.