Research Of The GPRS-based Short-term Load Forecasting System

Short-term load is one of the important works of scheduling, electricity plan and planning administration departments in the power system. To improve load forecasting technology favors reasonable plan for electricity management, grid operation maintenance and unit overhaul; and reduce electricity costs; favors rational planning of power construction; favors to improve the power system of economic benefit and social benefit. So the load forecast has become one of the important content of the power system management modernization.This paper firstly discusses the background of Short-term load system research, the status and the meanings of the research. Combined with GPRS wireless data transmission and the characteristics of the embedded system, using ARM9core, comes up with the scheme of power load system of embedded data acquisition terminal. It can collect the data for the load forecast in real-time and accurately, makes load data error greatly decreased, it laid a good foundation for short-term load forecasting. Established the short-term load forecasting model based on the BP neural network,and realized the forecast based on the historical load data. Detailed clarify the importance of the meteorological factors in short-term load forecasting. Mainly analyzed the factors on temperature prediction model of BP neural network, and compared the result with the predicted result of BP neural network models based only on historical load data, and the prediction precision obviously superior to the load forecast model which only consider the historical data.