Research On Power System Short-Term Load Forecasting Method Based On Neural Network

Power system load forecasting based on the case of a certain precision, consider the system operating characteristics, natural conditions, social conditions, regional economic development and other important factors. It uses the historical load value date which after a series of mathematical calculations to decide the particular moment load value in the future. Short-term load forecasting is the daily power load forecasting which form 1 day or 1 week in advance. The key issues of power Load forecasting is the technical problems and mathematical modeling problems of the forecasting. When the non-mathematical model method that based on neural networks, expert systems, artificial intelligence and fuzzy theory has been proposed, thus there have an effective method to solve the complex load forecasting problem of modeling.This thesis studied the issue of short-term load forecasting by neural network, and it improved the BP network and the Elman network forecasting method based on the original. With considering the impact of temperature on the load, it analyses the factors on the local load characteristics using historical data of a regional power grid in Heze City. Programming by MATLAB software, it establishes networking model and achieve simulation prediction, and the forecasting error of different networks mode and input nodes mode were analyzed and compared. At the same time a user interface (GUI) is also created to achieve human control, so the implementation of the control and analysis become more convenient.