Short Term Load Forecasting Based On EMD Decomposition Method

Short term load forecasting has been attracting much attention of people. It is very important in the power sector. Short-term load forecasting is related to the power system operation and production scheduling, an accurate load forecasting will help to improve power system security and stability, and to reduce the cost of electricity. With the establishment and development of the electricity market, short-term load forecasting is playing an increasingly important role.The rule of changes of short-term load which is not regularity is based on the load data and affected by many factors. Because of the difference of the load components, it is useful to improve the forecast accuracy for research if we extract the separate load components from the total load when we analyze and predict the load. The research in this article is based on Empirical Mode Decomposition, EMD will be used in the power load forecasting, a forecasting method based on EMD decomposition is proposed.Firstly, the methods of short-term load forecasting being used currently are summarized in this paper and the advantages and disadvantages of each method are discussed, the composition and characteristics of power system load have been discussed in the article, and the establishment of empirical mode decomposition model has also been proposed, this model combine EMD and ARMA, using the ARMA to predict the each of IMF components.Secondly, in order to address short-term load forecasting of meteorological factors on the results of prediction, this paper discusses the factors to consider the temperature of power system load forecasting, use intervention analysis model to remove the temperature of the load from the original load .Then, a district in 2008 spring and summer electricity load segment as an example, by using EMD decomposition, the load component is divided into several components. In the analysis of the characteristics of each component and are classified, the use of ARMA models to predict the future of each component in the value over time, add these forecasts, and then load in that period by the total forecast. This simulation results show that EMD for short-term load forecasting, the prediction accuracy has significantly improved and enhanced.At last, this article take weather factor as a interference then use a established intervention analysis model to strip out the temperature of the load from the original load, take the purified load sequence to repeat the processing of EMD decomposition and the ARMA model of each IMF Prediction, then get the forecasting value of the purified load, compare the results of this prediction with the same period without the intervention, simulation and calculation results show that the accuracy of load forecast which treated by the intervention of short-term forecasting model has further improved and enhanced.The simulation results strongly validate the proposed method and prediction models, it can effectively improve the accuracy of short-term load forecasting, power system it is a useful short-term load forecasting attempts and discussion.