A modeling and forecasting method of regional load of power systems based on trend decomposition

A modeling and forecasting method of regional load of power systems is proposed, which can consider the influences of various factors, such as temperature, rainfall rate, festivals, etc. The whole regional load is decomposed into two parts, or trend load and non-trend load, and then the models of these two kinds of loads are constructed respectively, which as a whole constitute the model of regional load. In detail, the trend load is used to describe the change quantity of load in accordance with the trend, and it could be calculated directly based on the historical load data without considering the influences of various factors. Comparatively, the non-tend load is used to describe the change quantity of load which is inconsistent with the trend, and it could be calculated by analyzing the relation between the non-trend change quantity of each factor and the corresponding non-trend change quantity of the load caused by this factor. In the paper, the model of non-trend load of each factor is named as the revisional model, and the revisional models of some dominant factors (temperature, rainfall rate and festival) are established by the models of ARMA, rough set, multiple proportions, respectively. Meanwhile, choosing an area of Jiangsu Province as an example, the proposed method is realized and the load forecasting is performed based on the constructed model. The results demonstrate the validity of the proposed method.