Short-term Load Forecasting For Region With Abundant Small Hydropower Based On Its Load Characteristics

Power system short-term load forecasting is an important task of the day-to-day scheduling of the power sector. Accurate load forecast, from the economic point of view, can help power enterprises to reasonable arrangements for the start and stop of the unit maintenance scheduling, reduce unnecessary rotating reserves, thereby reducing the cost of power generation; From a social point of view, it can improve the safety and stability of the power grid operation, ensure the user’s power quality, so as to ensure society’s normal production and life. Therefore, the accuracy of short-term load forecasting has become one important indicator to measure the modernization of the electric power enterprise’s management level.Small hydropower resources are widely distributed in China, as an economical and environmentally friendly renewable energy, it has contributed to not only in the increase of energy supply, improve the energy structure, protecting the environment and other aspects, and still played a unique role in the electricity emergency support. but for many region’s power sectors with abundant small hydropower, because of the regularity of the generation load of small hydropower is very different from the consumption load over the whole society, it cause trouble to improve the accuracy of load forecasting. Therefore, study the load characteristics, and research the load forecasting models suitable for applying to the regions with abundant hydropower, has an important practical significance.This paper starts from the composition of the load of the regions with abundant small hydropower, and decomposes the load supplied by power network into load of the whole society and small hydropower load, explore their regularity and study their internal law and external characteristics, with the different characteristics of the load curve, to lay the foundation for the choice of load forecasting model and consider the various relevant factors. Then select the appropriate forecasting method respectively, to establish a load forecasting model which comprehensive consider the influence of small hydropower generation load, and achieve load forecasting delicately:For small hydropower load, according to the different characteristics of the wavelet coefficients after wavelet decomposition, use the Elman neural network to predict the amplitude significant variance component and similar day weighted average model to predict the remaining components; And for load of the whole society, use the BP neural network to predict. Finally, composite the predicted load of the whole society with the results of small hydropower generation load to restore the load supplied by power network which the electricity sector required.At the last of the paper, cases are given to demonstrate the effectiveness and simplicity of the proposed method, which show a practical value to improve the accuracy of network supply load forecasting for abundant small hydropower regions.