基于BP神经网络的下一交易日出清电价预测

Method for forecasting next-day market clearing prices based on BP neural network

利用历史负荷和清算电价对未来时段出清电价采用3层BP神经网络模型进行预测,将出清电价看成一个多输入单输出系统。把一天中每个小时按输入影响大小分成五类,然后采用BP网络分别建模预测。并采用美国New England电力市场2002年的电价数据进行了训练和预测分析,最终得到比较理想的出清电价预测结果。

Forecasting the market clearing price （MCP） is the most essential task and the basis for any decision-making. The paper is to use the the historical load and price to forecast the future price. The structure of the neural network is a three-layer back-propagation（BP）network.The MCP was considered as a multi-input and single-output system. Every hour in one day is divided 5 types according to input＇s influence,Then,it establishes model and forecasts using BP network .The historical loads and prices of 2002 in next-day America New England energy market are used for training and forecasting.Results show that the proposed method is effective.