摘要
电力系统短期负荷预测直接影响电力企业的经济效益。对此,选择预测日前一天的电力负荷和阴晴、温度、湿度等气象特征数据作为网络的输入,预测日当天的电力负荷作为输出,建立了电力系统短期负荷BP神经网络模型。用历史负荷数据作为训练样本,对BP神经网络预测模型进行训练,用训练好的神经网络进行电力系统短期负荷预测。用真实历史数据对新疆某地区进行了电力系统负荷短期预测,结果表明,预测结果与实际值比较接近,1 d96个采样点的负荷预测平均准确率为98.45%。
Abstract： Short - term load forecasting for electric power system directly influences the economic benefit of electric power enterprises. The short - term load forecasting is established based on BP neural network for electric power system with the power load and meteorological data including weather, temperature and humidity before the forecas- ting day as input, and the power load at the forecasting day as output. The history load data is adopted as training sample to train BP network forecasting model. Then the trained network is used in short - term load forecasting. Besides, the real history data is used in the short -term load forecasting for a region in Xinjiang. The result shows that the forecasting is close to the practice with an average forecasting accuracy 98.45% of 96 sampling points in one day.
出处
《黑龙江电力》
CAS
2012年第6期439-441,445,共4页
Heilongjiang Electric Power
关键词
电力系统
BP神经网络
短期负荷预测
electric power system
BP neural network
short -term load forecasting
