采用神经网络进行电力系统短期负荷预测的一种降维方法

A Reduced Dimensions Method Applying Neural Network in Short Term Load Forecasting

提出一种采用神经网络进行电力系统短期负荷预测的降维方法。该方法将每天的L个时刻负荷数据降维处理为少数几个特征参数 ,利用人工神经网络预测待测日的这些参数 ,以及参照日负荷为依据进行负荷预测。该方法提高了预测精度和速度 。

A reduced dimensions method applying neural network is proposed for short term load forecasting.The method decreases 24 hours loads into 3 characteristic parameters and forecasts loads according to characteristic parameters and reference day's loads using neural network.The calculations show that the method is efficient for short term load forecasting.