基于粗糙特征量的短期电力负荷预测

针对负荷特征一直是实际电力负荷预测中的重大问题。提出了基于粗糙特征量的约简算法。通过对天气及负荷历史数据进行挖掘,找到负荷的关键特征,并与径向基网络结合建立了负荷预测模型。算例结果表明,与按经验选取输入的传统网络相比,预测准确度有了明显的提高,更适用于电力负荷预测。

Short-term Load Forecasting Based on Rough Characteristic-component Algorithm

The key characteristic of mining influence the load is always an important problem in power load forecasting.A reduction algorithm through rough characteristic-component algorithm is introduced.The key characteristics of the date of weather and history load data are discussed,and then a model combined with radical basis function neural network is established.Forecasting results of calculation examples show that the forecasting accuracy is obviously improved and more suitable for short-term load forecasting compared with traditional radical basis function neural network model that chooses input parameters in the light of experience.