融合实时因素的短期电力负荷预测方法

Short- term Power Load Forecasting Based on Data Processing

短期负荷预测对于电力系统安全经济运行有着重要的作用,支持向量机现已成功地应用在电力预测领域。提出一种基于实时气象因素的样本选择策略,首先利用日气象特征向量缩小样本集,然后基于实时气象因素利用FP-Growth算法选择与预测日相似的训练样本,最后建立支持向量机预测模型。最后通过实验表明,经过样本选择所建立起来的预测模型具有较高的预测精度。

Short-term load forecasting plays an important role in the economic operation of the power system security, the support vector machine （SVM） has been successfully applied in the field of electric load forecasting. This paper presents the sample selection strategy based on real-time weather factors. Firstly, we use the day meteorological feature vectors to reduce the sample set and then use FP-Growth algorithm selection based on real-time weather factors to select the training sample which similar to prediction day, and finally prediction model based on SVM is established. The experiment results show that the algorithm which based on sample selection has higher accuracy.