【摘要】 由于影响负荷的随机因素很多,而传统负荷预测方法-周期外推法主要考虑前几日同时段负荷瞬时变化的规律,故存在局限性,预测精度不高。从周期外推法出发,提出了一种基于神经网络的外推法修正模型:通过BP网络进行训练,找到随机因素(如天气等)对负荷的影响因子,然后将其作为周期外推法模型的修正项。此外,还对修正模型的结果进行了分析和调整。实例证明,此方法在一定程度上克服了周期外推法的缺陷,提高了短期负荷预测的精度。更多还原

【Abstract】 As a result that the load is influenced by many random factors,period extrapolation mainly considers the instantaneous rule of loads change,so the precision is not good with its localization.Based on general period extrapolation,a new method was proposed,which error was modified by BP network.Furthermore,the result of the modified model was analysed and modified. The example here shows that this method overcomes the disadvantage of general period extrapolation and improves the precision of the short-term load forecasting.更多还原