【摘要】 在电力系统负荷预测中,组合预测是一种较为有效的方法。该方法通常是采用对单个预测模型进行加权处理,要求参加组合预测的模型误差能保持稳定,但电力负荷预测结果的误差往往是非均匀性的。针对上述做法存在的问题,提出了基于人工神经网络的组合预测模型,利用人工神经网络对复杂非线性系统的拟合能力,通过网络训练自适应地调整各种预测模型的权重。同时,为了避免用常规语言建立人工神经网络负荷预测模型存在的模型结构复杂、训练时间长等缺点,利用Matlab神经网络工具箱建立组合预测模型,该模型不仅编程简单,而且收敛速度快。算例表明了该模型的实用性和有效性。更多还原

【Abstract】 Combined forecast is an effective method in power load forecast,which usually adopts weighted processing of each forecast model and requires stable model error in combined forecast.But the forecast errors are not well proportioned in power load forecast.An artificial neural network based combined forecast model is presented,which automatically adjusts the weight of each forecast model by network training for complicated non-linear system.The Matlab artificial neural network toolbox is used to avoid the complicated model structure and long training time in establishing the combined forecast model.The model is not only simpler in programming,but also quicker in convergence.The calculation example shows that it is practical and effective.更多还原