基于径向基神经网络的需求侧响应用户基线负荷计算方法

Calculation method of radial basis function neural network based on user demand side response baseline load

用户基线负荷是需求侧响应事件中用户响应性能的评估基础。用户基线负荷计算方法应当满足实时监测用户响应性能的要求。本文提出一种基于径向基人工神经网络的用户基线负荷计算方法。该方法考虑非工作日和可能的非正常用电日的影响,选择工作日中的典型日和事件发生前用户的用电情况构成模型的输入向量,训练神经网络。用训练完成的神经网络对假设的系统紧急情况发生时段的负荷进行预测,得到用户基线负荷曲线。实践表明,所提出模型具有一定的预测精度,满足实时性要求。

Customerbase line load(CBL)is the basis of evaluationof customer responsive performance in a DR event.CBL calculation methodscan beused to the monitor customer response performance inreal-time.This paper presents a CBL calculation methods based onradial basis artificial neural network.This method considers the impact of non-working days and some date with abnormal load,selects the load in some typical days and in several hours before the event to consist the model input vector, and trains the network. And then predict the load in an assumption emergency event.The prediction will be regard as CBL.It shows that the model proposed in this paperreacha satisfied accuracy,and meet real-time requirements.