基于深度神经网络和SoftMax分类器的台区负荷分类识别方法

Classification and Identification Method of Station Load Based on Deep Neural Network and Softmax Classifier

随着传统分类分析算法研究的不断深入,台区用电负荷模式的分类识别也在不断发展。提出了一种基于深度神经网络(deep neural networks,DNN)和SoftMax分类器的台区负荷分类识别方法,结合已有的典型负荷曲线特征库,实现对台区未知用户的负荷预测,为电网部门需求侧管理提供可靠的支撑。对某台区1200个用户负荷数据进行实证分析,结果表明,提出的分类方法在算法收敛性、计算时间以及预测精度等方面具有更好的性能。

With the continuous in-depth research of traditional classification analysis algorithms,the classification and identification of load patterns in power station area are also developing.Based on deep neural networks(DNN)and SoftMax classifier,a station load classification and identification method was proposed.Combined with the existing typical load curve feature database,the load forecasting for unknown users of the station areawas realized,which provides reliable support for DSM of power grid departments.The load data of 1200 users in a certain station were analyzed empirically.The results show that the proposed classification method has better performance in algorithm convergence,calculation time and prediction accuracy.