基于Fisher有监督判别的非侵入式居民负荷辨识方法

Non-Intrusive Household Appliance Load Identification Method Based on Fisher Supervised Discriminant

居民负荷监测是需求侧管理的重要部分,为实现居民非侵入负荷监测与跟踪,研究了非侵入式负荷监测系统物理架构与工作原理,提出一种基于家电负荷主用电特征的负荷辨识方法。以非侵入环境下不同家电负荷的实测电能数据为先验训练样本,利用主成分分析法对8种典型电器的负荷特征样本进行降维处理,得到最优辨识特征,获取综合评估值并结合样本排序图将负荷分为2类。结合Fisher有监督判别准则将2类数据投影到一维空间,实现不同类型负荷的分离。实验证明该方法对不同家电负荷分类和辨识具有良好效果,能够在获取负荷有效特征的基础上实现对不同负荷的有效分辨。

Load monitoring for residents is an important part of demand side management. In order to realize household appliance NILM（non-intrusive load monitoring） and tracking, this paper studied physical architecture of NILM system to see how it works, and proposed a load identification method based on main electrical feature of household appliance. Based on measured data of different loads, PCA（principal component analysis） was used to reduce dimensions of 8 types of typical loads＇ feature samples, and then optimal identification characteristics were obtained. Combining comprehensive evaluation value with sample sort diagram, loads were divided into two categories. Furthermore, feature space was projected into one-dimensional space to realize separation of different kinds of load with FSD（fisher supervised discriminant） method. Experimental results show that this method can effectively distinguish appliance based on effective load characteristics.