一种基于图信号处理的非侵入式负荷分解方法

A Non-intrusive Load Decomposition Method Based on Graphic Signal Processing

非侵入式负荷监测分解,指将用户总电能消耗分解为单独电气设备的消耗,从而有助于建立能源互联网,实现节能减排。提出一种基于图信号处理方法,采用目前大部分电表量测的有功功率为负荷特征。使用由图索引的离散信号表示负荷特征,并进行阈值筛选,数据聚类和特征匹配。通过对实测低采样频率的负荷数据集进行仿真分析,验证算法具有较好的识别率和鲁棒性,且可作为嵌入式系统的软件嵌入到智能电表上,而无需附加硬件成本。

Non-intrusive load monitoring and decomposition refers to the decomposition of total power consumption of users into consumption of individual electrical appliances,which could facilitate the establishment of energy Internet as well as realization of energy conservation and emission reduction.Based on graphic signal processing,a method was proposed to adopt active power measured through most electric meters at present as load characteristics.Discrete signal indexed through graphs was used to represent load characteristics,and threshold screening,data clustering and feature matching were performed.Results of simulation analysis of actual load data sets with low sampling frequency verified that the proposed algorithm had quite good recognition rate and robustness.It could be implanted into the smart electric meter as software of the embedded system without additional hardware cost.