Non-intrusive Load Disaggregation Based on Deep Learning and Multi-feature Fusion

Abstract

Non-intrusive load monitoring (NILM) is an important part of smart grid. In recent years, the deep learning method has been widely used in non-intrusive load dis-aggregation, but most of the current research only use low frequency active power signal for power disaggregation and does not consider the correlation of load power consumption patterns, which leads to load dis-aggregation can not achieve the desired effect. This paper presents a non-intrusive load disaggregation method based on deep learning and multi-feature fusion. In addition to the electric information of the load, the water and gas information of the load are also considered, and the correlation between the appliances power consumption patterns is studied. Finally, the performance of the proposed method is evaluated on the AMPds2 dataset. The results show that the proposed method can improve the load disaggregation effect.