Non-intrusive load monitoring of air conditioning using low-resolution smart meter data

Abstract

Rapid growth of air conditioning (AC) has caused large effect on safe and economical operation of power grid especially in summer. The research on the AC load monitoring is of great significance for improving the stability of power grid. Non-intrusive load monitoring (NILM) is gaining attention due to many attractive features, such as low cost, low-complexity and easy promotion. A NILM of AC model using low sampling rate smart meter data is proposed in this paper. Firstly, preliminary recognition is performed to detect all possible events. Secondly, several features describing the distinct difference between AC and other appliances are extracted. Support Vector Classification (SVM) model for further detection is built with the extracted features. Finally, simulations on real-world data with low sampling rate from 5 houses are carried out to verify the effectiveness of our model. The results indicate that the switch statuses of AC can be effectively identified by the proposed model.