Non-intrusive Load Monitoring for Home Energy Usage with Multiple Power States Recognition

Active study has recently been conducted on Non-intrusive Load Monitoring (NILM) system to construct a cost-efficient household energy management framework. The conventional method, however, is not easily applicable to appliances that exhibit a complex operating mode or multiple states of power consumption. In this paper, we propose a NILM system that considers especially multiple power states of appliance. The proposed system profiles the power state of appliances based on the distribution of the real power and apparent power of the device. We utilize the state transition probability between power states, and use Hidden Markov model (HMM) for power state recognition.