Hidden Markov Model based non-intrusive load monitoring using active and reactive power consumption

This work presents a residential appliance disaggregation technique to help achieve the fundamental goal in Non-Intrusive Load Monitoring (NILM) problem i.e. simple breakdown of energy consumption based on the appliance type in a household. The appliances are modeled using Hidden Markov Model by utilizing both their active and reactive power consumption data. The data was recorded by attaching Power Standards Lab PQube measurement device to the appliances. Granularity of the power readings of the disaggregated appliance matches with that of the reading collected for individual device. The accuracy of the model is compared with other models developed using only active power consumption of the appliances. The results using the proposed method are more effective and are found to predict a better output sequence for the appliances compared to model using only active power for modeling loads --Abstract, page iii