Load Disaggregation using Graph Signal Processing

In the era of the smart grid, energy disaggregation systems Non-Intrusive Load Monitoring(NILM) become the leading edge in the domain of Demand Side Management(DSM) which intends to provide a path for consumers to grab knowledge about their overall energy usage patterns and to devise appropriate energy strategies with limited smart meter installations. This paper proposes an initiative method on NILM called Non-Intrusive Device Level Load Classification(NIDLC), which classifies the appliances based on the load pattern using a technique of Graph Signal Processing(GSP). The data for load classification is collected from the embedded system built with the raspberry pi and made available in the Applied Electronics Laboratory of the EEE department of Thiagarajar College of Engineering, Madurai. The outcome of the proposed disaggregation is found to be significantly better than the values that were anticipated from the conventional process.