Non-intrusive Load Monitoring for Residential Customers Using Adaptive-Neuro Fuzzy Interface System and Fine Tree Classifier

Energy conservation has gained much attention among recent demand-side management strategies due to environmental concerns and rising energy costs. Load disaggregation can act as an enabler for energy conservation by informing about the individual energy consumption of various loads. Different studies have shown a significant increase in energy savings when consumers are aware of their energy-intensive appliances. This paper presents a Non-Intrusive Load Monitoring (NILM) technique for determining the appliance operating status from the aggregate demand data of a household. The proposed supervised learning model includes a multi-layer and multi-node adaptive-neuro fuzzy interface system for extracting the power consumption features of various appliances. A fine tree classifier learner is then employed for classifying the operating status of different appliances using the extracted features. The proposed technique was tested on publicly available United Kingdom Domestic Appliance Level Energy (UK-DALE) dataset. Different evaluation metrics are used to evaluate the results of the proposed technique and to compare its performance with other recent methods. The comparison shows that the results of the proposed technique are comparable with existing NILM techniques available in the literature.