Auto‐Cluster Quarter Hybrid Classification Algorithm Based on Instantaneous Active and Reactive Power

ABSTRACT Nonintrusive load monitoring (NILM) load classification algorithms' performance and effectiveness are evaluated not only based on their classification accuracy but also based on their computational cost. In recent literature a lot of hybrid approaches have been proposed in order to enhance load classification accuracy. These methods simply combine different NILM algorithms to enhance load classification accuracy, without taking into consideration the increased computational costs. In other words, these methods improve load classification accuracy but at the expense of speed and efficiency, which make them impractical to use in real life. This paper provides a review and a categorisation of hybrid methods in NILM load classification and proposes a novel hybrid approach based on the instantaneous p‐q load signature, which considers different characteristics of load signatures with the goal of computational cost reduction, which in itself can be classified as a new category of hybrid approaches.