Characterization of current switching transients for appliance identification

Abstract

Demand Side Management is a key concept within the Smart Grid vision to promote energy efficiency, load flexibility and interaction between the consumers and other power grid stakeholders. Disaggregated information requires an advanced load monitoring system of individual appliance consumption. A smart house is envisioned to include a Nonintrusive Load Monitoring (NILM) system to support demand side management and motivate the users to adopt energy saving practices. NILM systems use input electrical measurements taken at the energy meter point of a house and estimate the individual appliance operation and consumption through mathematical algorithms. Each appliance can be distinguished from others through a set of particular attributes namely load signatures that can be computed from transient signals, steady state signals or both. This paper aims to characterize current switching transients, for NILM applications and to discuss how they are affected by variation of factors such as point on wave of switching, network impedance, supply voltage distortion and sampling frequency of the meter. For that purpose, measurements of residential appliances of several categories are acquired and processed. The conclusion of this work is the assessment of suitability, robustness and efficiency of appliance identification based on current transients.