Second Order Volterra Filter for Appliance Modelling

Abstract

Availability of large quantities of residential electrical consumption data is bringing considerable attention towards load monitoring, load forecasting, load disaggregation and demand response. Load modelling is the first and most essential step in achieving all the above said tasks. Even though many appliance modelling schemes are presented in the literature, no considerably influential work has been done on modelling appliances under voltage fluctuating environment. Motivated by this fact, we present the design and analysis of a Volterra based appliance modelling scheme which can be used in a voltage fluctuating environment. Principles of Volterra filter, least mean square algorithm for Volterra filter coefficient approximation and applicability of Volterra filter for appliance modelling are discussed. Further, a case study is presented to validate and identify the performance of the model using a data set obtained from a real household. Obtained results show that, Volterra filter can be utilized as an efficient tool for appliance modelling in a supply voltage fluctuating environment. Finally, how Volterra filter modelling can be extended to achieve the non intrusive load monitoring task is discussed.