Statistical methodology to assess changes in the electrical consumption profile of buildings

Abstract

[EN] Many efforts have been made to define patterns, predict, and forecast energy use. However, changes in energy consumption may be studied in detail using various methodologies. This work presents a statistical methodology to assess changes in a facility consumption profile. Consumption patterns are obtained from a historical database of a predefined time interval, according to the type of day (day of the week, working or non-working), and an index that assesses change in the electrical consumption profile is proposed. Assessing these changes enables associating these values with possible events in a facility, which can serve to generate alarms in an energy management system, and reduce costs and maintenance periods. Additionally, a multi-criteria interpretation of the applied test table is presented that offers explanations and identifies possible causes of anomalous consumption. (C) 2018 Elsevier B.V. All rights reserved.