A statistical modeling approach to detect anomalies in energetic efficiency of buildings

• We propose a methodology to support decisions in the control of energy consumption.
• We obtain an electrical energy consumption model in the building.
• We define an efficiency index to control the installations and detecting deviations.
• We apply the methodology to a set of public hospitals.

This paper proposes a methodology that provides a tool to support decisions in the monitoring of electrical energy consumption. The purpose is to help the manager of a set of non-residential buildings, such as, for example, factories, hospitals or university buildings, to objectively compare the energetic performance of all the installations under its responsibility and take measures when necessary. Firstly, an energy consumption model is developed considering the variables that have an influence on energy consumption in the building and processing them using the appropriate multivariate statistical techniques. Then, an efficiency index is defined which allows the manager to monitor the evolution of the installations and detecting deviations that later can be investigated, from an energetic efficiency point of view. A case study is presented where the developed methodology is applied to the public hospitals in the region of Castilla y León (Spain).