Smart energy performance indicators for live historical and normative feedback systems




Journal Title

Journal ISSN



Volume Title





Peer reviewed



Communicating building energy performance to building users has been identified as a significant opportunity to support behaviour change. This research pursues the concept of continuous, automated feedback designed to support motivated building users to learn how their behaviour impacts building energy performance. Automated energy consumption data collection presents an opportunity to develop approaches for continuous feedback systems. However, energy performance is a complex notion and consumption data alone are not suitable to convey performance. In order to be of use, performance indicators designed specifically for providing feedback to building users must reflect changes in user behaviour which may be small relative to total consumption. A new building energy performance indicator is proposed based on the concept of continuous improvement. The indicator combines the benefits of historic and normative feedback by producing a normalised index of improvement or deterioration over time. The indicator is also well suited to communicating building energy performance in a user-friendly way. The indicator is based on a predictive consumption model fitted to data for a rolling baseline period. The scale of the indicator is defined in terms of the variation in baseline model residuals. This allows for a direct comparison between buildings on the basis of improvement or deterioration from the baseline performance. A direct comparison can be made even between very different buildings. A case study of five university buildings is presented. Consumption is predicted at half-hourly resolution using a variation of a standard variable degree day model. The indicator is calculated for each half hour beyond the initial 365-day baseline period on a rolling basis with a new baseline model being calculated each week. The indicator reflects even small changes to regular consumption patterns, both persistent and transient. Persistent changes are absorbed into the rolling baseline model after a few months. Critically the indicator is sensitive enough to detect small changes in consumption patterns and can be compared between buildings. As a feedback tool the indicator has the benefit of having a common scale and being comparable across buildings.



energy performance indicators, energy consumption modelling, energy feedback


Stuart, G. and Fleming, P. D. (2014) Smart energy performance indicators for live historical and normative feedback systems. Frankfurt: Improving Energy Efficiency in Commercial Buildings (IEECB) Conference 2014, pp. 400-414


Research Institute