Energy performance of buildings: A quantative approach to marry calculated demand and measured consumption

Panel: 8. Monitoring and evaluation: building confidence and enhancing practices

This is a peer-reviewed paper.

Authors:
Michael Hörner, Institut for Housing and Environment (IWU), Germany
Markus Lichtmeß, Goblet Lavandier & Associés Ingénieurs-Conseils S.A., Luxembourg

Abstract

Energy Performance Certificates (EPCs) for buildings are supposed to inform about the energy-efficiency of a building under standardised boundary conditions, such as room temperatures and climate parameters. In a way EPCs represent a virtual test rig for the building to demonstrate whether it complies with the requirements of the corresponding energy savings ordinance.

Measured consumption usually differs from the calculated demand in a typical manner, in particular when simplified calculation models as in EPCs are applied: For existing, energy inefficient buildings it tends to be lower, for refurbished or new buildings higher than calculated. Some call it prebound or rebound effect. Of course this deviation is not desirable since calculated energy saving and cost effectiveness of refurbishment measures are over-estimated.

These effects were analysed in two projects. The energy certificate database for residential buildings in Luxemburg, run by the Ministry of the Economy, contains 20,000 records, each consisting of 174 parameters per building, including the measured consumption. Well documented records consisting of hundreds of building parameters including measured consumption of 93 non-residential buildings have been gathered in the research project Teilenergiekennwerte von Nichtwohngebäuden (TEK) financed by the German Federal Ministry of Economic Affairs and Energy.

Statistical methods such as multiple linear regression and error calculus were applied to study the gap between measurement and calculation in detail. Significant, independent variables were identified in either project. Correction functions for an estimated consumption including a standard error were derived from these variables and the regression coefficients. The probability that the future consumption of the building will be in the specified range might be indicated in the EPC and contribute to a higher credibility of the document. This opens a new approach to calibrate calculated demands by measured consumptions.