Answers are grounded in this paper and cite their sources.
Energy performance certificates (EPCs) are the principal instrument for assessing and communicating building energy performance in Europe, yet studies testing EPCs against measured energy compare, almost without exception, whole-building totals. This study examines what that practice conceals. In a census of ten recently completed municipal buildings in Sandnes, Norway, every floor area, meter scope, and calculation basis was verified against primary sources, and EPC-calculated and measured delivered energy were compared at three levels of data resolution: whole-building for all ten buildings, energy-carrier for one, and end-use for one. Measured energy exceeded the calculation in eight of ten buildings; the portfolio-mean deviation is +28% (range − 8% to +102%). At end-use level, a building with an unremarkable aggregate deviation of +17% (net energy) deviated by −38% to +110% on its individual end uses, opposing errors cancelling at the meter. A cancellation index, C = 1 − |Δ|/D, quantifies the effect: C = 0.66, two-thirds of the deviation magnitude invisible in the whole-building total. A verified floor-area error alone would otherwise have made the same building appear compliant. Whole-building agreement is therefore not evidence that a building performs as calculated. Certificate validation needs the end-use resolution that building performance evaluation has long applied to diagnosis, and the study shows the cost of proceeding without it.
No takes yet. Share an insight, caveat, or question.
Syse et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: