A new approach to assess the building energy performance gap: Achieving accuracy through field measurements and input data analysis

The performance gap is defined as the difference between a calculated and measured quantity, and in buildings, it may refer to the energy performance or the indoor thermal conditions. According to the literature analysis, most studies start from simulation results and define methods and approaches to minimise the discrepancy against the measured values. This paper presents an alternative and innovative approach to the problem, starting with measurements in a fully instrumented and monitored living lab consisting of seven office rooms used to build and validate an accurate calculation model. The model is applied to observe how different input modes of the most relevant parameters affect the performance gap. The model exhibits high accuracy: the coefficient of variation of the root mean square error scores is 2.3 % for thermal free-floating and 10 % and 14 % for final cooling and heating energies, respectively. Depending on the single input variations, overestimation above 50 % and underestimation below 40 % are calculated for a given energy service. Results show that the weather data, occupancy profiles, related internal gains, and ventilation rates can significantly affect the performance gap. The outcomes of this field study call for new analyses aimed at generalising the achieved results and developing appropriate modes to input the relevant parameters to minimise the performance gap with limited calculation efforts. • Energy and environmental field measurements in a fully monitored living lab. • Implementation and calibration of a dynamic building model with a very low-performance gap. • Performance gap due to weather data input peaks at 50 %. • Performance gap due to occupancy and internal gains input peaks at 42 %. • Combination of input modes may increase the performance gap by 135 % or reduce it by 84 %.