What Matters the Most? The Role of Actual Occupancy Patterns and Automatic Model Calibration in Reducing the Building Energy Performance Gap in an Italian Case Study

Building energy models are increasingly used in energy renovation projects to identify the best retrofit strategy. However, a significant discrepancy between real and numerical building performances (“performance gap”) is generally observed, which can lead to an erroneous design of the energy retrofit measures. To reduce this gap, automatic model calibrations can be undertaken. This approach generally focuses on fine-tuning some “fixed” parameters to minimize an error function but often disregards the uncertainties in time-varying occupants’ behavior patterns. These latter are also commonly modeled through standardized profiles due to a lack of knowledge, then further increasing the performance gap, especially where occupants’ behavior may have a higher level of uncertainty, as in residential buildings. In this context, it is important to understand and quantify the impact of actual occupancy profiles on model accuracy also in comparison with that achieved through calibration. For this reason, this work compares, for a specific case study of social housing in Reggio Emilia (Italy), the performance gap reduction achievable through (i) common automatic calibration approaches; (ii) the modeling of the actual, experimentally observed, occupants’ behavior. The results reveal that modeling the actual users’ behavior decreases the error (RMSE) in indoor air temperature by 0.46 °C, i.e., more than the reduction obtained through the adopted calibration approaches (0.26 °C). In terms of energy consumption for space cooling, the performance gap without actual occupancy was significantly higher than that obtained for three monitored unoccupied apartments (AC always on), i.e., 10–15% against 1–4%. However, if the actual occupants’ behavior is modeled, the performance gap is reduced to the values obtained for the unoccupied apartments. This study highlights the importance of occupancy patterns in building energy modeling.