Impact of Occupants’ Behavior Uncertainty on Building Energy Consumption Through the Karhunen-Loève Expansion Technique: A Case Study in Italy

In Europe, the building sector is liable for 40% of the entire energy consumption (EC) and 35% of the total greenhouse emission. Building energy performance simulation (BEPS) tools are fundamental to assess the EC of both new buildings and energy retrofit intervention, and to verify the reaching of the requirements set by the national building energy standards. However, the results obtained from these tools are often unreliable due to the different assumptions that must be made in case of data input uncertainty, generating a “performance gap” between observed and predicted EC. Occupants’ behavior (OB) is one of the most difficult parameters to be estimated since affected by high uncertainty that may strongly affect the numerical results. However, the most recent BEPS tools neglect the existing uncertainty by modeling the occupant behavior through deterministic hourly-defined profiles. For this reason, in this work, the impact of OB uncertainties on EC is evaluated by applying a Karhunen-Loève Expansion (KLE) on deterministic hourly defined profiles. A typical Italian residential building is modeled and calibrated on EC data. Then, occupancy behavior-related profiles, such as heating setpoint, internal thermal loads, and windows opening, are randomly perturbed using the KLE technique. The results demonstrate that the heating setpoint patterns uncertainty has the highest impact on EC. Moreover, the more the energy performance of the building, the higher the impact of heat gains and losses caused by OB.