This thesis proposes an evidence-based bottom-up methodology to calibrate a building energy model starting at the zone level, and finishing with the whole building level. The calibration is based as much as possible on measurements taken from the existing Building Automation System (BAS). This study presents the calibration at the zone and air handling unit level.
First a literature review is presented, followed by the evidence-based bottom-up methodology. Next, the case study building is described, along with the extraction and analysis of the monitored data. Then the building model is created and calibrated at the zone level based on the supply air flow rate to each zone. The calibration at the air handling unit level is based on: i) the supply air flow rate leaving the air handling unit; ii) the supply air temperature and iii) the cooling coil load. The evaluation of the calibration quality is proposed to be performed in three stages: i) graphical representation; ii) statistical indices (RMSE, CV-RMSE, NMBE); and iii) paired difference statistical hypothesis testing. 
A sensitivity analysis is performed and it is found that the energy model is not sensitive to changes in the building envelope parameters, but rather to variations in internal loads. 
Two approaches for representing the schedules of internal loads are compared and the proposed approach, where the internal loads are derived from measured cooling load in each zone, is recommended.