Multi-Family Housing Thermal Parametric Model Calibration and Validation

Calibration and validation allow the designer to assess the effectiveness of the model reliability and evaluate and quantify whether it is acceptable for the proposed objective. Internal and external air temperature values were measured in the field to carry out the calibration, considering a multi-family housing building located in the city of São Luis, Maranhão, Brazil. For calibration, the indicators of normalized mean bias error (NMBE), the coefficient of variance of the mean squared error (CV[RMSE]) and the coefficient of determination (R²) were used. Twelve building models with different levels of detail were therefore constructed to test and study the interference of the modelling on the simulation results and its reliability compared to the values measured in the field. The results showed that the more complex models have better reliability, although it is possible to achieve calibration rates with simplified models.