Systematic Time-Based Study for Quantifying the Uncertainty of Uncalibrated Models in Building Energy Simulations



Journal Title

Journal ISSN

Volume Title


Texas A&M University


This thesis documents the usefulness and accuracy of uncalibrated simulations to determine for what end-uses these simulations should be used. The study was divided into three segments 1)comparison of the accuracy of two simulation models, massless and advanced, against measured data 2) comparison of the results from two simulations models, simplistic and massless, to determine the sensitivity of envelope shape and details for two weather conditions 3) identification of the parameters that have a significant impact on the simulation output. Five buildings were selected as the test sample. Four of the buildings were multi story commercial buildings. The fifth was a single-family residential house. For the first segment of the study two simulation models were created for all the buildings; the massless model with emphasis on the envelope using massless construction and typical values for system parameters and the advanced model with the inclusion of thermal mass and extensive as-built details of the systems. For the second part of the research the simplistic model was created having a single floor one-zone with glazing and conditioned areas equivalent to the massless model. The sensitivity analysis was done using the massless model and selected variables from the loads and systems as sensitivity parameters. By following the procedure mentioned, it was found that uncalibrated simulation models do not depict the real operating conditions of a building. For some cases the simulated values are higher than the measured data while for others they are significantly lower. The CV (RMSE) between the measured and simulated values ranges from 30 to 150%. From the comparison of the simplistic and massless model, it was concluded that the outer envelope shape and details have an impact on the heating and cooling energy use irrespective of the weather conditions. For internally load dominated buildings this impact is more on the heating loads than on the cooling loads. The conclusions from the sensitivity analysis were that outside air fraction and the total supply air have the most significant impact on the simulation output while thermal mass has a small impact.