Sensitivity analysis and validation of an EnergyPlus model of a house in Upper Austria
|W. Pereira de Jesus, A. Bögl, T. Natschläger. Sensitivity analysis and validation of an EnergyPlus model of a house in Upper Austria. pages 335-344, 6, 2014.|
|Buch||Pre-Proceedings of the International Conference on Sustainability in Energy and Building (SEB-14)|
This paper presents a heuristic sensitivity analysis and validation of an EnergyPlus model of a house in Upper Austria. We used sensitivity analysis to find out which material properties influence the simulation results the most. Previous studies about sensitivity analysis for simulating building systems have focused on quantifying the influence of input parameters in periods of time when the heating or cooling systems are also actuating on the building. The problem with that approach is that the influences of material properties and weather can be camouflaged by the actions of heating/cooling systems. In this work we study the influences of EnergyPlus input parameters and weather when the heating system is off. We use as a measure of evaluation the mean absolute error between the real and simulated temperatures. We modify an EnergyPlus weather file from a nearby location with updated data from a weather company and with predictions made by a regression decision tree. Using the best input parameters from the sensitivity analysis and the modified weather file, we were able to reproduce real measurements in a satisfactory way. In our case study the solar transmittance of the glass of the windows and the thermal conductivity of the bricks of the walls were the most influential input parameters. Our results also suggest that regression decision trees can be used to predict direct solar radiation for the weather file and improve simulation results.