Sensitivity Analysis and Validation of an EnergyPlus Model of a House in Upper Austria

Verification of energy related building simulations is crucial. In this paper we propose a sensitivity analysis method to find out the most influencing simulation parameters, and a machine learning based approach for predicting local weather inputs. Using the results from the sensitivity analysis and the predicted local weather, measurements could be reproduced in a satisfactory way. In our case study, the solar transmittance of the window glass and the thermal conductivity of the bricks were the most influential parameters. Our results also suggest that regression trees can be used to predict direct solar radiation for the local weather.