BIM integrated life cycle assessments (LCAs) pave the way for LCAs to enter early de-sign phases. Due to the information deficit, BIM models have not yet reached their full potential to provide adequate information for LCAs at an early design stage. These in-formation deficits must be closed by fuzzy assumptions, whereby uncertainties and var-iabilities must be integrated into life cycle assessments. Thereby, the BIM process can be accompanied by the optimisation of buildings according to ecological aspects in ad-dition to economic aspects. The idea behind this master thesis (MA) is to develop a statistical and probability-ori-ented calculation approach to make the uncertainties in BIM-based life cycle assess-ments in early design phases calculable. Global sensitivity analysis, coupled with uncer-tainty analysis, is an instrument for assessing the robustness of results, based on which well-founded decisions can be made. Monte Carlo analyses and experimental designs are regarded as promising approaches to deal with sensitivities and uncertainties in LCA. Due to the scalability and the large amount of data in LCAs, the Monte Carlo anal-ysis turned out to be suitable, although difficulties are mainly seen in analysing the ef-fects of individual input parameters on the uncertainty in the results. The concept was implemented in Python and tested with 13 ecologically relevant build-ing elements. The thickness, area and service life of the elements were sampled ac-cording to Monte Carlo and stored in matrices. Using matrix multiplications, a 3D result matrix could be produced, which contains the information state corresponding uncer-tainties and thus represents all possible results for the greenhouse potential and the non-renewable primary energy consumption of the reference building. Based on the re-sults matrix, an uncertainty and sensitivity analysis could be performed. The results showed that thicknesses and material variations are the biggest contributors to uncer-tainties in BIM based LCAs. In addition, the increase in the Level of Development (LOD) of a BIM model from LOD 100 to 200 to 300 leads to a significant improvement in the availability of information for LCA. Despite high uncertainties in early design phases, the calculated mean value converged relatively quickly to a final scenario. Furthermore, the increase of the information depth of the four most uncertain building elements led to an improvement of the accuracy of the results by more than 50 %. In this master thesis I agree with ALVAREZ et al. that a lot of research and development is needed to enable fully automated BIM based holistic LCAs in early design phases.     «

BIM integrated life cycle assessments (LCAs) pave the way for LCAs to enter early de-sign phases. Due to the information deficit, BIM models have not yet reached their full potential to provide adequate information for LCAs at an early design stage. These in-formation deficits must be closed by fuzzy assumptions, whereby uncertainties and var-iabilities must be integrated into life cycle assessments. Thereby, the BIM process can be accompanied by the optimisation of buildings according to ecolog...     »