Monte Carlo parameters in modeling service life: Influence on life-cycle assessment

Service life plays an important role that has a significant influence on life cycle impacts of the building operation phase, especially as operational energy impacts are being reduced through energy efficiency and renewal energy. In this context, uncertainty analysis of service life models is crucial for reliable building Life Cycle Assessment (LCA). The main purpose of this paper is to examine and discuss the influence of the uncertainties of service life models using statistical analysis of Monte Carlo (MC) simulation results of the service life of fifteen different building elements considering eleven different published maintenance, repair, and replacement models from seven countries. Descriptive statistics of the service life data and MC simulation are used to identify the variability, quantify the uncertainties, and identify the best fit distributions for service life data. The results highlight the differences between the models, indicating the need to choose the most appropriate for the purpose and context of the LCA or Life Cycle Cost (LCC) study, as this choice will likely affect the results. Describing the potential range of values is useful as it improves decision-making in LCA and/or LCC analysis. Three distributions were found to be suitable for MC simulation of service life: gamma, lognormal and Weibull. In general, for a 90% confidence interval, the lognormal distribution returns a shorter service life (between the upper and lower quartile) than gamma and Weibull.