A probabilistic-based approach to support the comfort performance assessment of existing buildings

Abstract

Building performance is focused on maintaining and increasing sustainability while enhancing occupants' comfort levels. Occupants' comfort depends on uncertain, interrelated personal, social and building factors. However, the relationships between these factors are not covered by performance assessment tools based on linear analysis approaches. This study proposes a probabilistic model based on Bayesian networks (BNs) to assess the comfort performance of a building. An extensive review and evaluation of the causal factors of building performance, supported by the results of a satisfaction survey, are the basis for the BN model. Sensitivity analysis is used to verify the BN model, and the proposed approach is tested on an existing building. Findings from this research can guide decision-makers in the facility management industry to assess and understand causal factors of occupants’ comfort to properly establish sustainable and cleaner production strategies.