Towards better prediction of building performance: a workbench to analyze uncertainty in building simulation

In this paper, the authors present the Georgia Tech Uncertainty and Risk Analysis Workbench (GURA-W), a software toolkit that explicitly captures uncertainty about the physical properties of the building and the energy models used to predict its performance. The workbench provides a UQ Repository, giving energy modellers direct access to previously quantified uncertainty distributions for a variety of parameters and models. The workbench also provides automatic identification and modification of parameter values in the input file for the simulation. Together with an intuitive user interface, these capabilities serve to increase the ease with which uncertainty and risk analysis is performed. As such, the methods become more accessible to the building design and retrofit profession at large, rather than being restricted to uncertainty analysis researchers. The predictions developed can serve as a basis for downstream risk-conscious design and retrofit decisions, for instance as part of contractual protocols for improved building performance.