Robustness analysis framework for computations associated with building performance models and immersive virtual experiments

Building performance models (BPMs) have been used to simulate and analyze building during design. While extensive research efforts made improve the of BPMs, little attention has given their robustness. Uncertainty is a crucial factor affecting robustness in which such effect needs be quantified through suitable approach. The paper offers analysis framework for BPMs by using perturbation techniques uncertainty input datasets. To investigate efficacy framework, generative adversarial network (GAN)-based was selected as case study light switch usages single-occupancy office simulated an immersive virtual environment (IVE). GAN analyzed comparing differences between baseline (i.e., BPM obtained from trained on non-perturbed dataset) perturbed Overall, significantly reduced when training datasets were structured transformation techniques. remained relatively robust additive perturbation. Additionally, sensitivity involves different magnitudes corresponding levels suggests that effective investigating data BPMs.