CURA Lab is pleased to announce the publication of a new research article in the Journal of Building Engineering, titled “A scalable probabilistic meta-model framework for accelerating Monte Carlo-based thermal performance analysis of small-to-medium-sized commercial buildings in six European regions under climate uncertainty.”
The study, authored by Ibrahim Tajuddeen, Hicham Johra, Eugénio Rodrigues, and Krzysztof Grygierek, introduces a novel data-driven framework designed to overcome one of the most critical limitations in building energy modelling: the computational burden of large-scale uncertainty analysis under climate change.
Traditional Monte Carlo approaches, while robust, are often prohibitively time-consuming when applied to high-dimensional problems and future climate scenarios. This work addresses that challenge by integrating Latin Hypercube Sampling, Bayesian Optimization, and advanced tree-based meta-models into a unified probabilistic framework. The proposed methodology achieves a remarkable 95–99% reduction in computational time while maintaining high predictive accuracy (R² > 0.90), enabling the generation of large stochastic datasets at unprecedented speed.
Applied to multiple commercial building typologies across six European climate regions, the study provides new insights into convergence behavior in uncertainty analysis. It demonstrates that reliable assessment of extreme outcomes (e.g., 5th and 95th percentiles) often requires significantly larger sample sizes than commonly assumed, highlighting the risks of relying solely on average performance metrics in climate resilience studies.
The results also identify key drivers of future building performance under climate uncertainty. In particular, the solar heat gain coefficient of south-facing glazing emerges as a dominant factor influencing cooling demand, while envelope thermal properties—especially in colder climates—remain critical for heating needs and overall comfort. These findings reinforce the need for balanced adaptation strategies that address both cooling and heating resilience in a warming climate.
Overall, the proposed framework represents a significant step toward scalable, high-resolution probabilistic assessmentof building performance, supporting more robust design decisions and evidence-based energy policy under deep climate uncertainty.
👉 The article is available open access: https://doi.org/10.1016/j.jobe.2026.115881.