This review paper provides a comprehensive overview of hygrothermal modeling with a focus on its application in mass timber construction, a sustainable and innovative construction approach. The first part of the review investigates the recent advancements in modeling heat and moisture transfer in wood-based materials, focusing on the key factors that influence moisture transport in wood. This section also covers the recent advances related to other bio-based building materials. The second part of the review discusses the latest findings in hygrothermal modeling applied to mass timber constructions, presenting various modeling approaches, tools, and simulation techniques. The discussion addresses how findings on heat and moisture transfer can enhance hygrothermal modeling of mass timber constructions, thereby improving the accuracy of simulations. Finally, the paper presents a thorough literature review on the application of machine learning in hygrothermal modeling. The emerging field of machine learning and its potential to enhance prediction accuracy for building assemblies is investigated. Key research gaps are identified with recommendations for future studies aimed at developing “white-box” simulation models tailored specifically for assessing hygrothermal performance of mass timber constructions. Such models could provide more reliable datasets, particularly useful for training black-box machine learning models. Additionally, the review highlights the potential of machine learning algorithms to accurately simulate complex heat and moisture transfers. It suggests further research to conduct a comparative analysis of algorithm performance and to explore integrating these “black-box” models with physics-based numerical models.

中文翻译：

---

大型木结构的湿热建模：最新进展和机器学习前景


这篇综述论文对湿热模型进行了全面的概述，重点关注其在大规模木结构建筑中的应用，这是一种可持续和创新的建筑方法。综述的第一部分研究了木质材料中热量和水分传递建模的最新进展，重点关注影响木材中水分传输的关键因素。本节还介绍了与其他生物基建筑材料相关的最新进展。综述的第二部分讨论了应用于大型木结构的湿热建模的最新发现，介绍了各种建模方法、工具和模拟技术。讨论讨论了热量和水分传递的发现如何增强大规模木结构的湿热建模，从而提高模拟的准确性。最后，本文对机器学习在湿热建模中的应用进行了全面的文献综述。研究了机器学习的新兴领域及其提高建筑组件预测准确性的潜力。确定了关键的研究差距，并为未来的研究提出了建议，旨在开发专门用于评估大规模木结构湿热性能的“白盒”模拟模型。此类模型可以提供更可靠的数据集，对于训练黑盒机器学习模型特别有用。此外，该评论还强调了机器学习算法在准确模拟复杂的热量和湿气传递方面的潜力。它建议进一步研究对算法性能进行比较分析，并探索将这些“黑匣子”模型与基于物理的数值模型相结合。