2025年8月1日，国际韧性基础设施研究中心（ICRI）邀请了英国皇家工程院院士、利兹大学计算机学院的Anthony G. Cohn教授，在岩土楼207室进行了学术交流。交流活动由ICRI主任黄宏伟教授主持，参与人员包括张东明教授、傅蕾博士、汪长松博士。傅蕾博士和汪长松博士分别作了专题报告，并与Cohn教授就各自的研究内容展开了深入且富有成效的讨论。

On August 1, 2025, International Joint Research Center for Resilient Infrastructure (ICRI) invited Professor Anthony G. Cohn—Fellow of the Royal Academy of Engineering and faculty member at the School of Computing, University of Leeds—for an academic exchange held in Room 207 of the Geotechnical Building. The event was hosted by Professor Hongwei Huang, Director of ICRI, and was attended by Professor Dongming Zhang, Dr. Lei Fu, Dr. Changsong Wang. Dr. Lei Fu and Dr. Changsong Wang each delivered research presentations and engaged in in-depth and productive discussions with Professor Cohn on their respective research topics.

图1 学术讨论

Fig.1 Academic Discussion

傅蕾博士作了题为《Ontology Modeling for Shield Tunneling: Design and Preliminary Validation》的报告。本报告提出构建一套面向盾构施工的语义本体，以系统表达地质环境、施工过程、参数设置与地表沉降之间的复杂因果关系。报告详细介绍了该本体构建的七个步骤，包括领域确定、概念体系设计、类与属性定义以及规则推理机制设计等，并借助现有成熟本体（如SWEET、GeoSPARQL等）进行了重用与扩展。通过引入SWRL规则语言，让本体具备智能推理能力，可实现沉降风险预警和参数调整建议的自动生成。最后，傅蕾博士展示了一个实际案例验证本体的应用可行性，并展望将本体与神经网络结合，开发兼具物理约束与数据驱动能力的混合建模框架，用于沉降预测与施工参数优化。该研究为实现知识驱动的盾构施工智能决策提供了新的思路和技术路径。

Dr. Lei Fu gave a presentation titled "Ontology Modeling for Shield Tunneling: Design and Preliminary Validation". The report proposed the construction of a semantic ontology tailored for shield tunneling, aiming to systematically represent the complex causal relationships among geological conditions, tunneling processes, parameter settings, and surface settlement. Dr. Fu detailed the seven development steps of the ontology, including domain identification, conceptual design, class and property definitions, and rule-based reasoning mechanisms. The ontology reuses and extends mature existing ontologies such as SWEET and GeoSPARQL. By incorporating SWRL rule language, the ontology is endowed with intelligent reasoning capabilities, enabling automated generation of settlement risk warnings and parameter adjustment suggestions. Dr. Fu presented a real-world case to demonstrate the ontology’s feasibility, and discussed the prospect of integrating the ontology with neural networks to develop a hybrid modeling framework that combines physical constraints with data-driven capabilities for settlement prediction and construction parameter optimization. This study provides a novel approach and technical pathway for knowledge-driven intelligent decision-making in shield tunneling.

图2 傅蕾博士学术报告

Fig. 2: Academic presentation by Dr. Lei Fu

汪长松博士作了题为《Hyperspectral Assessment of Concrete and Rock MassStrength in Tunnels》的报告。本报告介绍了一种利用高光谱成像技术结合人工智能算法，对隧道结构强度进行无损智能感知的新方法。针对混凝土衬砌，该研究通过机器学习建立了从光谱数据到抗压强度的预测模型，不仅实现了对裂缝等病害区域强度退化的可视化评估，还利用深度学习模型分析了混凝土早龄期强度的时空演化规律。针对掌子面围岩，报告创新性地提出了“高光谱能量熵”（HEE）方法来量化岩石损伤，并构建深度学习模型，实现了对围岩类型和强度的高精度同步识别与预测。该研究为隧道工程安全评价提供了一套快速、智能且无损的新技术手段。

Dr. Changsong Wang delivered a presentation titled "Hyperspectral Assessment of Concrete and Rock Mass Strength in Tunnels". The report introduced a new non-destructive and intelligent sensing method that leverages hyperspectral imaging technology and artificial intelligence algorithms to assess tunnel structural strength. For concrete linings, a machine learning-based predictive model was developed to estimate compressive strength from spectral data. This approach not only enables visual assessment of strength degradation in damaged areas such as cracks, but also utilizes deep learning models to analyze the spatiotemporal evolution of early-age concrete strength. Regarding the tunnel face surrounding rock, the study innovatively proposed the Hyperspectral Energy Entropy (HEE) method to quantify rock damage. A deep learning model was then constructed to achieve high-precision identification and prediction of rock types and strength. This research presents a rapid, intelligent, and non-destructive new technical solution for tunnel engineering safety assessment.

图3 汪长松博士学术报告

Fig. 3: Academic presentation by Dr. Changsong Wang

在交流过程中，Cohn教授针对傅蕾博士在本体构建细节及逻辑规则制定方面的工作，提出了许多中肯而富有启发性的建议；同时，还就汪长松博士开展的高光谱强度识别模型在现场应用中的验证问题，进行了深入探讨。两位博士通过此次交流获益良多，Cohn教授所提出的问题与建议为两位研究者提供了宝贵的思路和帮助。

During the discussion, Professor Cohn provided insightful and constructive suggestions on Dr. Lei Fu’s ontology development and logical rule design. He also held in-depth discussions with Dr. Changsong Wang on the validation of hyperspectral strength identification models in field applications. Both researchers greatly benefited from this academic exchange, and Professor Cohn’s questions and suggestions provided valuable inspiration and guidance for their ongoing work.

撰稿：傅  蕾

编辑：李  曦

校对：王  森

审核：申铁尧