MS005 - Data-Driven Methods and Digital Twin Applications in Geotechnical and Ground Engineering
Keywords: digital models, digital twins, geotechnics, infrastructures, real-time monitoring, scientific machine learning
The demand for development of sustainable and resilient geo-structures, as well as the repurposing and reuse of existing ones, presents significant challenges for geo-engineers and geo-scientists. The design complexities are increasing while optimising available resources. Computational modelling and data-driven approaches have unavoidably become fundamental tools in the design and back-analysis of geotechnical structures such as tunnels, deep basements, slopes, dams, retaining walls, and foundations. Recent advancements in numerical methods and data-driven techniques have revolutionised traditional approaches, enhancing the accuracy, efficiency, and reliability of geotechnical engineering-related projects, and extending their application beyond field experts to broader community of engineers and geo-scientists. This minisymposium aims to collect advanced data-driven models and digital twin approaches dealing with geotechnical and ground engineering problems.
This MS will cover a range of cutting-edge subtopics, including but not limited to:
• Application of physics-enhanced machine learning in geotechnical engineering.
• Model- and data- driven techniques including model update, inverse problems, fusion of models and data and virtual control for geotechnical engineering problems.
• Data-assimilation and digital twin applications, e.g., BIM-based models, for analysis, visualization and maintenance of infrastructure.
• Reduced order modelling for real-time predictive digital twins.
• Advanced and robust ML-based approaches for constitutive modelling.
• Advanced data-driven models for elastic and acoustic wave propagation problems.
REFERENCES
[1] Ninić, J. and Meschke, G., 2015. Model update and real-time steering of TBMs using simulation-based metamodels. Tun and Underground Space Tech, 45, pp.138-152.
[2] Xu, C.; Cao, B.T.; Yuan, Y. and Meschke, G., 2023. Transfer learning based physics-informed neural networks for solving inverse problems in engineering structures under different loading scenarios. Computer Method in Applied Mechanics and Engineering, 405:115852
This MS will cover a range of cutting-edge subtopics, including but not limited to:
• Application of physics-enhanced machine learning in geotechnical engineering.
• Model- and data- driven techniques including model update, inverse problems, fusion of models and data and virtual control for geotechnical engineering problems.
• Data-assimilation and digital twin applications, e.g., BIM-based models, for analysis, visualization and maintenance of infrastructure.
• Reduced order modelling for real-time predictive digital twins.
• Advanced and robust ML-based approaches for constitutive modelling.
• Advanced data-driven models for elastic and acoustic wave propagation problems.
REFERENCES
[1] Ninić, J. and Meschke, G., 2015. Model update and real-time steering of TBMs using simulation-based metamodels. Tun and Underground Space Tech, 45, pp.138-152.
[2] Xu, C.; Cao, B.T.; Yuan, Y. and Meschke, G., 2023. Transfer learning based physics-informed neural networks for solving inverse problems in engineering structures under different loading scenarios. Computer Method in Applied Mechanics and Engineering, 405:115852
