Modernise geotechnical investigation and analysis with machine learning

该项目旨在通过减少可变性和严格量化这些可变性来解决与岩土系统风险分析相关的无效性。预计它将在机器学习辅助风险分析和涉及随机变量的多相多物理多尺度问题的虚拟建模方面产生新的知识。预期成果是配备新功能的数据集和计算机工具，包括参数优化、不确定性量化、基于机器学习的替代模型和风险分析。这些工具将有助于弥合学术研究和工程实践之间日益扩大的差距，改变地质风险分析，优化复杂的施工流程。

The project aims to address the ineffectiveness associated with risk analysis of geotechnical systems by reducing variabilities and by rigorously quantifying such variabilities. It is expected to generate new knowledge in machine-learning-aided risk analysis and in virtual modelling of multiphase-multiphysics-multiscale problems involving random variables. Expected outcomes are datasets and computer tools that are equipped with new functionalities including parameter optimisation, uncertainty quantification, machine-learning based surrogate models and risk analysis. These tools will help to bridge the increasing gap between academic research and engineering practice, transform geo-risk analysis and optimise complex construction processes.