Congratulations to Assistant Professor Shi Chao on the award of MOE Tier 1 Call 1/2023 Award

Project Write-up

This project aims at developing a systematic knowledge-based and data-driven approach to optimize site planning and sampling processes for automatic digital twin construction and quasi-real-time geotechnical risk assessment. Routine site investigation schemes often specify equal sampling spacing with regular patterns, and the current guideline for site planning in Singapore is conceptual and relies heavily on engineering experience.

Beyond the current state-of-the-art, site planning can be formulated as a multi-stage constrained optimization problem that optimizes sampling locations based on subsurface heterogeneities inferred from an underground digital twin, which is automatically built from sparse data and constrained by geotechnical domain knowledge in a manner consistent with the current engineering practice. To ensure application and impact, new algorithms will be developed for the smart determination of initial investigation locations and subsequent multi-round sampling with full consideration of geotechnical domain knowledge.

The proposed method expects to transform the current experience-based site investigation procedures into an automatic process and will bring about a paradigm leap by exploiting subsurface heterogeneities and geotechnical domain knowledge for risk-informed site planning and optimization.