Academic Report on July.27,2019

Reporter Topic:Enhancing value of geotechnical sampling and instrumentation from the reliability perspective

Reporter: Professor Andy Leung, Hong Kong Polytechnic University

Time:10:00, July.27, 2019

Location:Conference Room 8213, Department of Engineering

     Prof. Leung is currently Associate Professor at The Hong Kong Polytechnic University. He graduated from The University of Hong Kong (BEng) and University of California, Berkeley (MS), before he obtained PhD degree at the University of Cambridge, UK. His research interests include soil-structure interaction, reliability of geotechnical and structural systems, probabilistic analysis approaches and risk management in infrastructure developments. He is currently the Secretary-General of Hong Kong Geotechnical Society, and had served on the Technical Committee for Code of Practice of Foundations of the Hong Kong Government. He has received awards including the HKIE Fugro Prize, Departmental Teaching Excellence Award, Dean’s Award for Outstanding Achievement in Research Funding, etc.

     Geotechnical engineering involves natural materials with significant variability. Soil sampling and field instrumentation are therefore key components in design and construction stages to reduce uncertainty or the associated risk levels. This presentation first explores the value of soil sampling, defined by its effectiveness in reducing system uncertainty, using quantitative sensitivity approaches extended for application in spatially correlated variables. The approach is illustrated through implementations in slopes and shallow footings on spatially variable soils, where the examples also allow the idea of sampling efficiency to be re-examined from two perspectives: one from classical soil mechanics and another from probabilistic analysis considering geotechnical uncertainty. In more complex engineering systems such as braced deep excavations, it is possible to utilize both soil sampling data and construction monitoring data to refine the predictions in subsequent construction stages. This presentation will also introduce an adaptive model updating approach for such purposes. Based on Bayesian learning, the approach considers various sources of uncertainty, including model uncertainty and spatial variability, which may lead to discrepancies between predicted and actual excavation responses. These are illustrated through re-analyses of a multi-stage braced excavation for a metro station construction project in Hong Kong. The presented approaches provide efficient modelling tools to extract additional value from soil sampling and field monitoring data, which also facilitate data-driven decision-making in geotechnical engineering.