Dr. Andy Y. F. Leung
BEng, MS, PhD (Cantab), PE, MASCE
Associate Head (Partnership) and Associate Professor
Department of Civil and Environmental Engineering
The Hong Kong Polytechnic University
Presidentof Hong Kong Geotechnical Society (HKGES)
Doctor of Philosophy in Engineering, University of Cambridge, UK
Master of Science in Engineering, University of California, Berkeley, USA
Bachelor of Engineering in Civil Engineering, University of Hong Kong
Phone: (852) 2766 6064
Fax: (852) 2334 6389
Email: andy.yf.leung@polyu.edu.hk ; yfleung@polyu.edu.hk
Research Interests
1. Novel sensing technologies for civil infrastructure
The Smart Geotechnology laboratory exploresvarious aspects of soil-structure interaction problems through novel sensing techniques and investigation of strain distribution and stress mobilizations in composite structures and foundation systems. The research focuses on the use of
high-performance materials, such as high-strength steel materials, in foundation engineering and underground construction. Advanced instrumentation techniques including those with fibre optics sensing are investigated and deployed in field and laboratory studies.
Figure (left): Laboratoryfor development and calibration of advanced sensing equipment (right) Fibreoptic sensing equipment based on Optical Frequency Domain Reflectometry, and examples of fibre optic cables for field deployment
Selected Projects
Straindevelopments in composite structural elements & high-strength steel in foundationsComposite structural elements, such asconcrete-filled steel tube utilizes the confinements provided by steel tube to enhance the performance of concrete column, and can have wide range of applications in superstructures and substructures. The strain developments and interaction between these elements are, however, not well understood. The high-resolution optical fibre optic sensors enable the strain developments can
be accurately captured, providing a new means to understand the interaction between concrete and steel tubular structures of different geometries.The use of high-strength steel asfoundation elements, such as socketed H-piles, is associated with clear benefits. With the abovementioned sensing technology, the strain developments in the steel H section, grout and surrounding geomaterials can be revealed. The available bond strengths between different components can be determined and the new data shed new insights to the behaviour of these foundation components, inorder to promote their application in construction.
Figure. Examples of composite structures and H-sectionspecimens with fibre optic sensors, and strain development patterns
Fibreoptic sensing in field investigation of soil-structure interaction
The research team actively exploresopportunities for field deployment of new sensing technologies in underground construction projects, with recent examples including the monitoring of deep excavation, foundation piles, soil nails, tunnels, etc. Field instrumentation provides information on the response of underground infrastructure in complex and variable ground conditions, and helps to verify design assumptions on phenomena of soil-structure interaction.
Figure: Site deployment of fibre optic sensing to investigate soil-structureinteraction in diaphragm walls (Lin et al. 2023)
2. Reliability of geotechnical systems
Reliability analyses aims to quantify the propagation of geotechnical uncertainty, such as uncertainty in loads and material resistances, into the uncertainty in system performance. The measure of reliability, i.e. reliability index β, is more meaningful and explicit than the factor of safety FS, because it describes safety by the number of standard deviations, separating the best estimate of FS from its defined failure value of 1.0. A popular approach to evaluate reliability index is the First-order reliability method (FORM). Our group has developed efficient spreadsheet implementation of FORM, which can be coupled with commercial software such as SLOPE/W, to handle slopes with complicated geometries and with soil nails reinforcement (Lo et al. 2022). The implementation has been applied to multiple slope cases in Hong Kong, providing practical insights on the slope reliability before and after slope improvement. Our group has also developed an efficient design chart approach to further simplify FORM analysis, which enhance reliability-based design of slopes. Utilizing the design chart approach, it is possible to reduce the amount of soil nails, at the same time improving the slope reliability.
Our group is also experienced in modelling of spatial variability of soil properties. Random field of soil properties can be simulated based on the spatial variability characteristics, which can then be applied in reliability analysis. Huang et al. (2020) coupled the conditional random field model with SLOPE/W software, with the conditional random field producing realisations that exactly match the observed data at the soil sampling locations. Lo and Leung (2017) coupled the conditional random field with finite difference modelling, to quantify the efficiency of various soil sampling strategies. Leung and Lo (2018) coupled random field model with the matrix condensation technique, which can effectively evaluate the uncertainty in differential settlement and tilting of a pile group under a superstructure loading. Mori et al. (2020) coupled the random field model with the smooth particle hydrodynamics, which can effectively evaluate the uncertainty in the runout distance during landslide.
Apart from uncertainties in loads and resistances, model uncertainty forms another important aspect in reliability analysis. Model uncertainty involves bias and randomness, where the bias indicates if the model systematically over-predicts or under-predicts the quantity in question; and the randomness indicates the variability in predictions from one prediction of that quantity to another. Our group is actively researching on characterizing and applying the model uncertainty in slope reliability analysis. This involves a through compilation of region-wide database of centrifuge slope cases, slope failure cases and soil nail pullout tests, together with probabilistic analyses on the data.
The test data serves as a valuable source on the variability of soil properties and subsurface domain, when interpreted in conjunction with the understanding of mechanics of soil and infrastructure system, could greatly assist the reliability analysis. Our group is focusing on the digitization of region-wide test database of triaxial tests and consolidation tests, which provides robust characterization on the distribution of shear strength parameters and permeability characteristics. Since a region-wide distribution may not match the distinctive site conditions, our group has formulated a hierarchical Bayesian model to narrow down the distribution of soil properties, based on the observed data of the site, in conjunction with characteristics such as particle size and site labels. Our group also is actively researching on the interpretation of region-wide test data with advanced constitutive models such as Norsand, with the aim of providing a robust distribution of model parameters for the reliability analysis of complex soil-structure interaction problems.
3. Numerical simulation in Geotechnical Engineering
Smoothed particle hydrodynamics (SPH) and its application in sinkhole formation
SPH is a mesh-free method, which was originally applied to astrophysical phenomena and was later extended to a vast range of problems in both fluid and solid mechanics. Instead of solving the governing equations on predefined grids, SPH solves the equations based on dynamic particle connections and therefore has obvious advantages in handling problems associated with large deformations and evolving free surfaces.
The study proposed a single-layer soil-water couped SPH updates the propagation of pore pressure through combination of volumetric strain and Darcy’s law, accounting for the momentum equation, soil constitutive behavior, and the development of pore pressure at each timestep of the simulation. The proposed method is validated by analytical solutions of consolidation. Since the proposed method describes all hydraulic and mechanical information in a type of particle, thereby balancing the computation cost and calculation accuracy.
Figure. Illustration of the concepts of SPH method: (Left) interactions between SPH particles and (Right) kernel function
Figure. Flowchart of the proposed method, where the boundary is dynamically detected for sinkhole formation simulation.
Figure. Failure process of sinkholes with different Young’s muduli at the same mass loss ratio R
Discrete Element Method (DEM)and its role in investigating micro soil mechanics
DEMis a popular tool for studying the displacements and forces of distinct particles and for explaining the macroscopic physical observation from the microscopic point of view. The Newton’s Law of Motion and embedded force-displacement relationship govern the behaviours of each particle. Aninteresting but counter-intuitive phenomenon called size segregation can be observed in shaking a can of nuts for example, where large particles go upward while small particles drop down, as shown in right figures (Jing et al., 2017).
The influence of particle shape on shear dilatancy can also beexplained through DEM. Simulations with particles from angular to rounded show opposite trend of dilatancy with shape at the same void ratio and relative density. Its mechanics is explained through a series of particle pairssubject to shearing, with different shapes, packing structures, contact orientations and relative displacement modes .
Figure. (Left) Sliding induced shear dilatancy and (Right) Rotate induced shear dilatancy
People
Current group memebers:
- Geotechnical reliability
- Databases in geotechnical engineering
- Machine learning and deep learning
- Bayesian methods
Abd Ullah
PhD Candidate
- Strain characterization in composite structural elements and the effective use of high-strength steel material in foundations and lateral support systems of deep excavation
Dr. Maoxin WANG
Postdoctoral fellow
- Geotechnical reliability and risk
- Earthquake engineering
- Slope stability and landslide hazard
- Soil-structure interaction
- Machine learning
Dingjian ZHANG
PhD Candidate
- Application of high strength steel into geotechnical projects, optimizing socketed H-piles, soil-pile-structure interaction
Yingxu, Dawn, HUO
PhD Candidate
- Particle shape effect on shearing of sands using DEM and control of water pipe failure due to corrosion
Hansong PANG
Research Assistant
- Geotechnical Risk and Reliability
- Uncertainty in soil properties
- Reliability analysis and design
- Slope stability
Xiaoyu, Emma, CHEN
PhD Candidate
- Single-layer soil-water SPH and its applications in sinkhole and landslides
- Evaluate and assessment of geo-hazards
Wenxi MENG
Research Assistant
- Probabilistic modelling and assessment of rock and soil properties and failure
Previous Students:
Dr. Wenfei, Felix, LIU
Dr.Liu obtained his PhD degree from PolyU in 2018. He is currently self-employed on the Mainland, China. He manages two companies in ocean engineering, shipping equipment and intellectual property services.
Dr. Xiaoguang Xie
Dr Xie now works at CCB Investment Funds Management
Dr. Lei HUANG
Dr.Huang obtained his PhD degree from PolyU in 2021. He now is a lecturer at the School of Civil Engineering andArchitecture at Wenzhou University. He research
is mainly about slope reliability analysis and risk assessment.Activity
Group meeting
Field trip
Seminar/workshop/conference
Christmas Lunch 2023
