dr. H. Cheng (Hongyang)

Dr. Hongyang Cheng specializes in multi-scale modeling of soils and Bayesian uncertainty quantification for geotechnical applications. His research bridges physics-based and data-driven approaches to advance soil mechanics, focusing on granular material behaviors—from quasistatic to free-flowing—and quantifying parameter uncertainties across particle to macro scales. Dr. Cheng serves as an editorial board member for Soils and Foundations, a nominated member of the Technical Committee 105 “Geomechanics: Micro to Macro”, and co-leads Working Group 1 of the COST Action ON-DEM.

Dr. Cheng supervises 3 postdocs/research engineers (2 completed), 7 PhD students (1 completed), and 8 MSc students (4 completed). As a collaborator and mentor, he actively engages with industry partners and advocates for Open Science. His research contributes to multi-scale risk assessment and optimization methods for geotechnical structures under extreme loading conditions. Click here to explore his recent NWO- and EU-funded projects on earthquakes and submarine landslides.

Expertise

• Earth and Planetary Sciences

  • Parameter
  • Calibration
  • Simulation
  • Discrete Element Method
  • Model

• Engineering

  • Element Method
  • Models
  • Discrete Element

Organisations

Dr. Hongyang Cheng’s research focuses on advancing the understanding and modeling of soils through two primary lines of inquiry: multi-scale modeling and Bayesian uncertainty quantification. His work aims to bridge the gap between theoretical soil models and practical risk assessments, particularly in geotechnical applications like dike and embankment safety.

Multi-scale Modeling: Dr. Cheng specializes in modeling granular materials (e.g., soils, rocks) across particle, meso, and macro scales. During his PhD, he developed Discrete Element Method models to understand fiber-reinforced soils and derive constitutive laws. As a postdoc, he improved direct numerical simulations of fluid-saturated soils by integrating microstructural and particle-fluid dynamics. His contributions include also generalized multi-scale formulations that better conserve momentum and energy, with applications to soil-structure interactions and transient granular flows.

Bayesian Uncertainty Quantification: With ten years of experience in Bayesian inference, Dr. Cheng developed methods like iterative Bayesian filtering for efficient parameter inference and spearheaded “GrainLearning”. GrainLearning combines physics-based modeling with ML, utilizing clustering algorithms and neural networks to improve inference and computational efficiency. This tool has earned recognition, including industry adoption and funding to expand its surrogate modeling capabilities.

Dr. Cheng is actively involved in multidisciplinary initiatives, such as the Lorentz Center workshop he organized on “Machine Learning for Discrete Granular Media,” which aimed to integrate physics-based and machine-learning approaches in computational geomechanics. He also leads working groups to promote open science within the DEM community. The application of his research extends beyond geotechnics to other fields, including laser sintering, pharmaceutical powder processing, and industrial granular material handling.

Publications

Research profiles

In the Civil Engineering & Management Department, Dr. Cheng teaches undergraduate courses on Soil Mechanics (second-year) and Geotechnics for Dike Design (third-year), equipping students with foundational knowledge in soil behavior and practical techniques for soil characterization in geotechnical applications. At the MSc level, he teaches GeoRisk Management (5EC), where students integrate probability theory, stochastic modeling, and numerical methods (FEM) to assess and manage risks in geotechnical engineering, by modifying and running a dike model (Python) for risk assessment.

During his postdoctoral tenure in the Thermal and Fluid Engineering Department, Dr. Cheng coordinated and lectured in three 5EC MSc courses: Multiphase Flows, focusing on the dynamics of particle-fluid systems; Granular Matter, exploring soil elastoplasticity; and Advanced Programming in Engineering, emphasizing image analysis techniques.

191158500 - APiE

201400194 - Granular Matter

201400300 - Multiphase Flows

Affiliated study programs

Courses academic year 2025/2026

Courses in the current academic year are added at the moment they are finalised in the Osiris system. Therefore it is possible that the list is not yet complete for the whole academic year.

Courses academic year 2024/2025

Courses academic year 2023/2024

Current projects

POSEIDON

Improve offshore infrastructure resilience against geohazards towards a changing climate

POSEIDON, funded by the MSCA Doctoral Networks 2022, is a collaborative project bringing together an interdisciplinary team of experts from academia and industry across the EU. It aims to train a new generation of doctoral candidates through a comprehensive program that combines advanced research and transferable skills. The project focuses on developing and integrating advanced numerical modeling and machine learning tools to address challenges in offshore geohazards and critical infrastructure. POSEIDON fosters cross-sectoral collaboration to equip researchers with the expertise needed for future advancements in the field.

Role: daily supervisor of 3 PhD students (2024-present); co-applicant; work package leader

FastSlip

Bridging Dynamic Fault Slip Multiphysics to All Relevant Scales of Induced Seismicity

Earthquakes often result in damage to buildings and infrastructure and sometimes loss of human lives. Induced earthquakes due to human activities like gas extraction, are the result of fast slip on powder-filled pre-existing faults in the subsurface due to a rapid breakdown of their strength. The physical mechanisms contributing to the rapid failure remain unclear. In this project we will perform experiments at scales of a single rock grain to millions of grains combined with computer models to investigate these weakening mechanisms. Our results will help to better constrain the hazard of induced earthquakes.

Role: supervisor of 1 Postdoc (2025-present); co-applicant

TUSAIL

Training in Upscaling particle Systems: Advancing Industry across Length-scales

Particulate matter is ubiquitous – from grains of sand on the beach and aerosols in our environment to powders in pharmaceutical pills and raw materials for industrial processes. Numerical models are fundamental to enhancing our understanding of the behaviours of such matter. However, there is always a trade-off: retaining enough detail as the model is scaled up from particle to system so that accuracy is maintained versus simplifying so that computation does not become intractable without adding value to the outcome. The EU-funded TUSAIL project is training a new generation of researchers to develop and apply novel methodologies of upscaling that ensure a scaled model retains the information necessary to capture system behaviour. Outcomes will be invaluable to numerous industries, significantly enhancing EU competitiveness and human capacity.

Role: daily supervisor of 1 PhD at University of Twente and co-supervisor of 1 PhD at University of Edinburgh

UNSAT

U-Net segmentation of 3D micro-CT images of rooted soils using label data from multi-physics simulators

This project developed a machine learning model to segment micro-CT images of vegetated soil, automating the assignment of soil, air, water, and plant root labels to pixels in 3D images. By streamlining the analysis of soil-vegetation interactions, the tool enhances the use of x-ray tomography for studying the effects of vegetation on soil properties. Building on a pre-existing manual image analysis pipeline, this automated solution aims to simplify complex data processing, benefiting researchers without extensive expertise in image processing.

Role: supervisor of 1 Postdoc and 2 Research Software Engineer; co-applicant

ON-DEM

ON-DEM

Particle-based simulations, such as those using the Discrete Element Method (DEM), are widely applied across disciplines to model diverse materials like sand, grains, and powders. With numerous DEM software packages available, selecting a suitable code can be difficult, even for experienced researchers. Open-source programs, however, offer benefits like accessibility, reproducibility, and transparency, avoiding the "black box" limitations of commercial software. This COST Action aims to unify the DEM community by sharing advancements, promoting best practices, and providing training and resources, including simulation examples, validation experiments, and data analysis tools. The initiative focuses on five themes: solving large-scale industrial problems, incorporating realistic physics, improving data visualization, standardizing methods, and enhancing commercial applications of DEM.

Role: vice-lead of Working Group 1 "passing through time and space scales"; second proposer

Finished projects

GrainLearning

An artificial brain for interpreting and accelerating physics-based simulations of granular materials

GrainLearning, funded by the Netherlands eScience Center, is a project focused on enhancing the "GrainLearning" software for simulating and analyzing granular materials. The software specializes in Bayesian uncertainty quantification, enabling the inference of model parameters from data. A key innovation is its integration of machine learning surrogates with physics-based models, allowing for faster and more efficient uncertainty quantification and optimization. This approach addresses challenges like deriving constitutive theories from particle-scale simulations and assimilating real-world observations into numerical models, with applications ranging from geotechnical engineering to particle technology. For more information, visit grainlearning.readthedocs.io.

Role: supervisor of 2 Research Software Engineers (2022-2024); main applicant

• In the column 'Lesson in Imagination' by U-Today, science journalist and illustrator Enith Vlooswijk explores how educators use metaphors to explain complex concepts. In the episode 'About Dikes and Sandcastles,' the work of Hongyang Cheng, assistant professor in the ET department of Soil MicroMechanics, is highlighted.
  https://www.utoday.nl/les-in-verbeelding/73108/over-dijken-en-zandkastelen
  

Address

University of Twente

Horst Complex (building no. 20)
De Horst 2
7522 LW Enschede
Netherlands

Navigate to location

Organisations