Geotechnical Resilience via Intelligent Design

Automating core logging: ML-based pipeline for jet grout columns

As part of the GRID project, Daniel Ryghseter (NGI) presents a machine-learning-based approach to automating core logging for jet grout columns at BOKU. The class covers image-based data processing, ML methods for automated feature recognition, data annotation, model training and validation, and integration into geotechnical project documentation — aimed at students, researchers and professionals in geotechnical engineering and digital construction.

// Oct 12, 2026 · 10:00 · Simonyhaus, Peter-Jordan-Strasse 65, Vienna

Workshop on AI & the Evolution of Geotechnical Engineering

Enrico Soranzo (BOKU) participated in a workshop on “AI and the evolution of Geotechnical Engineering” at Sapienza University of Rome, organised by AGI and the Department of Structural and Geotechnical Engineering. The workshop explored physics-informed approaches integrating machine learning, physical models and numerical simulation to improve prediction reliability, uncertainty management and geotechnical problem analysis, with application cases presented to support design decisions.

// Sep 21, 2026 · 10:00 · Aula 33, via Eudossiana 18, Roma

The Pore Radius Propagation approach: computational characterization of pore structures in granular media

A new GRID-affiliated paper in Computers and Geotechnics introduces the Pore Radius Propagation (PRP) approach — a computationally efficient method to estimate pore-size distributions from computed tomography (CT) datasets. Roberto Cudmani and colleagues at TUM, with Tongji University, validated it on open-access in-situ CT images of Hamburg sand and glass beads under unconfined compression and drainage–imbibition cycles, deriving water-retention behaviour and capturing the spatial evolution of pore structures under loading.

The framework also reproduced the microscale spatial distribution of capillary water under controlled suction, and cut processing time by ~16–50% versus conventional pore-morphology methods while keeping high predictive accuracy — a reliable basis for automated extraction of microstructural parameters.

Emilio Bilotta (UNINA) on GRID Secondment across China: PolyU Keynote & Tongji Research

From 8 to 13 August 2026, Emilio Bilotta (University of Naples Federico II) carried out a GRID secondment at The Hong Kong Polytechnic University, hosted by Prof. Zhen-Yu Yin, embedded in the First ICGdR World Congress on Geo-disaster Reduction. He delivered a keynote on earthquake-induced damage to tunnel linings in tectonically active zones, co-chaired Session 7, and visited PolyU’s geotechnical laboratories.

A follow-on four-week secondment at Tongji University (15 Aug – 9 Sep) focused on shallow tunnelling in dense urban environments and monitoring-informed numerical back-analysis, extending a deterministic framework towards physics-informed Bayesian updating and a path to probabilistic digital twins for underground construction. Through Tongji’s High-End Foreign Expert Program the visit also reached Guizhou, Chongqing Jiaotong, Xi’an Jiaotong and Beijing University of Technology, concluding with a seminar at Tongji alongside Prof. Dongming Zhang and colleagues.

Lei Chen (University of Leeds) Seconded to BOKU & HDAnalytics: ML for Global Lake Seasonality

Lei Chen from the University of Leeds began a three-month secondment at BOKU and HDAnalytics, building on his research Machine learning revealing changes in the seasonality structure of global lakes. The visit examines how climate change reorganizes the seasonal dynamics of global lakes and how machine learning can reveal the dominant drivers of these changes.

The work assesses global patterns and long-term changes in lake seasonality structure (LSS) from 2001–2023 by linking satellite-derived lake observations with lake ice, hydrological extremes, climate forcing and lake-property data. Over three months it moves from data preparation and framing, to a global assessment of LSS changes — with focus on lake ice in high-latitude regions and hydrological extremes in ice-free regions — to an interpretable machine learning model quantifying the relative importance of climatic, hydrological and lake-specific controls. Expected outputs include a harmonized global dataset, a global LSS assessment, an interpretable ML model, and a draft manuscript.

Prof. Francesca Ceccato & Xinglai Sun (UniPD) Complete Secondment at Tongji University

Prof. Francesca Ceccato and PhD candidate Xinglai Sun from the University of Padua — a new beneficiary of the GRID project — completed their secondment at Tongji University in Shanghai.

During the visit, Francesca delivered presentations on advanced geotechnical modelling and machine-learning algorithms for geohazard assessment and mitigation, and taught a hands-on course on Anura3D to Tongji’s doctoral students. The secondment sparked valuable exchanges on integrating advanced numerical modelling with artificial intelligence.

This collaboration strengthens the links between GRID’s European and Asian partners and supports innovative tools for geotechnical risk assessment and resilient design. Special thanks to Prof. Dongming Zhang, Prof. Jie Zhang, Prof. Mingliang Zhou and Dr Meng Lu for their warm welcome and stimulating scientific discussions.