Khin Nyein-Chan Kyaw received "Best Paper Award" at GEOMATE 2025

2025/11/19

GEOMATE_Khin — Awardee

Khin Nyein-Chan Kyaw

Faculty Supervisor
Prof. Shinya Inazumi

Conference name

The 15th International Conference on Geotechnique, Construction Materials and Environment (GEOMATE 2025)

Award
Best Paper Award

Title of Paper

Evaluation of Chemical Grouting in Heterogeneous Soils by Integrating FEM Analysis and AI Models

Purpose of the Research

The primary objective of this research is to reduce uncertainties in chemical grouting penetration behavior caused by ground heterogeneity, thereby enhancing the reliability of liquefaction countermeasures in seismically active regions including Japan, Myanmar, and Thailand. To quantitatively assess penetration risks posed by low-permeability zones, the study developed an integrated approach combining FEM analysis's detailed evaluation capabilities with AI's flexible prediction performance. This integration enables relatively rapid and highly accurate predictions under complex ground conditions, contributing to more efficient design and construction of ground improvement methods. The framework presented in this research aspires to serve as a foundation for advancing sustainable infrastructure disaster prevention initiatives across these earthquake-prone regions.

Research Summary

This distinguished research proposes an innovative evaluation method that integrates Finite Element Method (FEM) analysis with artificial intelligence (AI) models to examine the influence of low-permeability zones on chemical grouting penetration behavior in heterogeneous sandy soils. The FEM analysis successfully demonstrated that the proximity of low-permeability zones significantly affects grouting flow velocity and distribution patterns, enabling risk assessment through simplified regression equations. The AI models (neural networks and gradient boosting) achieved impressive prediction accuracy with R2 = 0.849 for penetration range. Notably, even when low-permeability zones comprised 5.5% of the area, average filling rates remained excellent at 94.5% (FEM) and 96% (AI). Under the most challenging conditions, these rates maintained 81% and 83% respectively, while AI predictions were completed in approximately 2 seconds, demonstrating strong potential for practical applications. This methodology represents a valuable contribution toward enhancing the reliability of ground improvement techniques.

Future Prospects

Looking ahead, the framework demonstrated in this research is expected to further promote the utilization of digital technologies in geotechnical engineering. The integrated FEM-AI methodology will be validated through field experiments and ultimately developed into a system capable of real-time prediction. Additionally, by incorporating unsaturated soil behavior and grouting rheological properties, the approach may be extended to accommodate a broader range of ground conditions. Through these efforts, this work aims to contribute to improving infrastructure disaster prevention reliability in earthquake-prone Japan, Myanmar, and Thailand, fostering sustainable social infrastructure development in these regions. With an eye toward international standardization, the research team is committed to advancing the foundation for the evolution of ground improvement methodologies that can benefit communities across Southeast Asia and beyond.

The Prize Winners