Developing Machine Learning (ML) Techniques to Predict Tunnel Performance and Stability (UTI-UTC 09)

This project explores the application of machine learning (ML) techniques to enhance predictive capabilities in tunnel performance and stability, particularly focusing on mitigating the risk of collapse during tunnel boring machine (TBM) operations. By leveraging geological and TBM operation data from past tunneling projects, the research develops and trains classification models—including multilayer perceptron (MLP), support vector machine (SVM), and random forest (RF) algorithms—to forecast collapse events with high accuracy. A novel contribution of the project is the introduction of the "influence zone" concept, enabling spatial prediction of collapse-prone regions ahead of excavation. The research demonstrates the feasibility of ML in tunneling safety and paves the way for real-time risk monitoring systems that can alert engineers to unstable zones, thereby improving construction planning, operational safety, and infrastructure reliability.

Language

• English

Project

• Status: Completed
• Funding: $11,276.00
• Contract Numbers:

  69A3551747118

• Sponsor Organizations:

  University Transportation Center for Underground Transportation Infrastructure

  Colorado School of Mines
  Golden, CO  United States  80401

  Office of the Assistant Secretary for Research and Technology

  University Transportation Centers Program
  Department of Transportation
  Washington, DC  United States  20590
• Managing Organizations:

  Colorado School of Mines

  1500 Illinois St
  Golden, CO  United States  80401
• Performing Organizations:

  Colorado School of Mines

  1500 Illinois St
  Golden, CO  United States  80401
• Principal Investigators:

  Mooney , Mike

  Gutierrez , Marte

• Start Date: 20211001
• Expected Completion Date: 20240501
• Actual Completion Date: 20240501
• USDOT Program: University Transportation Centers

Subject/Index Terms

• TRT Terms: Collapse; Excavation and tunneling; Machine learning; Predictive models; Risk assessment; Tunnel borers
• Subject Areas: Bridges and other structures; Data and Information Technology; Geotechnology; Highways;

Filing Info

• Accession Number: 01954496
• Record Type: Research project
• Source Agency: University Transportation Center for Underground Transportation Infrastructure
• Contract Numbers: 69A3551747118
• Files: UTC, RIP
• Created Date: May 7 2025 6:57PM