An automated system for inspecting rock faces and detecting potential rock falls using machine learning

The objective of the proposed research is to develop an automatic system for identifying blocks of rock that are prone to rock fall. Rock falls are a threat to the safety of residents, drivers, and transportation infrastructure at locations adjacent to steep rock cuts, including the South-Central states of Region 6. Furthermore, rehabilitation of transportation infrastructure after a rock fall is costly. A principal means to mitigate rock fall hazards is to detect and remove rocks that are prone to fall by manually inspecting and scaling existing exposed rock surfaces. Trained crews access the rock faces - often by rappelling over the edge from above or via portable lifts - and use pry-bars to strike the rock. The sound and feel of striking the rock are used to identify loose rocks that are then scaled. Rock inspection and scaling is high risk to the workers. The research team proposes a research program to utilize an automated tap hammer to strike rock surfaces and record the resulting reflected waveforms with a microphone. The response of the rock to the hammer tap will be interpreted in terms of the stability of the rock, similar to the conventional manual approach except that it is a less subjective measure. It is noted that a similar system has been developed for remote inspection of bridges by one of the project participants. It has been demonstrated that an automatic tap testing device can collect the acoustic impact response of concrete bridges automatically, and that these data can be used with machine learning classification methods to identify different structural states (i.e., damaged vs. nondamaged). The team will adapt and modify this technology for use in identifying potentially loose rock blocks on rock faces associated with transportation infrastructure. Beyond reducing the risks and costs associated with manual inspections, more consistent and useful data will be collected. In addition, future inspections can be repeated at the same location. By returning to the same locations on a periodic schedule, changes in the response of the rock face can be readily identified and used to focus attention and resources on these potentially problematic areas. The proposed project is divided into 3 tasks. The first two tasks involve technology development and the third task is related to implementation as summarized below. The objective of Task 1 is to quantify the response of rock with and without discontinuities (fractures, joints, bedding planes, etc.) to a tap hammer strike under controlled laboratory conditions. In this way, the potential of tap hammer technology to identify loose rock blocks, that is, blocks containing weak discontinuities, can be quantified. This task includes the development of the crank rocker mechanism, validation, and testing. The data collected in the laboratory will be processed through machine learning algorithms. The objective of Task 2 is to use the test tap test on field rock and correlate the tap test response to characteristics of the field discontinuities. Data collected in the field will be included in the data base that that is being processed through machine learning algorithms for data clustering. Task 3 will involve field implementation of the technology coordinated with the New Mexico Department of Transportation (NMDOT). The purpose of this task is to provide a field-based exploration of the new technology, identify performance limitations and barriers for implementation, and suggest recommendations for further development.

Language

  • English

Project

  • Status: Active
  • Funding: $ 100000
  • Contract Numbers:

    20GTUNM31

  • Sponsor Organizations:

    Office of the Assistant Secretary for Research and Technology

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

    Transportation Consortium of South-Central States

    Louisiana State University
    Baton Rouge, LA  United States  70803
  • Project Managers:

    Mousa, Momen

  • Performing Organizations:

    University of New Mexico, Albuquerque

    Department of Civil Engineering
    Albuquerque, NM  United States  87131-0001
  • Principal Investigators:

    Stormont, John

  • Start Date: 20200801
  • Expected Completion Date: 20220201
  • Actual Completion Date: 0
  • USDOT Program: University Transportation Centers Program

Subject/Index Terms

Filing Info

  • Accession Number: 01757537
  • Record Type: Research project
  • Source Agency: Transportation Consortium of South-Central States
  • Contract Numbers: 20GTUNM31
  • Files: UTC, RiP
  • Created Date: Nov 10 2020 10:01AM