Groundbreaking Full Waveform Inversion Testing Doesn’t Need to Break Ground

A drill crew was staking out a project site in south Florida, trekking over the terrain and drilling bore holes into the ground. The thuds of a metal tube being driven into the ground rang out in the morning air. The crew was conducting a Standard Penetration Test, a geotechnical test that determines important soil properties before construction begins. All was going according to plan until the drill reached about 75 feet underground and simply disappeared. When a drill disappears out of sight during a test, it’s usually not a good sign. In this case, the drill had found a sinkhole.

Sinkholes are underground anomalies that form when water erodes an underlying rock layer. This creates a void that cannot be seen from the surface and could eventually lead to a collapse.  When the rock layer is a material that dissolves easily with rainwater, like limestone, sinkholes are more likely to form. As such, sinkholes are prevalent in Florida as most of the state is situated on a bedrock of limestone.

When the FDOT crew lost their drill rod, the State Materials Office saw an opportunity to leverage research related to sinkholes. State Geotechnical Materials Engineer David Horhota pointed the crew to research conducted by principal investigators Dr. Khiem Tran and Dr. Michael McVay at the University of Florida.

Dr. Tran and Dr. McVay had been developing a new non-invasive test, one that did not require drilling. The technique is called Full Waveform Inversion (FWI), and they were optimistic that it could do for geotechnical engineering what medical imaging has done for medicine. Dr. Tran says, “At the hospital, we use ultrasounds and MRIs to identify the defects inside the body. Here, we do the same thing with what lies below the ground’s surface.”

Scratching the Surface on Full Waveform Inversion

There are many types of geotechnical tests that can be conducted during design and construction, such as Standard Penetration Tests and Cone Penetration Tests, both of which the FDOT crew that discovered this sinkhole had performed. However, there are limitations to this approach: these tests can give detailed information at individual locations but do not provide a reliable picture for an entire construction site. Inversely, non-invasive tests, which use electromagnetic and seismic waves, can provide data for larger areas but with less accuracy at point-specific locations.

The FDOT research team led by Dr. Tran and Dr. McVay wanted to develop a test that could balance these trade-offs and provide higher-resolution data for large areas. Their first step was to explore the viability of using the FWI technique. This approach is unique from other types of seismic tests, as it uses the entire measured seismic wave field. While traditional seismic tests cannot capture information for underground zones with soft material, FWI can. The benefit is higher-resolution data that enhances geotechnical assessments.

The research team used computer simulations to see if the FWI technique could work theoretically. Simulations showed that it was possible not only to detect the voids, but also to determine different soil types and whether voids were filled with water or air. This was crucial, as this type of information helps to determine what remediation measures are appropriate. Following computer simulations, the research team conducted field tests that showed promising signs for FWI and gave the team insight into improving their testing methods.

Finding Solid Ground: Getting Two-Dimensional Results

The second phase of research involved further developing the FWI software and field-testing a prototype of the equipment. The main equipment was an array of geophones, strung together as a 120-foot line of sensors. Seismic energy was created using either a sledgehammer or a propelled energy generator that sent waves along the line of sensors to produce a two-dimensional cross-section of the ground beneath the line.

Line of geophones (sensors) used in the field to run Full Waveform Inversion tests.   
Source: Detection of Sinkholes or Anomalies Using Full Seismic Wave Fields: Phase II Final Report

Large amounts of data could now be produced in the field, albeit with a time-consuming process to map results. However, by making some improvements to the software, the research team was able to reduce processing times from two hours to under 30 minutes. Final software improvements by the team included a more user-friendly interface.

Further field testing revealed a remaining challenge to reliable use of the software and technology. State Geotechnical Materials Engineer David Horhota says, “The issue was if they had a void that was too far off the line of receivers, the results might not detect it or, if it were detected, distort the actual location or depth.” The research team determined that using a grid instead of a line of sensors would avoid this potential for distortion. With solution in hand, the team advanced to the final project phase to literally cover more ground.

Going Deeper with Three-Dimensional Data

To move from two-dimensional results to three-dimensional (3D) results, the team needed more sensors. Many more sensors. Instead of a line of geophones, the sensors were laid in a grid. This provided more data but required software optimization for timely processing of the larger data sets. Once the software was developed, the resulting data allowed the team to reliably develop 3D renderings of features below the ground surface. These 3D renderings then inform the need for borings or other tests in specific locations before further design or construction activities are undertaken.


3D rendering of subsurface features at a test site, with dark blue zones representing anomalies.   
Source: Sinkhole Detection with 3-D Full Elastic Seismic Waveform Tomography Final Report

The success of this type of test has caught the attention of the Federal Highway Administration (FHWA), which highlights proven innovations through its Every Day Counts program. FWI testing will gain further exposure when it is referenced in FHWA’s EDC-5 A-GaME Case History Synthesis report, expected to be published later this year.

FDOT and the research team have high hopes for FWI testing. Reflecting on the possibilities, Dr. McVay says, “This technology has changed a lot in the past 20 years. There is great potential, especially with rapid increases in computational power. We’d like this to be a tool that is part of the regular toolset to visit a site.” In the meantime, FDOT is finding plenty of opportunities to use the technology to reduce project risk, manage costs, improve designs, and accelerate delivery. The benefits of FWI are anything but superficial.

Further Reading

Phase 1: BDK75-977-66 Detection of Sinkholes or Anomalies Using Full Seismic Wave Fields
Final Report | Summary

Phase 2: BDV31-977-29 Detection of Sinkholes or Anomalies Using Full Seismic Wave Fields: Phase II
Final Report | Summary

Phase 3: BDV31-977-82 Sinkhole Detection with 3-D Full Elastic Seismic Waveform Tomography
Final Report | Summary

What’s Really Down There? Researchers Develop New Methods for Sinkhole Detection
Research Showcase Article