Cavity Geometry Evaluation beneath Shield Tunnel Linings using Ground Penetrating Radar and Acoustic Sensing

Deep beneath the bustling streets of modern cities, a hidden world of infrastructure keeps society moving. Subways, utility tunnels, and underground passages are carved out by massive machines, leaving behind concrete rings that hold back the earth. To ensure these rings stay secure, workers fill the tiny gap between the new concrete and the surrounding soil with a thick, cement-like paste called grout. This backfilling is crucial; it supports the tunnel and prevents the ground from shifting. However, if this grout is not poured perfectly or if it leaks away over time, empty pockets of air can form behind the tunnel walls. These hidden cavities are dangerous because they weaken the structure, potentially leading to sudden collapses, yet they are invisible from the inside of the tunnel where the concrete looks perfectly smooth.

For decades, engineers have tried to find these empty spaces without breaking the tunnel apart. They have used tools like ground-penetrating radar, which sends radio waves into the ground to look for changes in material, and thermal cameras that look for temperature differences. But these methods have a blind spot when it comes to reinforced concrete, the kind used in most modern tunnels. Inside these walls, a grid of steel bars, known as rebars, creates a chaotic signal that often hides the very defects engineers are trying to find. When the radar waves hit the steel, they bounce back in a confusing mess, making it nearly impossible to tell if a reflection comes from a steel bar or a dangerous hole behind the wall. This uncertainty leaves tunnel operators guessing about the safety of their infrastructure.

A team of researchers at Korea University and Konkuk University decided to test a different approach by combining the traditional radar with a simple, yet often overlooked tool: a microphone. Their goal was to detect these hidden voids and estimate their dimensions. To do this, they built a large-scale model of a tunnel lining in a laboratory. They constructed a thick, reinforced concrete plate to mimic the tunnel wall, complete with a realistic grid of steel bars spaced about 38 centimeters apart. Beneath this, they placed a second concrete plate to represent the ground. On this lower plate, they carved out four distinct empty spaces to simulate the cavities: two were square and two were circular, with widths and diameters ranging from 30 to 60 centimeters. They filled the rest of the gap with mortar to act as the backfill, leaving only the specific cavities empty.

The researchers then ran two different tests across the surface of the reinforced plate. First, they used a ground-penetrating radar unit, dragging it along specific lines to scan the area. The radar images showed the expected results: the steel bars created strong, curved reflections that dominated the picture. While the radar could detect that something was different in the areas with cavities, the signals from the empty spaces were tangled with the signals from the steel bars. The researchers found that they could use a mathematical technique called a wavelet transform to try and separate these overlapping signals, and they could see that the signal strength increased as the cavities got larger. However, the steel bars made it very difficult to clearly define the edges of the holes or determine if a cavity was square or round; the study concluded that identifying the shape of the cavity remains challenging and that the accuracy of dimension estimation using radar might decrease.

Next, the team turned to the microphone. Instead of sending radio waves, they used a specialized hammer to tap the concrete surface, creating a sharp sound and a vibration that traveled through the plate. A sensitive microphone, held just two centimeters above the surface, listened to the air particles vibrating in response to the tap. The logic here relies on how a plate behaves when it has a hole underneath it. When the concrete is solid and fully supported by the ground, it is stiff and resists bending. But when there is an empty cavity beneath it, that section of the plate becomes flexible, like a trampoline. When tapped, the flexible section vibrates differently, producing a sound with a different character than the solid sections.

The results from the microphone were strikingly clear. The researchers measured the "flexibility" of the plate at every point along their survey lines. Where the concrete was solid, the flexibility remained low and steady. But as soon as the microphone passed over a cavity, the flexibility jumped significantly. The data showed a sharp rise in flexibility exactly where the holes were located, allowing the team to pinpoint the edges of the cavities with high precision. Unlike the radar, the microphone was barely bothered by the steel bars inside the concrete; the sound of the plate bending over the hole was distinct enough to ignore the noise of the steel. The team found that they could not only detect the presence of a hole but also comparatively estimate its size. Larger cavities caused a greater increase in flexibility, and the microphone could clearly distinguish between the 30-centimeter and 60-centimeter holes.

Perhaps the most revealing finding was how the shape of the cavity affected the sound. When the researchers compared the square holes to the circular ones, they noticed that the area of increased flexibility was wider over the square holes than over the circular ones of the same width. This happened because the square shape maintains its full width along the entire length of the survey line, whereas the circular shape tapers off at the edges, creating a narrower zone of flexibility. However, the study noted that while the acoustic sensing method showed a clear advantage in delineating cavity edges, the flexibility differences associated with detailed cavity shapes were negligible. This subtle difference meant that while the researchers could infer the size of the cavity, they could not reliably distinguish between a square or circular shape based on flexibility alone. While the radar struggled to separate the signal of the hole from the signal of the steel, the microphone provided a clean, direct map of where the ground was missing.

The study concludes that while ground-penetrating radar is a useful tool, its ability to map the exact geometry of cavities behind reinforced concrete is limited by the interference of steel bars. In contrast, acoustic sensing using a microphone offers a more reliable way to detect these hidden voids and estimate their dimensions. By listening to how the tunnel wall vibrates, engineers can identify the boundaries of a cavity with much greater accuracy than by looking at radar images alone. This suggests that in the future, combining these two methods could give tunnel inspectors a much clearer picture of the hidden dangers lurking behind the walls, ensuring that the underground arteries of our cities remain safe and sound.