Ultrasonic thickness measurement tells you how much metal is left in a submerged structure without taking it out of the water. The diver holds a probe against the piece and the instrument calculates the thickness from the echo's time of flight. The difficulty isn't in taking the reading — it's in preparing the surface and knowing whether the number is credible.
A visual inspection tells you whether a structure has corrosion. It doesn't tell you how much section has been lost, which is exactly what an engineer needs to know to decide whether the piece is still fit for purpose. That question is answered by ultrasonic thickness measurement, one of the best-paid and least understood tasks in underwater inspection work.
The principle is simple. The probe sends out an ultrasonic pulse that passes through the material, bounces off the far face and comes back. The instrument measures how long that takes and, knowing the speed of sound in that material, calculates the thickness. What complicates matters isn't the physics — it's the real world: coatings, uneven corrosion, rough surfaces and an operator wearing gloves in murky water.
It's a technique that opens doors. A diver who can also inspect and measure is worth more than one who only does mechanical work, and it's one of the clearest routes for progression within the underwater inspector profile.
How the equipment works
The instrument is a digital thickness gauge in a watertight housing. There are two common setups: the self-contained unit the diver carries, with its own display and direct reading, and the surface-readout unit, where the diver handles only the probe and the inspector sees the result on deck. The second is preferable when the person interpreting the reading isn't the one in the water, which is the more common case.
One advantage of working underwater: there's no need to apply couplant gel, because the water itself acts as the couplant between the probe and the piece. That simplifies the job and removes a common source of error found in surface work.
Single-echo versus multi-echo
This is the distinction with the biggest practical consequences. In single-echo mode the instrument measures from the transmit pulse to the first back-wall echo, so if there's paint or a coating, its thickness gets added to the metal's and the reading comes out high. Measuring correctly in this mode means getting down to bare metal.
In multi-echo mode, or echo-to-echo mode, the instrument measures the interval between two consecutive back-wall echoes. Since the path through the coating is common to both, it cancels out, and the result is the thickness of the metal even with the paint still on. This is huge underwater, because it lets you measure without destroying the coating and without spending hours grinding.
Multi-echo has a cost: it needs the far face to return clean, repeatable echoes, and that fails when the back wall is heavily corroded, irregular, or the piece is too thin. In those cases you have to fall back on single-echo with full preparation, and then you really do need to clean down to bare metal.
The part that decides the result: the surface
The most common mistake in this technique isn't about equipment, it's about preparation. A layer of marine growth, a crust of loose rust or a rough surface prevents decent acoustic contact and produces readings that either don't appear or come out wrong. Preparing properly takes far longer than measuring.
Cleaning is done with hand or power tools, and the aim is a firm, reasonably smooth surface with no loose corrosion products left on it. One especially treacherous issue is pitting corrosion: if the inner face is pitted, the reading depends on whether the probe lands on a pit or beside it, and two points a centimetre apart can give very different values.
The answer to that isn't measuring more precisely, it's measuring more often. You lay out a grid of points over the area of interest and take readings at all of them, and what you report is the set as a whole, with its minimum and its spread. A single value doesn't describe a corroded structure.
Calibration and traceability
Before signing off on a measurement campaign, you need to calibrate the instrument against a reference block of known thickness and the same material you're about to measure, because the speed of sound changes with the material. Calibration is repeated at the start, periodically during the work and at the end, and it's that final check that lets you decide whether the day's data is valid.
Temperature also has an effect, a small one but not zero, and it's worth keeping in mind in cold water. The same goes for recording which probe was used, since dual-crystal probes perform better on corroded back walls, while single-crystal probes are the ones that allow multi-echo mode.
Then there's traceability, which is what separates a useful data point from a stray number. Every reading has to be tied to an identifiable position: element, face, elevation from a fixed reference and orientation. If another team wants to know next year whether the corrosion is progressing, they need to measure in exactly the same spot. Without a reference, the campaign is useless for comparison.
Common mistakes that invalidate a measurement
The first is accepting the first reading that comes up. A thickness gauge almost always gives you a number, even when coupling is poor, and that number can be a multiple or a fraction of the real one. The correct practice is to look for stability: repeat the placement several times and keep the value that recurs, not the first one that appears.
The second is measuring with the probe tilted. The beam has to enter perpendicular to the surface; a noticeable tilt lengthens the path and distorts the echo. With gloves, current and poor visibility, keeping the probe square to the surface is harder than it sounds, and it's a skill you have to train.
The third is not distinguishing between general loss and local loss. A structure can have an acceptable average thickness and still have a pit that compromises it, or the other way round. Reporting only the average hides exactly the defect that matters.
And the fourth, the most costly, is measuring where it's convenient instead of where it matters. The critical zones are usually the worst ones: the tidal and splash zone, the joints, the areas where the cathodic protection is failing, and the points where sediment builds up. If the inspection plan says those are the points, then those are the points, however awkward they are.
Where it's used and what the data is for
The most classic use is on fixed offshore structures: platform tubular members, braces and nodes, where the aim is to track how corrosion develops over the years and feed structural integrity calculations. There the measurement isn't a one-off check but a time series, and its value depends entirely on always measuring at the same points.
The second major use is naval: hulls, tanks and appendages, where an underwater measurement can avoid or shorten a dry-docking. And the third is civil and industrial: quay piles and sheet piles, submerged pipelines, gates and metal fittings on coastal installations, where the question is usually whether a piece will last until the next scheduled shutdown.
In every case, the diver doesn't interpret the data. Whether a section loss is acceptable is decided by the asset owner's engineering team, applying the criteria of the standard or code the structure was designed to. The inspector's job is to deliver defensible measurements, with their preparation, calibration and location documented, so someone else can audit them.
It's also worth knowing what this technique doesn't answer. Thickness measurement tells you how much material is left at the point measured; it doesn't detect a crack that doesn't reduce thickness, nor does it tell you whether a closed tubular member has flooded through a crack, which is a different check with its own equipment. Confusing what each technique measures is a common mistake when defining an inspection scope, and it's usually discovered late, once the report no longer answers the question that prompted the campaign.
What qualification you need
Two things combine here. On one side, the commercial diving qualification required by the country where the job is, without which you don't enter the water. On the other, a qualification in inspection or in NDT, which is what shows you know how to take and interpret a reading.
Internationally, the most recognised route for a diver is underwater inspector, with certification schemes that combine visual inspection and thickness measurement. The requirements, validity and renewals depend on the body issuing the certificate, so it's worth checking them directly before enrolling in anything. You can see the general picture on the page about certifications.
It's worth the effort. It's a specialty that pays better, that puts less physical wear on the body than purely mechanical work, and that extends your working life, because the value lies in judgement rather than in brute strength. If you're starting out, the logical sequence is to train at a recognised school, build up real working hours, and add inspection later, once you're moving comfortably underwater.
Frequently asked questions
Do you need to remove the paint to measure thickness underwater?
Not always. In multi-echo mode the instrument measures between two consecutive back-wall echoes and the coating cancels out, so you can measure without removing it. It's only when that mode doesn't give reliable readings, because of corrosion or irregularity on the far face, that you need to prepare down to bare metal and measure with single-echo.
Do you need couplant gel underwater?
No. The water acts as the couplant between the probe and the piece, which removes a step and a common source of error in measurements taken in the dry.
Why do two nearby readings give different values?
Almost always because of uneven corrosion on the far face. If the inner surface is pitted, the result depends on whether the probe is sitting on a pit or beside it. That's why you measure on a grid and report the set together with its minimum value, not a single isolated reading.
What qualification do you need to take these measurements?
A commercial diving qualification valid in the country where the job is, plus an inspection or NDT qualification that demonstrates competence in the technique. The specific requirements and their validity are set by the certifying body, so it's worth checking with them directly.
Does thickness measurement replace visual inspection?
No, it complements it. Visual inspection locates where the problem is and measurement quantifies how much section is left. The usual approach is to use the first to direct the second toward the areas that really matter.
Published 6 September 2026 · By Arthur More, professional saturation diver
Photo: U.S. Department of Defense Current Photos · dominio público via Openverse
This article is part of our in-depth guide to underwater welding: the pay, the training and the myth — the complete picture, with sources.
This article is part of our in-depth guide to underwater welding: the pay, the training and the myth — the complete picture, with sources.
This article is part of our in-depth guide to underwater welding: the pay, the training and the myth — the complete picture, with sources.
This article is part of our in-depth guide to underwater welding: the pay, the training and the myth — the complete picture, with sources.