UT Corrosion-Rate Trending: Make Every Thickness Reading Comparable

Correct arithmetic can still produce the wrong trend

A two-reading corrosion rate starts with a difference in thickness over elapsed time. That arithmetic is useful only after the inspection team qualifies both inputs. Move one point around the pipe, shift a grid by half a cell, compare a local scan minimum with an older spot reading, or change the sound path without reconciliation and the result may describe measurement variability or a different patch of wall rather than metal loss with time.

The UK Health and Safety Executive describes ultrasonic thickness measurement as a point measurement. It also ties measurement position and grid pitch to the corrosion damage of concern. That matters because broad thinning, isolated pitting and directional erosion do not occupy the same footprint. A sparse point plan may track general thinning and miss a narrow pit. A dense scan may find a lower local value simply because it examined more surface.

Treat change detection as a measurement-system problem. Pacific Northwest National Laboratory examined process variability, operator methods and differences between reported minima, averages and raw UT data in its Hanford work. Hanford tanks differ from process piping, yet the practical lesson holds: matching units do not make two plausible report values comparable.

Define the same metal before collecting another value

A tag such as TML-14 or CML-203 is an identifier, not a complete location. A repeatable location needs four layers: the exact asset and component, a stable datum, local coordinates and a bounded area that states what was actually examined. Build those layers into the work pack and report so another crew can reconstruct them without relying on faded paint or memory.

On pipe, use an owner-approved reference such as a numbered girth weld, branch centerline or permanent datum, then record axial distance and circumferential position with an explicit clock convention and viewing direction. On a vessel shell, combine course, seam, nozzle or fixed datum with elevation and circumferential coordinates. Retain units, sign convention, orientation, drawing revision and photographs that show the datum and examination boundary.

FHWA bridge guidance requires thickness locations to be tied to reference points that can be readily re-established and retained for future inspections. A process plant follows its own governing program, but the location-control principle transfers cleanly: a dimension from a stable physical feature survives longer than a chalk cross or an unreferenced screenshot.

Match the comparison type to the damage

Point-to-point comparison is the narrowest case. Use it when the physical point can be re-established, both readings are valid and the expected damage morphology makes that point meaningful. Even then, the procedure must address probe placement, local roughness and measurement variability. A point beside a pit is not a measurement of the pit.

Scan-to-scan or area-to-area comparison can be stronger for distributed or spatially variable damage, but only when the windows, resolution basis, exclusions and reported statistic remain compatible. A minimum, mean and percentile answer different questions. Expanding the scan window increases the chance of finding a lower local value. Do not call that difference wall loss until a qualified reviewer reconciles the populations.

Point-to-area comparison needs special care. An old spot and a new scan minimum may both be valid measurements, yet they do not represent the same statistic. Keep them in the history, label the distinction and let the governing procedure decide whether any equivalent comparison is defensible. If it is not, the scan becomes the new controlled baseline.

Re-establish the datum, not the old paint mark

Before surface preparation, photograph the reference feature and verify it against the drawing or previous record. Recreate the coordinate origin from that feature, then mark the planned window. If insulation, coating renewal or maintenance removed the old marks, do not estimate their former location. Rebuild the coordinates from the retained dimensions.

Use the same orientation every time. A three o'clock position changes if one report looks with flow and another looks against flow. State the viewing direction, clockwise convention and whether axial distances run upstream or downstream. For vertical equipment, state whether elevation is measured from a course seam, foundation datum or another fixed reference.

The bounded area belongs in the record with the nominal point. That boundary tells the next reviewer whether a prepared patch, grid or encoded scan covered the same wall. Where support steel, a weld crown, nozzle, coating transition or unsafe access interrupts the boundary, carry the interruption as a coverage exception rather than shifting the route silently.

Qualify the complete UT measurement chain

Location identity does not rescue an invalid ultrasonic response. Confirm that the material and geometry suit the method, the surface and coupling are acceptable, the correct velocity basis is used and the back-wall response meets the approved procedure. Record the instrument, probe, acquisition mode, settings, calibration and verification references, couplant, surface preparation, coating condition and temperature context needed for review and repetition.

ISO 16809:2025 covers principles for ultrasonic thickness determination. ISO 16831:2025 separately addresses characterization and verification of equipment used for thickness determination. That separation is useful in the field: an instrument can pass its verification while a particular reading still fails because the surface, geometry, coupling, material response or signal acceptance is unsuitable.

ASTM E797/E797M-21 covers manual contact pulse-echo thickness measurement. Use it only where it belongs; do not transfer a manual procedure or its limits to an encoded or robotic acquisition by assumption. The project procedure and NDT authority must define the applicable configuration, checks, acceptance rules, repeat-reading treatment and uncertainty convention.

Keep the raw evidence that the contracted scope requires: A-scans, encoded scan files, verification records, rejected responses, images and setup records where available. A report value stripped from its signal, setup and validity state is difficult to challenge and harder to repeat. Agree the retention set before mobilization rather than discovering after the survey that the decision needs data that were never saved.

Reset the baseline when the metal or surface changes

A repair, replacement, grinding campaign, weld build-up, deposited lining, coating renewal or geometry change can break continuity. Preserve the pre-change history and record the event, date, extent and supporting documents. Then establish a post-change baseline. Do not subtract across the event as if the same wall and sound path remained untouched.

Some changes need qualified reconciliation rather than an automatic reset. Surface preparation may remove scale and improve coupling without changing nominal wall, while aggressive grinding may remove metal. A coating change can alter the measurement approach. A relocated support can expose one area and conceal another. The responsible reviewer decides whether records remain comparable and documents the basis.

Process changes also matter even when the metal is physically unchanged. A new injection point, flow regime, temperature cycle or water chemistry can move the credible damage zone. Keep the old trend for history, but redesign coverage around the revised mechanism hypothesis instead of extending a familiar point plan into the wrong location.

Use MagRover where the route supports valid acquisition

MagRover uses magnetic adhesion on ferrous surfaces and can collect UT spot readings, UT scans and visual evidence on a suitable route. PetroBot publishes 10 inches OD and above as the first pipeline screen and describes its pipeline service as providing location-correlated inspection data. That screen starts a suitability review; it does not guarantee access, coverage or repeatability on a specific asset.

Check material, diameter, curvature, coating and surface condition, temperature, magnetic adhesion, probe seating, welds, attachments, route continuity, access, power, water for sensor couplant, tether management and recovery. Define the inspection question, datum, scan boundary, validity rules and required files before deployment. Site classification and operating controls remain part of the project review.

Do not infer a coordinate accuracy, automatic registration method, through-coating envelope or named-feature traversal capability from the crawler's general description. Where the route cannot preserve valid coupling or reach the required wall, mark the boundary and hand the gap to another access plan or NDT method.

Run the repeat survey as a controlled acquisition

Start with a pre-job data review. Load the previous drawing, location register, photographs, raw files, setup details, coverage exceptions, repairs and open dispositions. Flag records that lack a reconstructable datum or valid signal basis before they influence the new route. The field team should know which locations repeat a qualified history and which locations establish a new baseline.

At the asset, confirm identity and as-found configuration. Re-establish the datum, orientation and boundaries. Photograph the setup area before and after surface preparation. Verify the equipment and system as required by the approved procedure, record environmental and surface context, then acquire the planned points or scans without moving the origin to make access easier.

Check quality while remobilization is still possible. Review signal validity, coordinate completeness, achieved coverage, rejected responses, inaccessible zones and unexpected configuration changes. Repeat only under the procedure's rules. Never replace a failed point with a nearby convenient value under the same identifier.

At closeout, preserve lineage. Link every reported value to its asset, location, acquisition file, validity state, setup and campaign. Record who acquired, reviewed and released it under the owner's competence system. The data package should allow an independent qualified reviewer to reconstruct what happened without reverse-engineering filenames.

Pass a comparability gate before subtraction

Run the gate in order. First confirm asset and physical location. Then check for repair, replacement or other configuration change. Qualify both measurement records against the applicable procedure. Compare coverage windows and statistics. Reconcile exceptions and the approved uncertainty treatment. Only then release the pair to the role authorized to calculate and use the trend.

A failure does not erase data. It changes the state. Keep both records, document why they cannot support one trend and decide whether to repeat the examination, use another method, widen the area, create a new baseline or escalate for engineering review. Calling a pair non-comparable is a quality decision, not a missing result.

Keep every evidence state visible

A blank cell is ambiguous. It might mean no work was planned, access failed, the signal was rejected, the component changed or a file was lost. Use explicit states so the history cannot turn absence into apparent coverage. At minimum, distinguish planned, acquired-valid, acquired-rejected, inaccessible, changed configuration, non-comparable history, qualified comparison and engineering disposition.

These states protect the next decision. A rejected response can trigger surface preparation or another technique. An inaccessible area can drive an access change. A changed component starts a new baseline. A qualified comparison may enter a corrosion calculation, but the calculation and disposition still belong to the authorized inspector or engineer under the owner's program.

Use Visualize to organize the inspection history

Visualize is PetroBot's beta platform for asset history, thickness mapping, 3D visualization and report export. It can organize campaigns and spatial context when the supplied data model retains the fields the project needs.

The platform does not remove the comparability gate. Require asset identity, datum, coordinates, coverage, acquisition mode, statistic, validity state, physical-change events, file lineage and review status in the buyer's data contract. Confirm the actual import and export behavior for the project rather than assuming that every recommended field is a standard schema.

Qualified roles still validate UT, reconcile campaigns, apply uncertainty rules, calculate rates and decide inspection interval, remaining life, fitness for service, rerating or repair. A thickness map displays evidence; it does not prove that two pixels or two values represent unchanged metal under compatible conditions.

Interpret the trend with mechanism and consequence

Once the data pass the gate, review the result against the expected mechanism, neighboring locations, coverage limits, operating history and prior campaigns. A single apparent acceleration may justify a targeted repeat before it justifies a new long-term rate. A stable average can still hide a local minimum if the damage is pitting or erosion concentrated outside the repeated points.

API's public announcement for the fifth editions of API 570 and RP 574 places inspection-location selection, result evaluation, follow-up, fitness-for-service, rerating and repair within the piping integrity framework. API 510 provides corresponding pressure-vessel context. Where the owner and jurisdiction adopt those documents, authorized roles apply the licensed requirements and site procedures.

Keep the responsibilities explicit. The NDT authority controls the examination procedure and signal-validity basis. Inspectors and data reviewers qualify the record within their assigned roles. Corrosion and integrity engineers evaluate mechanism, trend use and engineering action. Operations supplies process history and controls safe access. PetroBot's acquisition and data tools support that chain; they do not own the final disposition.

Make safety and recovery part of the data plan

A repeatable coordinate does not justify an unsafe route. Assess operating temperature, pressure boundary condition, releases, site classification, work at height, dropped objects, surface preparation, utilities, tether path, access restrictions and emergency response under site controls. Define stop-work and recovery actions before equipment moves onto the asset.

Plan for loss of adhesion, coupling, communication or power without improvising around live equipment. Decide where the route ends, how the device is recovered and which areas will remain inaccessible. A clearly recorded no-data zone is safer and more useful than an unplanned maneuver that compromises people, equipment or evidence.

Personnel competence must match the method, asset and decision. ISO 9712:2021 describes a qualification and certification framework for industrial NDT personnel, while API's QUTE-TM program uses hands-on performance demonstration for manual, non-encoded thickness examination. Apply the owner's required competence system without transferring a manual qualification to robotic work or claiming a certification that has not been established for the people and configuration used.

Specify the repeat-inspection package before mobilization

A concise scope prevents most history gaps. State the decision, damage hypothesis, asset population, stable datums, coordinate conventions, examination boundaries, procedure, equipment and probe configuration, surface preparation, verification checks, coverage and validity rules, required raw files, evidence states, review roles and closeout format. Include the events that force a new baseline or engineering review.

Finish the scope with exceptions. Name concealed or inaccessible areas, weld and attachment restrictions, alternate techniques, recovery limits and the owner for each unresolved zone. Then require an as-achieved coverage record instead of accepting the planned grid as proof of completed work.

Frequently asked questions

Technical references

• UK HSE — Inspection and Non-Destructive Testing guidance
• FHWA Technical Advisory T 5140.31 — re-establishable NDE measurement locations
• API — fifth-edition API 570 and RP 574 announcement
• API — API 510 11th-edition announcement
• ISO 16809:2025 — Non-destructive testing: ultrasonic thickness determination
• ISO 16831:2025 — Verification of ultrasonic equipment for thickness determination
• ASTM E797/E797M-21 — Manual ultrasonic pulse-echo contact thickness measurement
• ISO 9712:2021 — Qualification and certification of NDT personnel
• API QUTE-TM — Qualification of Ultrasonic Testing Examiners
• PNNL — Evaluation of UT Wall Thickness Measurements and Measurement Methodology