Dead-Leg Corrosion Inspection: From Sparse UT Spots to a Traceable Mapping Workflow

The four readings that missed the thinnest area

The UK Health and Safety Executive documented a catastrophic rupture in an 8-inch vertical relief-line dead leg. Four locations had been routinely tested, yet highly localized internal wall loss developed between them. The line released an estimated 75 tonnes of hot hydrocarbon when it failed. HSE's lesson is specific and practical: for a high-risk dead leg, reliance on spot ultrasonic testing alone is not adequate, and the examination regime should test a significant proportion of the component for remaining wall thickness.

This was not a PetroBot inspection or a MagRover case. The failed line was also below PetroBot's currently published 10-inch-and-above pipe-OD screening boundary. Its relevance is the evidence gap: a correct value at one thickness-monitoring location does not validate the wall between locations. A useful plan therefore moves through five controlled steps—identify risk features, define coverage, collect qualified data, declare uncertainty, and send the result to the responsible inspector or engineer.

Define the dead leg before selecting an inspection technique

A process-piping dead leg is a branch or section in which flow is absent, stagnant or much lower than in the connected system. Examples include blinded branches, normally closed bypasses, spare-pump lines, relief piping, high-point vents, drains, bleeders, level bridles and redundant piping that remains connected. The label alone does not establish the damage mechanism: service chemistry, temperature history, phase behavior, deposits, orientation and operating cycles still matter.

Start with a dead-leg register linked to the line number and current isometric. Record the branch origin and termination, orientation, material, nominal size and schedule or design basis supplied by the owner, insulation and heat tracing, service, operating history, status and prior inspection data. Verify that the line still exists and that a supposedly isolated branch has not changed duty. HSE places elimination of non-essential dead legs ahead of continued inspection; where removal is practicable, the owner should evaluate it rather than turning repeated examination into a permanent substitute for design action.

Why localized wall loss defeats sparse monitoring points

Contact UT measures the travel time response from a small interrogated region. It can provide reliable local thickness evidence under a qualified procedure, but it does not describe unsampled wall. An equal-spaced set of acceptable readings can therefore coexist with a severe minimum between the points. Thermal gradients, liquid-vapour interfaces, condensate locations, deposits, low points, intermittent wetting, mixing zones, clamps and supports are reasons to challenge the coverage plan—not proof that corrosion exists at a predicted position.

The planning question is not merely ‘how many readings?’ It is whether the proposed valid-data footprint can address the credible localized pattern and consequence. When the morphology may change sharply, broader scanning or locally denser characterization can reduce the gaps, subject to probe response, surface condition and geometry. Any remaining gap is residual uncertainty that must survive into the deliverable and the owner's decision.

Start with the integrity question and governing program

Define the decision before mobilization. Baseline characterization, anomaly follow-up, interval support, pre-turnaround planning, repair verification and data for an engineering assessment require different coverage and traceability. Identify the governing owner procedure and applicable piping code, the responsible inspector, NDT authority, minimum-thickness basis, acceptance route and escalation contacts. Where API 570 and API RP 574 form part of a process-piping program, the owner and inspector remain responsible for applying their current requirements; this guide does not reproduce paid clauses or certify compliance.

Keep facility process piping distinct from regulated transmission pipelines and in-line inspection. ‘Pipeline inspection’ is PetroBot's service taxonomy, while this article's engineering scope is external examination of accessible process piping. Also separate the decisions: screening prioritizes further work; spot UT measures selected positions; a qualified scan collects a denser set of thickness evidence; interpretation considers signal and condition; engineering disposition determines continued service, repair, replacement or further assessment.

Build a risk-feature map before building a UT map

Review the isometric, branch orientation and elevation together with expected liquid level, temperature profile, process chemistry, condensation, deposits, insulation, heat tracing, supports, clamps, drains, injection or mixing points, prior values and known leaks. Mark each feature as a planning hypothesis. A predicted interface or low point deserves attention, but only acquired evidence can establish where the measured minimum lies.

Define a coordinate convention another inspector can reconstruct: line and isometric ID, a fixed datum, axial distance, clock position and named component or obstruction references. Mark priority bands and likely inaccessible zones before field work. The field record must preserve deviations when surface condition, geometry or site controls prevent the planned route. PetroBot publicly describes location-correlated inspection data, but does not publish the correlation mechanism or spatial accuracy; this guide therefore makes no encoder, coordinate or automation claim.

Choose the method mix, not a robot by default

The responsible inspection and NDT team should select methods from the damage question, pipe size and duty, geometry, temperature, surface condition, access, required sensitivity and consequence. A robot is one possible carrier for camera and UT data; it is not the starting assumption. Screening methods can help prioritize areas, while a different method may be needed to characterize a local indication or reach behind insulation, supports and complex features.

MagRover may fit when the pipe is ferromagnetic, its exposed surface and travel path are accessible, magnetic adhesion can be maintained, the UT or visual area can be prepared, the OD meets the published 10-inch-and-above boundary, and the project has a defined acquisition objective. The boundary is a first screen only. Current public evidence does not establish traversal of elbows, reducers, tees, welds, flanges, clamps, supports or branches, or inspection through intact insulation and cladding.

Qualify deployment prerequisites and safety controls

Before acquisition, verify the material, OD and curvature, coating, roughness, scale, temperature, magnetic adhesion, surface preparation, probe seating, couplant control, route, tether behavior and access around the whole intended area. Setup verification, reference material or velocity, calibration checks, repeatability and validity rules belong in the approved procedure. Curved or non-concentric geometry, pitting and rough surfaces can change the contact and back-wall response; an unstable or inconsistent signal must be rejected or escalated rather than converted into a convenient number.

The task risk assessment must address process hazards, isolation or operating state, stored energy, leaks, hot surfaces, falling objects, work at height, access equipment, dropped objects, tether management, couplant collection, electrical controls, communications and recovery. Shutdown, insulation removal, scaffolding and live-line work are project-specific. No hazardous-area, ATEX, PESO or intrinsic-safety claim is made here because an approved certificate and exact project configuration were not available for this article.

Stop or reassess for loss of adhesion, unstable coupling, temperature outside the approved project envelope, unsafe site conditions, inaccessible geometry, inconsistent back-wall response or unverified coating behavior. Robotic access can change where personnel stand, but it does not eliminate competent field staff, site permits, manual prove-up or every exposure.

Use a controlled acquisition workflow

Begin with an as-found visual record of the accessible route, surface, coating, supports and obstructions. Confirm the datum and orientation before collecting thickness data. Perform the procedure-defined setup checks, then follow the approved scan or spot plan while linking visual media, UT values and exceptions to the same location convention. If the route changes, record the deviation at the point it occurs rather than redrawing the planned map after the fact.

At each position or scan segment, assign a validity state. Preserve attempted data that was rejected because of coupling, surface, geometry or signal criteria. Confirm a low or abrupt value with the approved repeat process; where the procedure and geometry support it, locally densify around the indication. Retain the original and repeat result, setup state, surface condition and exact location. Crack-like responses, severe pitting, geometry echoes and inconsistent signals require an appropriate qualified reviewer and may need complementary NDT.

Turn achieved coverage into a reconstructable data set

A credible coverage record distinguishes axial and circumferential intent, attempted route, accepted readings and unresolved gaps. Do not interpolate across a clamp, rejected signal or uninspected band to make a complete-looking map. ‘Full coverage’ is defensible only when the approved scope, achieved route and validity rules support that phrase. A no-data zone beside a high-risk feature is a decision item, not an empty cell to ignore.

The following states are engineering guidance for a decision-ready record; they are not claimed as automatic MagRover software fields or a standard PetroBot deliverable. The exact schema, raw-data retention and correlation method must be confirmed in the quotation and approved procedure.

Control data quality and confirm meaningful minima

Data quality is a chain, not a cleanup step. Record the scope revision, equipment and procedure identifiers, operator and reviewer roles, material and geometry inputs, surface and coating state, temperature and couplant conditions, setup checks, datum, route, reading status, visual-media IDs and limitations. Use consistent units and preserve precision as reported by the qualified system. If historic positions, setup or coating state cannot be reconstructed, do not force a corrosion-rate calculation from non-comparable values; establish a new baseline or seek engineering direction.

An apparent minimum needs confirmation because coupling loss, roughness, curvature, coating behavior and geometry echoes can mimic or obscure a response. Repeat according to the approved procedure and use another probe, access condition or NDT method when needed. Do not silently replace the original result with a cleaner number. The record should show what changed, which result was accepted, who reviewed it and what uncertainty remains.

Interpret thickness evidence without crossing into disposition

A qualified thickness map can show where measured wall is lower across the valid sampled area. It cannot, by itself, establish whether loss originated internally or externally, identify the active corrosion mechanism, characterize a crack, or prove that an inaccessible area is sound. Correlate the thickness evidence with process service, temperature and phase history, external visual and coating condition, prior comparable data and any complementary inspection.

The responsible owner, API 570 inspector or other governing integrity role supplies the required minimum-thickness basis and decides corrosion rate, remaining life, inspection interval, rerating, repair, replacement or fitness for service. API RP 571 may inform damage-mechanism review, and API 579-1/ASME FFS-1 may be selected for an engineering assessment, but citing those documents does not make PetroBot's acquisition code-compliant or turn the scan into an engineering conclusion.

Specify a decision-ready dead-leg inspection deliverable

PetroBot's controlled pipeline-service material supports high-level outputs including UT readings, visual records, corrosion or metal-loss observations, inspection notes and a maintenance-planning report. A project may need a more detailed evidence package. The fields below are recommended scope requirements, not a promise that they are included in every standard PetroBot deliverable; confirm them contractually before mobilization.

Keep four classes visibly separate. A surface condition seen in a referenced image is an observation. A valid UT result at a referenced position is a measurement. A statement that the pattern may merit expansion is interpretation. A decision to continue service, inspect again, repair or assess is owner disposition. That separation lets a reviewer audit what the instrument established and prevents recommendations from being mistaken for measured fact.

• Asset and line identifiers, isometric revision, owner-supplied material, size, schedule and design inputs
• Scope objective, applicable procedure and program references, responsible acquisition and review roles
• Coordinate legend with planned-versus-achieved coverage and every deviation
• UT data with units, location, validity status, confirmed minima and required setup or verification records
• Linked visual stills or video, surface and geometry observations, and obstruction references
• Rejected results, inaccessible and uninspected zones, stop-work events, limitations and residual questions
• Prioritized follow-up without invented acceptance limits, repair instructions or fitness conclusions

Pre-quotation checklist and practical next step

A useful request for quotation gives the inspection team enough information to challenge robotic suitability before it promises a method. Unknowns can be recorded as survey risks; hidden assumptions cannot. Ask who owns insulation removal, surface preparation, access, couplant control, isolation and work-at-height provisions, and identify the condition that would trigger another access method or NDT technique.

• What integrity decision must the inspection support, and who approves acquisition, interpretation and disposition?
• Which current line list, dead-leg register and isometric define the component and datum?
• What are the pipe OD, material, schedule or nominal thickness, service, temperature, orientation and operating history?
• Where are insulation, coating, supports, clamps, welds, branches, valves and other route constraints?
• Which credible damage patterns and risk features drive planned axial and clock-position coverage?
• What procedure, personnel-qualification system, setup checks, validity rules and complementary methods govern the work?
• How will valid, rejected, inaccessible and uninspected coverage be correlated and reported?
• What site, area-classification, isolation, access, tether, couplant, recovery and stop-work controls apply?
• Which deliverables and raw or retained records are contractually required?
• What result or coverage gap triggers expanded inspection, engineering assessment or physical modification?

Frequently asked questions

Technical references

• UK HSE CEMHD2-2019 — Catastrophic rupture of dead-leg pipe-work
• API standards plan — API 570, RP 574, RP 571, RP 580, RP 581 and API 579-1/ASME FFS-1
• API — API 570 fifth-edition announcement
• ISO 16809:2025 — Non-destructive testing: ultrasonic thickness determination
• ASTM E797/E797M-21 — Manual ultrasonic pulse-echo contact thickness measurement
• ISO 9712:2021 — Qualification and certification of NDT personnel