Case Study: FCC reactor shell, 14 ft ID, 1¼Cr-½Mo, 28-year service — Petroleum refining

Asset & context

Petroleum refining — US Gulf Coast (Texas). FCC reactor shell, 14 ft ID, 1¼Cr-½Mo, 28-year service. Crew of 3 for 4 days.

The challenge

The 28-year-old fluid catalytic cracking (FCC) reactor at a 240,000 bpd Gulf Coast refinery was three months into its 5-year inspection interval when a Level II UT technician flagged anomalous thickness readings on the lower shell course. Spot UT showed 9.8 mm remaining wall against a nominal 16.6 mm — a 41% loss concentrated in a 600 mm band roughly 1.2 m above the reactor cone-to-shell weld. The unit was scheduled to restart in 96 hours. The operator needed an inspection that could (a) map the full extent of wall loss, (b) characterize whether this was internal high-temperature H2S/H2 attack or external CUI, and (c) give the API 510 inspector enough data to file a fitness-for-service determination under API 579-1/ASME FFS-1 §5.[1][2]

Approach

Method selection: why PAUT corrosion mapping won over conventional UT

Conventional 0° UT with a 2.25 MHz dual-element probe is the workhorse for spot thickness checks, and it is what flagged the anomaly. But spot UT cannot bound a corrosion patch — it only tells you what is under the probe. With 96 hours on the clock and a need to defend the wall-loss footprint to a third-party API 510 inspector, the inspection lead called for phased-array corrosion mapping (PAUT-CM) using encoded raster scanning per ASME V Article 4 Mandatory Appendix IV and API RP 577 §11.[3][4]

PAUT-CM gives full-volumetric coverage of the scan footprint at a typical 1 mm × 1 mm resolution, with C-scan output that maps minimum remaining wall directly onto the shell. The crew specified a 5 MHz, 64-element linear-array probe with 0.6 mm pitch — selected because the 32 mm aperture clears the resolution requirements in ASME V T-451.1 for the 16.6 mm nominal wall.[3] The probe was paired with an Olympus OmniScan X3 64 channel unit and a wireless scanner crawler rated for 80°C surface temperature, which mattered because the shell was still at 71°C four hours after steam-out.

A secondary call was made to deploy TOFD on the shell-to-cone girth weld in parallel. The wall-loss patch was inside the heat-affected-zone band of the most recent weld repair (2019 turnaround), and TOFD under ASME V Article 4 Mandatory Appendix III gives the highest probability of detection for HAZ cracking driven by high-temperature hydrogen attack (HTHA) — a damage mechanism specifically called out in API RP 941 §5 for 1¼Cr steel above 232°C.[5][6]

Procedure: raster scan, gridding, and reference points

Surface prep was the first time-sink. The external shell carried 3-4 mm of intumescent fireproofing under cladding. The crew cut a 1.2 m × 1.5 m inspection window centered on the spot-UT anomaly, removed cladding, scraped fireproofing to bare steel, and abraded to a 3.2 µm Ra surface finish per the procedure qualification under SNT-TC-1A §6.4.[7] Surface temperature at start of scan was 64°C; the procedure was qualified up to 80°C contact, so no cooling delay was needed.

The scan plan gridded the window into six 200 mm × 250 mm tiles. Each tile was raster-scanned with the probe in linear electronic scanning mode at 32-element aperture, 8-element step, generating a C-scan at 1 mm encoder resolution. Total scan time per tile: 7 minutes. Calibration was performed on a step-wedge block traceable to NIST per ASME V T-462, with 80% screen-height response verified on a 1.5 mm flat-bottom hole side-drilled reference reflector.[3]

A 250 mm × 250 mm overlap zone between tiles was scanned twice to verify positional repeatability. Drift between repeat scans came in under 0.2 mm at the deepest pit — well inside the ±0.5 mm acceptance criterion the procedure called for. The crew completed all six tiles plus the TOFD girth-weld scan in a single 11-hour shift.

Findings: a high-temperature H2S attack signature, not CUI

The C-scan composite showed a 480 mm × 320 mm patch of internal-side wall loss with minimum remaining wall of 9.4 mm — slightly worse than the spot-UT had indicated. The loss profile was smooth, dish-shaped, and concentrated on the inside diameter. No external pitting was visible after fireproofing removal, and no moisture or chloride contamination was found behind the cladding. This rules out CUI per NACE SP0198 §A2.2 — CUI signatures are external-side, irregular, and chloride-stained.[8]

The damage signature matched high-temperature sulfidic corrosion per API RP 939-C §4.2: smooth, broad, internal-side metal loss in 1¼Cr-½Mo above 260°C in a hydrogen-bearing sour stream.[9] Operating temperature at this band was 312°C and H2S partial pressure was 0.18 MPa — well inside the corrosion envelope mapped in the McConomy curves. The 2019 weld repair likely disturbed the chromium-enriched sulfide scale, accelerating local thinning in the HAZ.

TOFD on the girth weld came back clean — no HTHA cracking, no HAZ cracks, no lack-of-fusion. This was critical: HTHA would have forced an immediate shutdown under API RP 941 §6.3, regardless of remaining wall.[5] With cracking ruled out and the corrosion mechanism characterized, the path forward was a fitness-for-service Level 2 assessment under API 579-1 §5.4 using the local thin-area (LTA) method.[2]

Disposition: FFS Level 2, run-to-turnaround with monitored intervals

API 579-1 Level 2 FFS on a 480 mm × 320 mm LTA with 9.4 mm minimum remaining wall, against an MAWP of 2.4 MPa and a t_min of 7.8 mm calculated per ASME VIII Div 1 UG-27, returned a remaining strength factor (RSF) of 0.91. API 579 §5.4.2.2 allows continued operation at MAWP when RSF ≥ 0.90, with periodic re-inspection.[2][10] The unit was cleared to restart at full MAWP with a 12-month PAUT-CM re-inspection interval written into the integrity operating window (IOW).

The inspector also wrote three corrective actions into the next turnaround scope: (1) replace the affected shell section with 347H stainless (sulfidation-resistant per API RP 939-C Table 1) on a 1.5 m × 1.5 m insert plate; (2) install three permanent UT sensors at the LTA centroid for between-turnaround monitoring; (3) revise the corrosion control document to flag the post-weld region as a higher-priority CML.[9]

Total inspection cost: $47,500. Avoided cost: a 5-day unplanned shutdown at this FCC unit averages $2.1M/day in lost throughput and a further $1.1M in catalyst replacement if the unit comes down hot. Conservative cost avoidance was $4.2M against the worst-case scenario where the spot-UT reading was treated as a shutdown trigger without further characterization.

Methods deployed

Defects found

Outcome

Fitness-for-service Level 2 assessment under API 579-1 §5.4 returned RSF = 0.91. Reactor cleared to restart at full MAWP with 12-month PAUT re-inspection. Repair scoped into next turnaround as 347H insert plate. Avoided 5-day unplanned shutdown.

Cost avoidance

$4.2M against worst-case unplanned shutdown ($2.1M/day FCC throughput loss × 2 days + $1.1M hot-catalyst replacement, per the operator's 2023 reliability planning data)

Certifications required

Crew qualifications: