Vacuum Transfer Line Wall Loss: Naphthenic Acid or Sulfidation (and What to Build Back With)?
An 8 in carbon steel vacuum transfer line carries vacuum tower bottoms to the visbreaker at 680°F. Feed assays give TAN 4.6 (potentiometric, whole crude) and 2.9 wt% sulfur; the line has a designed velocity of 14 ft/s at the elbow and downstream straight run, and some low-velocity sections near the tower outlet. After 5 years the long-radius elbow has thinned from 0.322 in to 0.145 in, with sharp-edged, deep grooves and a smooth, bright, apparently scale-free metal surface. The low-velocity upstream straight run shows broad, more uniform thinning with a dark, adherent, layered scale. UT readings one elbow back show only 0.006 in loss. No hydrogen probe data are available. The unit plans to replace the elbow during the next shutdown and is debating between 9Cr-1Mo and a Mo-bearing stainless for the whole line.
1. Assign the mechanisms to the two damage locations and justify the assignment from conditions and morphology. What does the elbow-to-straight-run contrast tell you?
Read the worked answer
Two mechanisms are acting simultaneously, and the morphology separates them:
- Elbow and downstream high-velocity run: naphthenic acid corrosion. Sharp-edged grooves, a smooth/bright metal surface with no protective scale, localized to 14 ft/s regions, at 680°F with TAN 4.6. Naphthenic acids attack metal directly and strip the sulfide scale that would otherwise protect the surface; the products are oil-soluble, so nothing builds back. Attack is strongly velocity/shear sensitive, which is why it concentrates at the elbow and just downstream of it and not in the low-velocity run.
- Upstream low-velocity straight run: sulfidation. Broad uniform thinning with an adherent, layered, dark scale is the sulfidation signature: H2S from the 2.9 wt% sulfur reacts to form iron sulfide, and the rate is largely temperature- and H2S-partial-pressure-driven, much less velocity sensitive.
The 0.006 in one elbow back versus 0.177 in at the failed elbow is the decisive clue: a purely sulfidation-driven rate would not vary by a factor of ~30 over one fitting. The contrasting severity is a velocity effect, which points to naphthenic acid as the dominant local mechanism and tells you the thinning is not a uniform 'high-sulfur crudes' problem - it is a flow-regime problem that will recur wherever shear is high.
2. Evaluate the 9Cr-1Mo versus stainless debate for this line. What material would you specify where, and why is a single alloy for the whole line the wrong question?
Read the worked answer
9Cr-1Mo is the correct answer to sulfidation and the wrong answer to naphthenic acid.
- For sulfidation, chromium content governs resistance and 9Cr is a large step up from carbon steel at 680°F - appropriate and economical for the low-velocity, scale-forming upstream run.
- For naphthenic acid at TAN 4.6 and 680°F, chromium content in the ferritic range is not the controlling variable. 9Cr offers little protection; Mo-bearing austenitic stainless (316L/317L) or nickel-alloy cladding (625-type) is the recognized choice, with 317L or 625 preferred as TAN and velocity rise. The high-velocity elbow and the downstream run are therefore stainless or clad territory.
So the correct answer is a split specification by mechanism and location: 9Cr (or better) where sulfidation dominates at low shear, and 317L/625-clad where naphthenic acid plus shear dominates - with the transition located by a velocity/shear survey, not by drawing a line at a flange. Two further constraints to state: velocity must be reduced or geometry improved at the elbow regardless of alloy (no alloy is immune to very high shear), and the material choice must be validated against the actual TAN of the worst credible feed, since 4.6 is an assay value and not a guarantee. Confirm that the stainless choice is compatible with any downstream sulfur/hydrogen environment, since 300-series is not acceptable in some high-temperature H2/H2S services.
3. What inspection and monitoring program would you implement across the line and in the unit, given that one elbow showed 30 times the loss of its neighbor?
Read the worked answer
- Profile the whole circuit, not the known bad fitting. Dense UT grid (or RT profile) at every elbow, tee, reducer, and 2–3 diameters downstream of each, plus a baseline grid on the straight runs. The 30× neighbor-to-neighbor contrast means the corrosion is distributed unpredictably relative to geometry - sample spacing must be tight enough to catch a locally severe fitting.
- Record fluid velocity and shear per segment from the hydraulic model and correlate with measured loss; re-rate or re-pipe segments where velocity is out of the acceptable range for the selected alloy.
- Monitor TAN and sulfur on each crude slate, not just on the annual assay - naphthenic acid corrosion risk moves with feed changes, and a TAN excursion on top of a thin elbow is a short path to a loss-of-containment event.
- Track temperature, since naphthenic attack requires a threshold temperature (roughly >400°F) and sulfidation is strongly temperature dependent; any increase in transfer temperature accelerates both.
- Set the re-inspection interval from the measured rates, with the worst-fitting rate (not the average) used for the calculation, and treat any elbow within a defined margin of minimum required thickness as a scheduled replacement rather than a monitored item.
- Do not use general-thickness coupons as the primary indicator - they will confirm sulfidation and completely miss the naphthenic grooving.