Cracking in a Sour Water Flash Drum: SSC, HIC, or SOHIC?
A sour water flash drum, SA-516 Gr. 70, 0.75 in shell, fully post-weld heat treated at 1150°F, has operated 6 years at 165 psig. Vapor analyzes 2.2 mol% H2S (partial pressure about 3.6 psi); the water phase is pH 7.6 and contains ammonia and phenols, with no detectable cyanide. A wet fluorescent magnetic particle inspection during a turnaround found linear indications at a nozzle-to-shell weld toe. Follow-up ultrasonic shear-wave examination found planar, surface-parallel, stacked reflectors linked by short through-thickness segments, forming a ladder pattern normal to the welding residual stress. Hardness surveys over the weld, HAZ, and base metal read 165–185 HBW. A separate 3 in diameter blister with a surface-breaking fissure was found on the shell course well away from any weld.
1. Rank the candidate wet H2S mechanisms for this drum and justify the ranking with the specific evidence given.
Read the worked answer
Most likely: SOHIC (stress-oriented hydrogen-induced cracking), with HIC/blistering present as its precursor. Least likely: SSC.
- HIC/blistering is confirmed at the shell course - surface-parallel planar reflectors and a blister are the classic signature of atomic hydrogen diffusing into the steel, recombining at inclusions/laminations, and pressurizing internal planes. It is not hardness dependent.
- SOHIC is the right label at the weld. The distinguishing feature is the through-thickness, ladder-like array of HIC cracks aligned normal to the residual/applied tensile stress, located in the HAZ near a weld where restraint is highest. SOHIC requires both a HIC-susceptible steel and high stress, which matches a nozzle weld in a PWHT'd drum.
- SSC is unlikely. Carbon steel SSC in wet H2S is a hardness-controlled phenomenon, generally requiring weld/HAZ hardness above about 22 HRC (roughly 237 HBW). At 165–185 HBW the hardness route is essentially closed off.
- Conditions are consistent with wet H2S cracking per the common screening charts: pH 7.6 with roughly 3.6 psi H2S partial pressure falls in the region where HIC/SOHIC are credible; cyanide is absent, which removes a strong promoter but does not protect.
The practical consequence: because SOHIC is present, this is a steel-quality and stress problem, not just a hardness problem, and it is far less forgiving than isolated SSC.
2. The maintenance planner proposes grinding out the indications at the nozzle and returning the drum to service, arguing that PWHT and low hardness make the drum 'not susceptible.' How do you respond, and what does the cracking pattern imply about the remaining population?
Read the worked answer
Reject the argument on two counts:
- PWHT and hardness do not address HIC/SOHIC. PWHT relieves some residual stress and lowers hardness, which is the right control for SSC - but HIC and SOHIC occur in steels at normal hardness. What governs susceptibility is steel cleanliness: sulfur level, inclusion type and shape, calcium treatment/desulfurization practice, and carbon equivalent. A clean, HIC-resistant plate may be essentially immune; a dirty, banded plate could crack at the same hardness. Ask for the original mill certificates, sulfur content, and inclusion control practice before accepting any 'not susceptible' conclusion.
- Grinding is not a repair for SOHIC. The ladder of cracks is a stacked array that extends normal to the surface; local grinding removes the visible indication while leaving the through-thickness linked array in the remaining wall. The correct response is to size the indication through-wall, and where the depth or density is significant, cut out and repair or replace the section with a demonstrated HIC-resistant steel and a qualified weld procedure with hardness verification.
The unaffected shell-course blister is the more important finding. It demonstrates that the steel itself is susceptible independent of weld stress, so the nozzle is not an isolated defect. Assume the population is affected and inspect the entire drum rather than repairing one joint.
3. Define the inspection and integrity plan you would require before this drum is returned to wet H2S service.
Read the worked answer
- Full-coverage examination, not sample-based. Wet fluorescent MT on all welds (both sides where accessible) after grit blasting, plus manual or automated UT shear-wave/TOFD on weld HAZs and heat-affected base metal. HIC is a planar, surface-parallel target that is missed by volumetric RT in many orientations and by spot UT.
- Blister and lamination survey of shell courses and heads, including UT scanning for internal planar reflectors away from welds, with sizing and mapping of every blister found.
- Fitness-for-service assessment of any remaining indications (API 579-type evaluation), plus a written acceptance basis for continued service - 'acceptable to operate' with documented sizing and a re-inspection interval, not an undocumented pass.
- Root-cause material review - sulfur, inclusion shape control, and banding from the original plate; if the plate is not HIC-resistant, plan replacement or upgrade rather than perpetual repair.
- Process controls: reduce H2S partial pressure where feasible, control pH, keep water phase from concentrating, eliminate cyanide sources, and verify no dead legs or low-flow zones where sour water can accumulate.
- Re-inspection interval shorter than the observed time to crack and calibrated to the deepest retained indication; document a monitoring basis if any indication remains in service.