A Single High Hardness Reading in a Wet H2S Separator
A 1-1/4 in. SA-516 Gr. 70 wet-H2S separator was PWHT'd at 1150 °F for 75 minutes. A portable Leeb hardness gauge, converted to HBW, gives 168 on the long seam, 152 on shell plate, 196 on a repair weld, and a single reading of 214 in the heat-affected zone of a 4 in. nozzle-to-shell weld, taken on a ground but curved surface near the toe. The client specification invokes NACE MR0103 with a 200 HBW maximum for carbon steel welds, while the fabricator's engineer argues that ISO 15156 permits 250 HV and that a Leeb reading is not a valid test. The fabricator proposes to grind the 214 location until the gauge reads below 200 and then paint.
1. Which criterion governs, and is the 214 reading a valid basis for rejection?
Read the worked answer
The contract governs. The specification invokes NACE MR0103, which sets a 200 HBW maximum for carbon steel welds and heat-affected zones. The engineer's ISO 15156 argument is a real distinction, since that standard permits higher hardness for carbon steel, but a less restrictive general standard cannot override a stricter contractual requirement. If the engineer believes the specification is wrong, the route is a documented concession, not a unilateral reading of ISO 15156.
That said, one converted Leeb reading is not a hardness survey. Rebound instruments are sensitive to surface finish, curvature, component mass and rigidity, and to the conversion correlation used; a ground but curved surface near a toe is close to a worst case. Conversion scatter between HL and HBW is large enough that 214 may not be 214.
Disposition: neither accept nor reject on that value. Re-prepare the surface, take several readings per location, use the instrument's own verified correlation, and confirm the suspect area with a second method (UCI or a field metallographic replica evaluated on a bench instrument) before deciding.
2. If the elevated hardness is confirmed, what remedies are available and what must each one address?
Read the worked answer
Local re-PWHT of the nozzle weld is the metallurgically correct fix: it tempers the martensite and lowers hardness at the fusion line and coarse-grained zone where the reading sits. It must be engineered, not improvised: heaters and insulation around the nozzle and shell, thermocouples on the weld and on the shell away from the heater, a hold inside the qualified range, and controlled heating and cooling. Afterward, re-survey hardness and check whether the new cycle pushes aggregate time past the 80% threshold in QW-407.2.
Grinding is only acceptable if it truly removes the hard microstructure. A hard zone can extend a tenth of an inch or more below the surface and along the fusion line, so polishing until the gauge reads low may simply move the probe to softer metal while leaving hard material at the toe. Grinding is also limited by remaining wall thickness against design plus corrosion allowance, and the excavation must be smoothly blended.
Either way, the cause matters: wet H2S service makes an unrelieved hard zone a sulfide stress cracking site, so the fix must remove the condition, not the measurement.
3. Why might only the nozzle heat-affected zone be hard, and what would you review to find out?
Read the worked answer
The pattern is diagnostic. A seam that reads 168 while a nozzle reads 214 suggests that region did not see the same thermal history. Candidate causes: the nozzle weld was made or repaired after the furnace cycle, which is a common field failure; the reinforcement lagged the furnace so its actual temperature was below the recorded shell temperature; the nozzle weld used a different, higher-heat-input or higher-carbon-equivalent procedure; or the reading is an artifact of curvature and surface condition at that specific location, which the shell readings were not.
Review the welding and repair chronology against the furnace chart, the thermocouple location map, the WPS and heat input used for the nozzle weld, base metal certificates including carbon equivalent, and any post-PWHT attachment or repair welding. Compare the same nozzle with a second method before concluding. If the weld was made after PWHT, the answer is straightforward: heat treat it, and fix the work sequence so post-PWHT welding is controlled in future.