Hi Folks,
I’m troubleshooting a puzzling safety-device failure and would appreciate the community’s insight.
Plant & Duty
-
Unit: Continuous deodoriser for palm oil (7 horizontal vessels, vacuum stripping).
-
Heating medium: HP steam loop (~50 bar g @ 240 – 260 °C) from a natural-circulation boiler.
-
Rupture disc: REMBE BT-KUB-1, DN25, Inconel®, rated 95 bar g @ 298 °C (batch 2202109). Installed upstream of the safety valve on the boiler outlet; vent stub rises ~4 m, then to atmosphere.
-
Transmitter: 54 PI 240 (0 – 100 bar g, 1 s scan rate).
Trend data show repeated shutdown sequences where pressure falls to the transmitter’s low cut-off (≈ 0.3 bar g) for 6 – 18 h, then ramps slowly back to 60 – 70 bar g during startup.
Two hypotheses
-
Vacuum-induced low-cycle fatigue
-
During cooldown the closed steam circuit pulls near-full vacuum (transmitter can’t read below 0 bar g).
-
The reverse-acting dome flexes inward; next startup reverses the load to +70 bar g.
-
After a dozen cycles the burst point drifts down to ~20 bar g and the disc finally opens.
-
Evidence: identical pressure cycles every shutdown; no collateral pipe damage; fracture surface shows beach-marks.
-
-
Single water-/steam-hammer spike
-
Condensate in the 4 m vent stub is accelerated by incoming steam, creating a very short 100 + bar pulse that the PI tag misses.
-
The disc ruptures cleanly at the spike, PI plot only captures the 20 bar tail.
-
Evidence: some startup traces show 30–50 bar oscillations; hammer in similar lines has reached 5–10 × design pressure in literature.
-
-
Have you seen reverse-acting discs derate by >70 % purely from vacuum/pressure cycling? How many cycles did it take?
-
Can a short hammer pulse really shear a KUB-type disc without leaving dents or pipe-support damage nearby?
Thanks in advance for your thoughts!