Thermal oil fire from a failing pump bearing, water mist did not work (Pride of Hull)
EL-0129 · Solved · rating 0 · occurred 20 Oct 2020
Source: /cases/el-0129
What happened#
Ro-ro passenger ferry, outbound in the Humber. Thermal oil (Texatherm 32, winter set point 170 °C) carried heat from the economisers and a boiler. Circulation pump #1 was on duty. The fire panel showed zone "Oil treatment pumps", relayed to the ECR as the "fuel treatment room", so the EOOW went there. The fire was in the aft ER.
Hi-Fog water mist did not control it. The ship lost power, anchored, and the crew put the fire out with CO2. Nobody was hurt.
Diagnosis#
- Seat of the fire: pump #1 (
55 kW,3565 rpm). - Impeller-end bearing failed first. The drive-end bearing collapsed and its outer race spun in the housing (about
1,200 °C), then fused to it. - Mechanical seals and circlips failed, so oil reached the hot spot.
- Impeller-end bearing seat pitted and possibly slightly worn. Bearings were genuine Allweiler.
- Only
432 hsince last overhaul. - Oil flash point
135.5 °C(Sept 2020), below the140 °Climit (new:220 °C). Not the ignition cause, but a missed warning.
Root cause#
The impeller-end bearing failed for a reason the examination could not confirm; the pitted seat may have affected the fit. The drive-end bearing then collapsed and its spinning outer race made a hot spot far above the oil's auto-ignition temperature. Failed seals gave the oil a path to it.
Pump #1 had 12 overhauls in 15 years for bearings or seals (pump #2: three). In 2017 the overhaul interval was cut from 18 to 12 months; failures continued. Monthly duty/stand-by changeovers were not always done, so #1 ran far more hours (changed to 6000 running hours on 14 Oct 2020).
Other findings (not causes): after a 2017 failure the spray guard was boxed in, the pumps overheated, and the side panels were removed, which may have limited containment. The Sept 2020 oil report gave no warning; Chevron said the label did not identify the equipment. Earlier results had flagged it.
Timeline / response#
Times UTC+1.
| Time | Event |
|---|---|
| 20:35 | Fire alarm. EOOW goes to the "fuel treatment room". |
| — | Low-flow alarm; changeover to pump #2. |
| — | Hi-Fog M7 starts after 121 s. Smoke triggers M3, M4, M5. Too little pressure for mist. |
| 20:44 | Propulsion lost. Anchored about 10 min later. |
| 20:47 | BA team 1 in. No thermal camera, fire not found. |
| 21:11 | Team 1 out. Team 2 finds the fire 6 min later. |
| 21:23 | CO2 released. |
| — | Two BA entries to forward ER to drain economisers so a generator can run. |
| 00:30 | BA team enters CO2-flooded aft ER to reset breakers. |
| 07:15 (21 Oct) | Power restored; propulsion 10:00. |
Lessons#
- Repeat failures have a cause. Find it (here a pitted seat) and work with the maker.
- Change over duty/stand-by pumps as planned and compare running hours.
- Label oil samples with the equipment name and trend flash point.
- Keep spray guards complete. If a boxed-in pump overheats, fix the cooling. With side panels off, the guard may not contain a spray.
- Know your water mist limits. The pump was sized for one section (
195 L/min); extra sections dropped pressure below mist level. Pumps were not on the emergency switchboard, and the426 Lbreak tank emptied in 2 min because the FW pump needed a local reset. Put this in the fire check cards. - Zone names must match the location; first BA team takes the thermal camera.
Re-entry after CO2 can let air in and re-ignite the fire. Use a door you can keep closed behind you.
Comments · 0
No comments yet.