A Growing Bearing Defect on the Plant's Only Air Compressor
Overall vibration on the drive-end bearing of the plant's only process-air compressor has risen from a baseline of 2.2 mm/s to 4.1 mm/s over eight weeks, roughly 0.3 mm/s per week on the last five readings. The envelope spectrum shows a growing peak at the outer-race defect frequency, bearing housing temperature has climbed from 65 C to 78 C, and the last oil sample reported iron at 45 ppm against a 30 ppm flag, with ISO particle count up two codes. The compressor runs 24/7 and feeds the only air header; a failure would stop two of three production lines. A reconditioned spare motor is in stores. A new motor is a six-week lead time. A four-day planned shutdown is two weeks out; the next one after that is five months away.
1. Interpret the data. Is this a sensor artifact or a real defect, and how much running life is realistically left?
Read the worked answer
Three independent technologies agree, so this is a real, progressing outer-race bearing defect, not instrumentation:
- Vibration level: 2.2 to 4.1 mm/s in eight weeks, with an accelerating trend of 0.3 mm/s per week.
- Envelope/demodulation spectrum: a growing peak at the outer-race defect frequency is direct evidence of a spalled raceway - it typically appears before the overall level moves much.
- Temperature: +13 C at the housing is a lagging but serious indicator; friction is now generating measurable heat.
- Oil analysis: iron at 45 ppm against a 30 ppm flag plus a two-code particle increase confirms active wear debris generation.
Life estimate, stated honestly: a naive straight-line extrapolation to a 7.1 mm/s alarm gives about ten weeks, but the trend is not linear - a spalled bearing degrades faster as debris circulates and heat rises. The realistic planning window is weeks, on the order of two to four, not the five months to the next shutdown. The temperature rise is the part of the data most likely to fail suddenly and without further warning.
2. What do you recommend, and what interim controls do you put in place?
Read the worked answer
Recommendation: replace the motor at the four-day shutdown in two weeks using the reconditioned spare. Do not run to failure and do not stretch to the five-month window.
Reasoning:
- There is no header redundancy: an unplanned failure stops two of three lines plus whatever the air-down recovery costs. That risk dwarfs a planned four-day outage.
- The spare is already in stores, so the lead time that would normally force a run-to-failure decision does not apply. The asset with the long lead time is the air supply, not the motor.
- Two weeks is inside the credible remaining-life window, with margin.
Interim controls:
- Increase condition monitoring from monthly to daily vibration and temperature readings, trended against a written limit (e.g., overall 5.5 mm/s or housing 85 C triggers escalation).
- Pre-stage the spare, rigging, alignment tools, and a work plan; confirm the craft labor and the shutdown permit in advance.
- Brief operations on what an air loss means and what the safe shutdown sequence is; have a contingency plan for the two lines.
- Keep the oil clean - filter or change oil to slow debris-driven acceleration, but treat this as a delay tactic only, not a repair.
3. Suppose the housing temperature reaches 92 C next week, before the shutdown. What is the decision, and how does it change the earlier plan?
Read the worked answer
The written trigger has been crossed, so the plan changes from 'scheduled' to 'controlled now':
- Do not wait for the shutdown. At 92 C the bearing is close to seizure, and a seized drive end can damage the rotor, the stator, or the coupling - turning a motor swap into a compressor rebuild.
- Choose the outage rather than have it chosen for you. Coordinate a planned, short-notice swap using the spare and the pre-staged plan. A managed interruption of a few hours, announced to production, is far cheaper than a failure at 3 a.m. that also takes out the header.
- Escalate the decision, do not absorb it. Production planning, maintenance, and operations management decide the timing together; the reliability data defines the risk, but the business decides the exposure.
- What does not change: the diagnosis, the spare, and the need to replace rather than rebuild in place. What changes is the tolerance for monitoring - the 'watch it another week' option is gone.
The general principle the exam is testing: condition data creates decision triggers, and the trigger must be written down before the trend is alarming, so that the decision is made on evidence instead of on the pressure of the moment.