Bearing Failure Diagnosis: Evidence, Causes and Prevention

Beginner → Intermediate → Advanced learning path

Prerequisites: basic rotating-equipment awareness, approved isolation practices and familiarity with the machine's normal operating duty. A learner should finish able to:

• Beginner: recognize warning signs and record operating context without declaring a cause too early.
• Intermediate: collect repeatable vibration, temperature, lubricant and process evidence at defined points.
• Advanced: build and test competing fault hypotheses, preserve removed-part evidence and define post-repair acceptance checks.

1. Confirm the symptom before naming the cause

“Bearing noise” is a symptom statement, not a diagnosis. First identify which bearing position is involved, when the symptom occurs and whether it changes with speed, load, temperature, product, belt condition or another operating variable. Record the asset ID, drive and non-drive end, bearing arrangement, speed, process load, operating hours, recent work and alarm history.

Rule out obvious external sources such as a loose guard, rubbing fan, damaged belt, soft foot, coupling problem, process cavitation or structure-borne vibration from adjacent equipment. A stethoscope, vibration tool or thermal camera can support screening when used by a competent person, but one reading should not be treated as proof of a specific failure mode.

2. Establish a comparable operating baseline

A trend is useful only when measurements are comparable. Mark permanent measurement points where practical and use the same instrument, sensor orientation, mounting method, operating state and units. Record speed and load with each route reading. Comparing a fully loaded machine at normal speed with an unloaded machine during warm-up can create a false trend.

• Use the healthy sister machine or the same bearing position's historical data as context—not as an automatic alarm limit.
• Record whether the machine is starting, steady, changing load or stopping.
• Note ambient conditions, recent washdown, lubrication and maintenance interventions.
• Retain raw readings and instrument settings where the monitoring program requires them.

Alarm limits, acceptable temperatures and vibration severity depend on the machine, bearing, speed, mounting, measurement method and applicable site/OEM criteria. This guide intentionally provides no universal numeric trip limit.

3. Collect evidence through four channels

Vibration and sound

Collect readings in the approved radial and axial directions at repeatable points. Overall vibration can show that machine condition changed, while spectrum, envelope or high-frequency analysis may help a trained analyst distinguish bearing-related activity from imbalance, misalignment, looseness, gear mesh or electrical influences. Check sensor mounting and repeat an unexpected reading before escalating the diagnosis.

Temperature

Trend the same surface point under comparable duty and allow for normal warm-up. A rise may result from excess or insufficient lubricant, excessive load, wrong fit, seal friction, misalignment or heat transferred from the process. Temperature alone cannot identify which mechanism is present. Verify emissivity, reflected heat, distance and access limitations when using infrared measurement.

Lubrication and contamination

Verify lubricant identity, quantity, compatibility, relubrication method, interval and storage/handling controls against approved requirements. Inspect for leakage, discoloration, water, particles, purge condition and blocked relief paths without introducing contamination. Preserve a representative sample when analysis is justified; label the sampling point, date, operating condition and whether the sample was taken before or after fresh lubricant was added.

Process, mechanical and electrical context

Review load, speed, motor current, belt tension, coupling condition, alignment history, shaft/housing fits, seal condition and recent process changes. On VFD-driven motors, suspected electrical erosion needs evidence such as characteristic surface damage plus a competent electrical assessment; do not infer shaft-current damage from VFD presence alone.

4. Build competing hypotheses from the evidence

List plausible causes before deciding which one is most likely. Then identify what observation would support or contradict each cause:

• Lubrication problem: lubricant condition, purge path, relubrication history, temperature trend and contact pattern should agree with the hypothesis.
• Contamination or moisture: seal condition, ingress path, washdown exposure, corrosion or indentation evidence and lubricant sample should connect the damage to a credible route.
• Misalignment or incorrect fit: alignment records, shaft/housing measurements, load-zone pattern, fretting, creep or uneven wear should support the conclusion.
• Overload or process influence: correlate the event with production load, jams, belt tension, hydraulic force, fan fouling or another measurable duty change.
• Installation damage: review tools, force path, heating control, cleanliness, handling, internal clearance and commissioning evidence.
• Electrical erosion: preserve surface evidence and involve an appropriately competent person to evaluate grounding, common-mode effects and the OEM-recommended mitigation.

A repeated failure at the same position should widen the boundary of investigation. Include shaft geometry, housing condition, foundation, lubrication delivery, seals, coupling, driven equipment, operating sequence and maintenance method. Replacing the bearing with a larger one or shortening the interval is not a root-cause correction unless the engineering evidence supports it.

5. Preserve evidence during shutdown and removal

Photograph the arrangement before disturbance and mark orientation, bearing position and load direction. Record fastener condition, fits, clearances and lubricant distribution as required by the approved procedure. Use the correct puller, press, induction heater or fitting tool so removal does not create fresh damage that is later mistaken for the original fault.

Do not clean, grind or discard the failed bearing before the investigation owner decides what evidence is needed. Protect rings, rolling elements, cage, seals and lubricant from mixing or corrosion. If the failure has safety, warranty, insurance or repeat-loss significance, maintain identification and handover records and seek OEM or specialist analysis.

6. Correct the cause and verify the repair

Translate the supported cause into a controlled action: correct the lubricant and delivery method, restore sealing, repair shaft or housing fits, align the train, correct belt load, remove looseness, resolve the process overload or apply an OEM-approved electrical mitigation. Document any parameter or design change through site change control.

After commissioning, verify guarding, lubrication, rotation, noise, leakage, temperature, vibration, motor current and process performance at agreed operating stages. Capture an initial post-repair baseline and repeat the measurement after a defined running period based on criticality and OEM/site requirements. A quiet startup is encouraging, but closure requires evidence that the original symptom and its cause have not returned.

Minimum bearing investigation record

• Asset, bearing position, bearing designation and failure/event date.
• Speed, load, process state, ambient condition and operating hours.
• Measurement point, direction, instrument, units and reading timestamp.
• Vibration, temperature, sound, motor-current and lubricant observations.
• Recent lubrication, alignment, installation, seal and process-change history.
• Photographs before removal, orientation marks and removed-part observations.
• Competing hypotheses, supporting/contradicting evidence and confirmed cause status.
• Corrective action, approval, commissioning checks and post-repair trend review.

Limitations and escalation

Screening instruments and visual patterns do not replace competent vibration analysis, dimensional inspection, lubricant analysis, electrical investigation or metallurgical examination when those are needed. Escalate rapid deterioration, unstable vibration, smoke, severe heat, damaged guarding, looseness, shaft/housing damage, repeated failure, safety risk or uncertainty about continued operation. Site/OEM shutdown criteria take priority.

Authoritative further reading

• SKF: Root Cause Bearing Damage Analysis — damage-analysis scope, lubrication, installation and environmental factors.
• SKF: Bearing-housing condition monitoring — periodic and continuous monitoring context.
• Fluke: Bearing monitoring — monitoring causes, symptoms and route-based versus online approaches.
• NTN: Bearing technical training topics — installation, fits, lubrication and failure analysis.

External references support further study. Always use the current manual and technical limits for the exact bearing, housing, lubricant and machine.