Low Vibration, Severe Bearing Damage: How 10kHz+ Signals Revealed a Hidden Bearing Failure

A roller press had been operating for years, while its low-frequency vibration RMS remained stable for several months without triggering any alarms.

However, during equipment inspection and disassembly, engineers discovered severe fatigue spalling on the outer raceway of a critical support bearing:

The spalling area reached approximately 30 cm in length, 15 cm in width, and 6–7 mm in depth.

More importantly, throughout the entire degradation process, regular low-frequency vibration waveform and spectrum analysis did not show obvious abnormal characteristics.

The key indicators that captured the bearing degradation process at an earlier stage were high-frequency acceleration signals above 10 kHz and high-frequency envelope spectrum analysis.

This case provides a complete view of bearing damage progression in a low-speed, heavy-duty industrial equipment — from initial defect initiation to severe spalling — and demonstrates the value of high-frequency vibration analysis for early detection of bearing degradation.

Equipment Background: Hidden Bearing Damage in Low-Speed, Heavy-Duty Equipment

This case comes from a roller press installed in a cement production line.

As a critical asset in cement production, the roller press operates under low-speed and high-load conditions for extended periods. Its key supporting components typically use heavy-duty, high-capacity double-row spherical roller bearings.

The transmission system mainly consists of:

•  Motor system;

•  First-stage parallel gearbox;

•  Second-stage planetary gearbox;

•  Roll shaft support bearings.

Due to complex operating conditions and significant load variations, early-stage damage in heavy-duty bearings typically features:

•  Low initial impact energy;

•  Weak fault characteristics;

•  Delayed response in low-frequency vibration signals.

A total of 16 vibration sensors were installed to continuously monitor the motor, gearbox, and critical roll shaft bearings. RONDS Remote Diagnostic Center provided 24/7 condition monitoring and analysis services.

Based on equipment structure parameters, operating conditions, and characteristic frequency analysis, the fault location was identified as the floating roll shaft non-drive end bearing (measurement point 8H).

The rotational frequency at this bearing position was only 0.303 Hz, making early-stage damage detection particularly challenging using regular low-frequency vibration indicators.

From Initial Pitting to Severe Spalling: Tracking the Complete Bearing Degradation Process

Based on long-term continuous monitoring data, including trend analysis, time waveform analysis, spectrum analysis, and envelope spectrum analysis, the entire bearing degradation process was divided into three stages.

Stage 1: Early Outer Race Pitting (Sep 17, 2020 – Apr 30, 2021)

During this stage:

•  The overall high-frequency acceleration trend remained stable;

•  Weak impacts corresponding to BPFO (Ball Pass Frequency Outer Race) occasionally appeared in the time waveform;

•  Slight noise floor energy accumulation appeared above 10 kHz in the high-frequency spectrum;

•  BPFO and its harmonics were already visible in the >10 kHz envelope spectrum, but with relatively low energy.

Meanwhile:

•  No obvious abnormality was observed in the low-frequency acceleration waveform;

•  No clear bearing fault characteristics appeared in the low-frequency spectrum.

At this stage, only initial fatigue pitting had developed on the outer raceway. The impact energy was too weak to significantly influence low-frequency vibration signals.

However, high-frequency acceleration analysis had already captured the weak impacts generated by the early-stage defect.

Stage 2: Outer Race Spalling Development (Apr 30, 2021 – Aug 13, 2021)

As the damage continued to develop:

•  The high-frequency acceleration trend began to increase gradually;

•  Periodic impact clusters related to BPFO appeared in the time waveform;

•  Individual impacts gradually merged into continuous rounded impact clusters.

Spectrum analysis showed:

•  Continuous increase of noise floor energy above 10 kHz;

•  Significant growth of BPFO and harmonic energy in the high-frequency envelope spectrum.

However:

•  The low-frequency acceleration waveform still did not show stable impact characteristics;

•  Bearing fault frequencies remained difficult to identify clearly in the low-frequency spectrum.

At this stage, localized pitting on the outer raceway gradually expanded and merged, forming larger spalling areas.

When rolling elements passed through the damaged area, double-peak or multi-peak impact responses occurred. However, because the impact energy was still mainly concentrated in the high-frequency range, the low-frequency vibration response remained limited.

Stage 3: Severe Spalling Growth and Failure Evolution (Aug 13, 2021 – Sep 24, 2021)

As the bearing entered the severe degradation stage, the damaged outer raceway area continued to expand. Rolling elements passing through the spalled area generated stronger and more complex impact responses.

High-frequency signals continued to increase:

•  High-frequency acceleration levels continued rising;

•  Numerous sharp impact peaks appeared within the original rounded impact clusters;

•  Multiple compound impacts occurred within a single rotation cycle.

Meanwhile:

•  Noise floor energy above 10 kHz continued increasing, with a wider frequency coverage;

•  The number and energy of BPFO components and harmonics in the high-frequency envelope spectrum increased significantly.

As damage severity increased, low-frequency signals finally began to show identifiable bearing fault characteristics:

•  Intermittent BPFO-related impacts appeared in the low-frequency waveform;

•  BPFO and its harmonics became observable in the low-frequency spectrum.

However, when bearing faults become clearly visible in low-frequency vibration, the bearing has usually already entered a severe damage stage, and the optimal maintenance intervention window may have been missed.

This degradation process demonstrates that bearing damage does not occur suddenly. Instead, it develops progressively from weak impacts, to spalling growth, and finally to severe damage.

High-frequency analysis can capture and continuously track this degradation process at an earlier stage, while low-frequency vibration characteristics typically become apparent only after the damage has significantly progressed.

Inspection Verification: 30 cm Severe Spalling Damage

During maintenance inspection, the floating roll shaft non-drive end bearing was found with severe outer raceway spalling:

•  Length: approximately 30 cm;

•  Width: approximately 15 cm;

•  Depth: approximately 6–7 mm.

The inspection findings were highly consistent with the vibration analysis results, validating the effectiveness of high-frequency analysis for early bearing degradation detection.

Why Didn’t Low-Frequency Vibration Provide Early Warning?

Regular vibration monitoring typically focuses on:

•  Velocity;

•  Low-frequency acceleration;

•  RMS trend variation.

These indicators are highly effective for identifying:

•  Unbalance;

•  Misalignment;

•  Mechanical looseness;

•  Advanced mechanical abnormalities.

However, for early bearing fatigue damage:

The initial defect generates extremely short-duration, high-frequency impact signals.

As damage progresses:

•  High-frequency impact energy gradually increases;

•  Eventually, the structural response becomes strong enough to influence low-frequency vibration.

Therefore:

Stable low-frequency vibration does not necessarily mean a healthy bearing. It may simply indicate that the damage has not yet developed enough to affect overall vibration levels.

Key Insights for Industrial Maintenance Teams

(1) Do not rely only on overall vibration levels to evaluate bearing health

For large low-speed heavy-duty equipment, even bearings with significant spalling may maintain stable low-frequency RMS values for an extended period.

(2) High-frequency signals above 10 kHz provide an important window for early bearing damage identification

High-frequency acceleration waveform analysis combined with high-frequency envelope spectrum analysis can capture impact characteristics during bearing degradation and continuously track the progression from initial pitting, spalling development, to severe damage.

(3) Waveform evolution helps identify bearing damage stages

Weak intermittent high-frequency impacts → Early pitting initiation;

Rounded continuous impact clusters → Spalling formation and expansion;

Multiple sharp compound impacts per rotation → Severe spalling development and high-risk condition.

(4) Continuous trend analysis provides more value than a single spectrum snapshot

Long-term high-frequency envelope spectrum energy trends can continuously record the degradation process from healthy operation to severe damage.

Conclusion

For large low-speed heavy-duty equipment in industries such as cement, mining, and steel, major reliability risks rarely occur suddenly. Instead, they develop progressively over long-term operation.

Regular vibration indicators primarily reflect:
If the equipment has already developed an abnormality?

High-frequency analysis above 10 kHz helps further determine:
If the equipment is progressing toward potential failure?

Through continuous condition monitoring, intelligent analysis, and expert diagnosis, enabling a transition from reactive maintenance to predictive maintenance, can we ultimately unlock the true value of condition monitoring.