Mechanical Monitoring

Bearing Condition Monitoring

Bearing condition monitoring tracks the health of rolling element bearings through vibration analysis, acoustic emission, temperature, and lubrication assessment to detect developing faults before they reach a critical stage.

Key Concept

Bearing failures have a predictable progression. They start with lubrication film breakdown or a surface defect, progress through measurable vibration at bearing defect frequencies, then into elevated broadband vibration and temperature, and finally to catastrophic failure. Monitoring detects the early stages, not the final one.

Why Bearing Condition Monitoring Matters

Bearings are among the most common failure points in rotating machinery. A failed bearing in a critical pump, fan, or motor can stop a production line with no warning if the failure develops faster than a periodic inspection schedule can detect.

Condition monitoring provides continuous or periodic visibility into bearing health, giving maintenance teams time to plan a replacement at a scheduled outage rather than reacting to an unplanned breakdown. The cost difference between a planned bearing change and an emergency repair that includes collateral damage to a shaft, seal, or housing is typically an order of magnitude.

The practical argument for bearing monitoring is straightforward: bearings fail in a recognizable sequence. That sequence produces measurable signals at each stage. Capturing those signals early enough to act on them is the entire purpose of the discipline.

Rolling Element Bearing Failure Modes

ISO 15243 classifies rolling element bearing failure modes into six groups. Understanding each mode matters because the appropriate monitoring method and the earliest detectable signal differ by mode.

Fatigue

Sub-surface origin fatigue begins as a crack below the raceway surface, driven by cyclic Hertzian contact stress. It progresses until the crack reaches the surface and produces spalling. Early detection opportunity: high-frequency envelope analysis captures the impacting as the spall develops. Surface origin fatigue starts at surface irregularities, often from inadequate lubrication or surface damage, and produces similar spalling with a shorter time-to-failure. Ultrasound detects the earliest impacting before vibration spectra show clear defect frequencies.

Wear

Abrasive wear results from hard particles in the lubricant grinding the rolling surfaces. It produces a gradual loss of geometry and increased clearance rather than a discrete defect. Oil analysis detecting metallic particles provides the earliest indicator. Adhesive wear (smearing) occurs when the lubricant film collapses under high sliding forces and metal-to-metal contact transfers material between surfaces. Relubrication intervals that are too long and high loads at low speeds are common causes.

Corrosion

Moisture ingress corrosion pits the raceway and rolling element surfaces, which then act as fatigue initiation sites. Environments with washdown, steam, or high humidity require sealed or specially greased bearings along with monitoring for early surface damage. Fretting corrosion develops at the bearing-to-housing or bearing-to-shaft fit surfaces when micromotion occurs under load; it appears as reddish-brown oxide debris and reduces the interference fit over time.

Electrical Erosion

Electrical erosion, known as fluting, results from stray electrical currents passing through the bearing. The current discharges across the lubricant film and erodes the raceway surface in a characteristic washboard or fluting pattern. It is increasingly common in variable frequency drive applications where high-frequency switching generates shaft voltages. Fluting produces a distinctive broadband noise with periodic modulation; insulated bearings or shaft grounding rings are the prevention, and monitoring detects the resulting surface damage through vibration signature changes.

Plastic Deformation

Overload indentation occurs when static or dynamic loads exceed the bearing's capacity, permanently deforming the raceway at rolling element spacing intervals. Each indentation acts as a periodic impactor at the ball pass frequency. Debris indentation from contaminated grease or oil produces irregular denting across the raceway and generates elevated broadband vibration noise.

Fracture

Ring fracture and cage fracture are typically the result of installation errors, severe overloading, or advanced fatigue. Cage fracture in particular can lead to rapid, catastrophic failure as rolling elements lose guidance. Monitoring that has reached Stage 4 signatures should be treated as an immediate-replacement situation to prevent this outcome.

The Bearing Fault Frequency Approach

Each component of a rolling element bearing generates vibration at a predictable frequency when it develops a defect. The defect impacts another surface once per revolution of the defective component, producing a series of impulses at a characteristic rate. These rates are calculated from shaft speed and bearing geometry.

The four principal bearing defect frequencies are:

BPFO: Ball Pass Frequency Outer Race

BPFO is the rate at which rolling elements pass over a defect on the outer race. Because the outer race is stationary in most installations, this is typically the most frequently detected bearing defect frequency.

Formula: BPFO = (N / 2) × f × (1 − (d/D) × cosθ)

BPFI: Ball Pass Frequency Inner Race

BPFI is the rate at which rolling elements pass over a defect on the inner race. Because the inner race rotates with the shaft, BPFI signatures are modulated at shaft frequency and its harmonics, which helps distinguish them from BPFO signatures.

Formula: BPFI = (N / 2) × f × (1 + (d/D) × cosθ)

BSF: Ball Spin Frequency

BSF is the rotational rate of a rolling element about its own axis. A defect on a rolling element generates impacts at twice BSF (once per contact with each race). BSF signatures are typically lower in amplitude than race defect signatures and can be harder to identify clearly in a spectrum.

Formula: BSF = (D / 2d) × f × (1 − ((d/D) × cosθ)²)

FTF: Fundamental Train Frequency

FTF is the rotational rate of the bearing cage. Cage defects generate vibration at FTF and are often accompanied by modulation of other bearing frequencies. FTF is a sub-synchronous frequency, typically 0.38 to 0.48 times shaft frequency for most bearing geometries.

Formula: FTF = (f / 2) × (1 − (d/D) × cosθ)

Variable Definitions

Bearing manufacturers publish these geometric parameters for every part number. Most condition monitoring software calculates defect frequencies automatically from the bearing catalog once the bearing number and running speed are entered.

Calculate BPFO, BPFI, BSF, and FTF for your specific bearing geometry and shaft speed.

Open Bearing Fault Frequency Calculator

Fault Progression Stages

Bearing deterioration follows a recognizable progression across four stages. Each stage is associated with different monitoring signals, different urgency levels, and different recommended actions. Understanding which stage a bearing is in determines how much time remains before intervention is required.

Stage 1 Stage 2 Stage 3 Stage 4 Earliest signs Defect confirmed Accelerating Near failure Condition Severity Time → P Potential failure F Functional failure ← P-F Interval → DETECTION WINDOW Ultrasound / AE Vibration (defect freq.) Temperature Broadband noise Bearing P-F interval model. ISO 10816 zone references. Detection window onset varies with application severity.

Stage 1: Earliest Detectable Changes

The first signal is a rise in ultrasound amplitude as the lubricant film begins to break down or a surface defect starts producing low-energy stress waves. Standard accelerometer-based vibration measurement shows no meaningful change at this stage. Overall vibration levels remain within normal range.

Detection requires ultrasound measurement in the 20 to 60 kHz range. The dBµV amplitude at the bearing housing rises above baseline, often before any defect frequency is identifiable in a standard vibration spectrum. At Stage 1, the recommended action is to check lubrication condition and apply relubrication if ultrasound amplitude normalizes after greasing. If it does not normalize, schedule further investigation.

Stage 2: Defect Frequencies Appear

BPFO, BPFI, BSF, or FTF begin to appear in the vibration frequency spectrum at low but measurable amplitude. The overall vibration level remains within ISO 10816 Zone A or B for the machine class. The bearing is degrading but is not in imminent danger of failure.

Stage 2 is the most actionable point in the progression. The defect is confirmed, the time to failure is still measured in weeks to months depending on load and speed, and a planned replacement can be scheduled for the next available maintenance window. For most industrial bearings operating at moderate load and speed, Stage 2 detection provides 2 to 6 months of planning time, though this range varies considerably with application severity.

Stage 3: Amplitude Growth and Sidebands

Defect frequency amplitudes increase. Sidebands appear around the defect frequencies at shaft frequency intervals, indicating that the defect is modulating the bearing vibration with each shaft revolution. The noise floor across the spectrum rises as broadband energy increases. Temperature at the bearing housing may begin to increase measurably above its historical baseline.

Overall vibration is typically entering ISO 10816 Zone C at this stage. The bearing requires replacement at the next available opportunity. Continued operation through Stage 3 is possible but carries increasing risk of accelerated progression to Stage 4. Monitoring frequency should increase to daily or continuous.

Stage 4: Near Failure

The defect has grown large enough that the sharp, discrete defect frequency peaks begin to disappear into a rising broadband noise floor. The spectrum loses its clean harmonic structure and becomes dominated by broadband vibration energy. Temperature rises sharply. Overall vibration enters ISO 10816 Zone D.

At Stage 4, the bearing requires immediate replacement. The time-to-failure at this stage can be hours to days, not weeks. If the asset is critical, operation should be stopped at the first opportunity. If shutdown is not immediately possible, monitoring should switch to continuous real-time surveillance with a clearly defined abort criterion.

Monitoring Methods Compared

No single measurement technique detects bearing faults at every stage with equal effectiveness. A complete bearing monitoring strategy combines methods based on the target fault stage, the shaft speed, and the lubrication system.

Method What It Detects Best Fault Stage Slow-Speed Suitability
Vibration (acceleration, high-frequency envelope) Defect frequency signatures, sideband structure, broadband noise rise Stage 2 to 3 Moderate. Below 100 RPM, bearing defect frequencies shift below practical detection thresholds and signal energy decreases significantly.
Ultrasound (airborne or contact, 20 to 60 kHz) Lubrication film quality, early impacting from surface defects Stage 1 to 2 Excellent. Stress waves from impacting are detectable at very low shaft speeds where vibration analysis is ineffective.
Temperature (RTD, thermocouple, infrared) Friction heating from lubrication failure or advanced defect Stage 3 to 4 Good. Temperature rise from friction is detectable regardless of shaft speed, though it is a late-stage indicator.
Oil analysis Wear metals in lubricant, contamination, lubricant degradation Stage 2 to 3 Good for oil-lubricated bearings. Not applicable to grease-lubricated bearings without dedicated grease analysis.

For assets running below 100 RPM, ultrasound and temperature are the primary techniques. For assets running above 600 RPM, vibration analysis with envelope detection is the most information-rich method. Between 100 and 600 RPM, combining ultrasound and vibration provides the best coverage.

Lubrication as the First Line of Defense

Lubrication-related causes account for the majority of premature bearing failures. The failure mechanism varies: insufficient grease volume leaves the rolling surfaces in contact without an adequate film; the wrong grease type may not provide the required viscosity at operating temperature; contamination introduces abrasive particles into the contact zone; and relubrication intervals that are too long allow the existing grease to oxidize and lose its film-forming properties.

Ultrasound measurement is the most practical tool for confirming that a bearing is receiving adequate lubrication. The procedure is straightforward: take an ultrasound reading at the bearing housing before starting relubrication, add grease in small increments while monitoring the dBµV level continuously, and stop adding grease when the ultrasound amplitude returns to or below the pre-lubrication baseline.

This approach prevents two failure modes simultaneously. Under-lubrication causes the expected film breakdown and accelerated wear. Over-lubrication raises internal bearing temperature as excess grease is churned by the rolling elements, which degrades the lubricant and can damage seals. Ultrasound-guided lubrication eliminates guesswork and gives a direct, real-time confirmation that the lubricant has reached the contact zone.

Establishing lubrication routes driven by ultrasound readings rather than fixed calendar intervals is a meaningful upgrade from time-based greasing. Bearings that operate under low load or intermittently will show stable ultrasound signals for extended periods. Bearings under high load, in contaminated environments, or running at elevated temperatures will require attention sooner than a calendar-based schedule would indicate.

Setting Alarm Levels

Alarm levels for bearing monitoring should be based on baseline measurements taken when the bearing is known to be in good condition. Absolute alarm thresholds derived from general tables work as a starting point but tend to produce either missed faults on quiet machines or nuisance alarms on inherently noisy ones.

Vibration-Based Alarms

For defect frequency amplitude at specific bearing fault frequencies in an envelope spectrum, a widely used starting point is:

For overall broadband vibration, ISO 10816 zone boundaries provide a machine-class-adjusted cross-check. Zone A to B transitions may prompt increased monitoring frequency; Zone C triggers a planned replacement; Zone D triggers immediate intervention.

Ultrasound Alarms

Ultrasound alarm levels are set relative to a clean baseline reading at the bearing housing with the asset at normal operating temperature and load:

These starting values are widely cited in ultrasound practitioner guidance and work well for most industrial bearings. Adjust based on accumulated experience with each asset class, particularly for machines with inherently high ultrasound levels from process-related sources such as pneumatic components or cavitating pumps.

Temperature Alarms

Temperature alarms are most effective when set relative to each bearing's operating baseline rather than an absolute temperature threshold. A rise of 10 to 15 degrees Celsius above the established running temperature at consistent load warrants investigation. A sustained rise of 20 to 25 degrees above baseline at normal load is a danger indicator. These are relative changes; an absolute cutoff (such as 85 degrees Celsius for a standard grease-lubricated bearing in free air) provides a safety backstop but should not substitute for trend-based alarming.

Continuous Bearing Fault Frequency Monitoring

Tractian's sensors continuously monitor bearing housing vibration at bearing defect frequencies, trending BPFO, BPFI, BSF, and FTF amplitude automatically to detect fault progression from Stage 1 onward.

See Mechanical Monitoring
mechanicalmonitoring.com is an educational resource for maintenance and reliability engineers. All alarm thresholds, formulas, and stage descriptions are general guidance. Consult bearing manufacturer documentation, ISO standards, and qualified engineering judgment for site-specific applications. Tractian sensor systems are a commercial product of Tractian; references to Tractian technology on this site reflect the sponsoring organization.