Bearing condition monitoring is the practice of tracking vibration – usually measured at the bearing housing – and temperature trends to identify signs of rolling-element bearing degradation, such as raceway spalling or lubrication-related distress, before functional failure. In induction motors, one of the best-documented applications, bearing faults are consistently reported as the largest single failure category.
When budgets are limited, prioritize bearing monitoring on assets whose failure would stop production, using an asset-criticality assessment rather than component type alone.
Why Bearing Monitoring Is a Core Predictive Maintenance Program
Bearing faults account for roughly 41% to 44% of induction motor failures, more than any other single component, including stator windings and rotor faults, according to a peer-reviewed summary combining IEEE, EPRI, and academic motor reliability surveys. These surveys support the importance of bearing monitoring in induction motors; they should not be generalized to all rotating equipment.
The financial exposure is not abstract. Siemens’ 2024 Senseye survey estimates – by extrapolating its research to Fortune Global 500 industrial organizations – that unplanned downtime costs nearly $1.4 trillion annually, equivalent to 11% of revenue. The report also gives an average of 27 lost hours per plant per month.
Separately, Siemens reports that live Senseye deployments have shown a 50% reduction in unplanned downtime, a 40% reduction in maintenance costs, and payback within three months. These are vendor-reported deployment outcomes, not universal benchmarks. The P-F interval for a developing bearing fault may range from weeks to months, but it depends on sensor placement, sampling strategy, load, speed, operating conditions, and fault progression.
The Six Ways a Bearing Actually Fails
Not every bearing failure looks the same, and treating them as one problem wastes diagnostic effort. ISO 15243:2017, the international standard bearing manufacturers use to classify damage, sorts failures into six categories based on visible appearance: rolling contact fatigue, wear, corrosion, electrical erosion, plastic deformation, and fracture. fracture and cracking.
ISO 15243 distinguishes moisture corrosion from frictional corrosion; frictional corrosion includes fretting corrosion and false brinelling. These mechanisms differ in appearance and corrective action.
One detail from bearing manufacturer SKF is worth sitting with: roughly 90% of bearings outlive the machine they were installed in. This statistic does not mean that all premature failures share one cause. Rather, it reinforces the need to investigate lubrication, contamination, mounting, alignment, loading, and operating conditions. Vibration monitoring detects symptoms and probable fault locations; root cause may require complementary evidence such as lubricant analysis, operating data, or post-removal inspection.
From Vibration to a Bearing Health Score: How Condition Monitoring Works
Envelope analysis demodulates the high-frequency resonance excited by repetitive impacts and reveals characteristic defect frequencies in the envelope spectrum. These include Ball Pass Frequency Inner Race (BPFI), Ball Pass Frequency Outer Race (BPFO), Ball Spin Frequency (BSF), and Fundamental Train Frequency (FTF); their expected values depend on bearing geometry and rotational speed.
Early bearing impacts often excite a machine-specific structural resonance in a high-frequency carrier band. The appropriate demodulation band must be selected from the sensor bandwidth and measured spectrum; 2-6 kHz is not a universal range. After demodulation, BPFI, BPFO, BSF, and FTF components are evaluated in the lower-frequency envelope spectrum, ideally with speed or order information for variable-speed assets. The P-F interval may range from weeks to months in suitable applications, but it is not a guaranteed lead time.
Edge processing can compute RMS, peak, kurtosis, crest factor, filtered spectra, envelope features, and anomaly scores at the sensor, reducing bandwidth and alert latency. Fault labels and severity trends should be treated as diagnostic outputs whose accuracy depends on model validation, coverage of operating conditions, and access to supporting waveform evidence.
Standards Relevant to Bearing Monitoring Programs
Several ISO standards contribute to a formal bearing monitoring program, and they answer different questions.
| Standard | Scope | Relevance |
| ISO 15243:2017 | Classifies bearing damage and failure modes by visible characteristics: rolling contact fatigue, wear, corrosion, electrical erosion, plastic deformation, fracture (ISO) | Gives a common vocabulary for what a failed bearing actually shows, used by major bearing manufacturers for failure analysis |
| ISO 20816-1:2016 | General conditions and procedures for measuring and evaluating vibration on rotating and non-rotating parts of complete machines (ISO) | Provides general vibration-evaluation principles. Machine-specific numerical criteria come from the applicable parts of ISO 20816, and the series does not itself diagnose rolling-element bearing faults. |
| ISO 17359:2018 | General guidance for establishing condition-monitoring programs | Program design and implementation framework |
| ISO 13373-2:2016 | Vibration-data processing, analysis, and presentation | Time- and frequency-domain diagnostic methods |
ISO 15243 supports post-damage classification; ISO 20816 supports whole-machine vibration evaluation; ISO 17359 addresses program design; and ISO 13373-2 covers signal processing and vibration diagnostics. None specifies a particular sensor or validates an AI classifier. A monitoring program should therefore combine applicable standards with asset-specific baselines, alarm validation, model-performance evidence, and engineering review.
Architecture: Wired or Wireless, One Analytics Layer
The sensor choice comes down to one question: is the asset already wired for monitoring, or does adding a sensor mean running new cable? For motors, pumps, and gearboxes on an existing production line with structured cabling, the wired DS-Track product page states less than 1 ms of on-device anomaly-detection latency and lists wired industrial communications.
For difficult-to-reach or previously unwired assets, DS-Track AIR can reduce cabling work.
Either way, the data lands in the same place. DS-Insight aggregates readings across the fleet into a single dashboard, applies smart alarm filtering so a one-off spike does not trigger a false callout, and builds a per-asset health index a reliability manager can sort by risk.
What Field Diagnostic Data Can Show
In one coupling misalignment case on an industrial water pump, FFT combined with phase-difference analysis picked up rising RMS vibration consistent with a developing misalignment fault.
A separate bearing case, drawn from Delphisonic’s field diagnostic reports on rail equipment rather than industrial assets directly, used envelope spectrum analysis to pick up BPFI harmonics staged across four progressive severity levels. The same bearing-kinematics equations apply, but signal expression and detectability vary with speed, load, mounting, structural response, and background noise.
Building a Bearing Monitoring Program
Start with the assets where a bearing failure stops production entirely, not the ones that are simply expensive to replace. A conveyor gearbox bearing that halts a line is a different priority than a redundant fan motor, even if the fan motor costs more.
Layer in envelope analysis against BPFI, BPFO, BSF, and FTF from day one rather than starting with broadband severity alone. Broadband trends can flag an overall change, while envelope analysis can help localize bearing-related components. Neither method alone determines remaining useful life without a validated prognostic model.
If you’re building out a bearing condition monitoring program for industrial assets, Delphisonic has been engineering this kind of sensor and analytics stack since 2012, on the principle that machines should speak before they break. Request a demo to walk through what a rollout would look like on your plant floor.
Frequently Asked Questions (FAQ)
Bearing condition monitoring: causes, standards, and deployment questions
Bearing condition monitoring is the practice of tracking vibration and temperature signals from a machine’s bearings to detect wear signatures, such as race spalling or lubrication breakdown, well before the bearing reaches functional failure, instead of waiting for a scheduled inspection or an unplanned stop.
Bearing faults account for roughly 41% to 44% of induction motor failures according to combined IEEE and EPRI motor reliability surveys, more than any other single component, including stator windings and rotor faults.
Envelope spectrum analysis on BPFI, BPFO, BSF, and FTF frequencies can reveal a developing bearing fault well before it becomes audible or causes a functional failure. The frequency band used depends on the sensor and the machine’s structural resonance rather than one fixed range, and the resulting lead time varies with sensor placement, operating conditions, and fault progression, commonly reported as weeks to months rather than a guaranteed interval.
ISO 15243:2017 classifies bearing damage by its visible characteristics after the fact, ISO 20816-1:2016 sets general vibration-evaluation principles, ISO 17359:2018 covers condition-monitoring program design, and ISO 13373-2:2016 addresses vibration signal processing and diagnostics. None of these standards specifies a particular sensor or validates an AI-based fault classifier.
A wired sensor makes sense when the asset already has structured cabling in place, since it offers continuous, low-latency monitoring. A wireless sensor is generally the better fit for retrofitting older equipment or hard-to-reach assets, since it avoids the cost of running new cable through existing machine guarding and conduit.
References
- Yousuf, M. et al., “IoT-based health monitoring and fault detection of industrial AC induction motor for efficient predictive maintenance,” SAGE Journals / Measurement and Control (2024), citing IEEE and EPRI motor reliability surveys – https://journals.sagepub.com/doi/10.1177/00202940241231473
- Siemens / Senseye Predictive Maintenance, “The True Cost of Downtime 2024” – https://assets.new.siemens.com/siemens/assets/api/uuid:1b43afb5-2d07-47f7-9eb7-893fe7d0bc59/TCOD-2024_original.pdf
- International Organization for Standardization, “ISO 15243:2017 – Rolling bearings — Damage and failures — Terms, characteristics and causes” – https://www.iso.org/standard/59619.html
- International Organization for Standardization, “ISO 20816-1:2016 – Mechanical vibration — Measurement and evaluation of machine vibration — Part 1: General guidelines” – https://www.iso.org/standard/63180.html
- SKF Evolution, “Understanding the ISO 15243 – Bearing damage modes and classifications” – https://evolution.skf.com/understanding-the-iso-15243-bearing-damage-modes-and-classifications/
- International Organization for Standardization, “ISO 17359:2018 – Condition monitoring and diagnostics of machines – General guidelines” – https://www.iso.org/standard/71194.html
- International Organization for Standardization, “ISO 13373-2:2016 – Condition monitoring and diagnostics of machines – Vibration condition monitoring – Part 2” – https://www.iso.org/standard/68128.html
- Delphisonic, “DS-Track product specifications” – https://delphisonic.com/product-ds-track/
