Bearing Failure Modes: A Visual Field Guide

Bearing failure modes are the distinct ways a rolling-element bearing degrades in service. The international standard ISO 15243 groups them into six classes: rolling contact fatigue, wear, corrosion, electrical erosion, plastic deformation, and fracture and cracking. Each leaves a recognizable mark that points back to a specific root cause.

That last point is why a failed bearing is worth ten minutes on the bench instead of a trip to the scrap bin. The damage pattern is a signed confession: fluting on the raceway means stray electrical current, evenly spaced dents mean a shock load during handling, a dull frosted band means the lubricant film broke down. This guide walks the six ISO 15243 modes, what each looks like, the cause it implicates, and how to turn a post-mortem into a fix. It pairs naturally with the bearing defect frequencies guide, vibration tells you a bearing is failing and which surface; the failure mode tells you why.

What are the six ISO 15243 bearing failure modes?

ISO 15243:2017 classifies damage that occurs while a bearing is installed and running (it excludes manufacturing defects). Six primary modes, each with subcategories, are shown below.

• Rolling contact fatigue. The mode bearings are engineered to eventually reach. Cyclic stress under the rolling elements initiates cracks that grow until material flakes away (spalling). Subsurface-initiated starts below the surface at inclusions; surface-initiated starts at the surface where the lubricant film is inadequate. Appearance: flaking, spalling, pitting in the load zone.
• Wear. Progressive material removal. Abrasive wear comes from hard particles (dirt, debris) lapping the surfaces to a dull, frosted finish; adhesive wear (smearing) comes from metal-to-metal transfer when the lubricant film collapses under sliding.
• Corrosion. Moisture corrosion (rust) from water or condensation in the lubricant; frictional corrosion which includes fretting corrosion at the fit surfaces and false brinelling on the raceway from micro-movement while the machine sits idle and vibrating.
• Electrical erosion. Stray current passing through the bearing. Excessive-voltage erosion melts tiny craters; current-leakage erosion produces the classic washboard fluting pattern across the raceway. Common on VFD-driven motors without proper shaft grounding.
• Plastic deformation. Permanent indentation without material loss. Overload deformation and true brinelling come from static shock loads (a hammer blow during mounting, a dropped assembly); debris indentation comes from particles rolled into the raceway, each leaving a dent with a raised edge.
• Fracture and cracking. Forced fracture from overload or rough mounting; fatigue fracture from cyclic stress beyond endurance; thermal cracking from frictional heat when sliding overwhelms rolling. Rings crack, flanges break away.

How does one bad bearing damage cause another?

Most failed bearings carry more than one mode by the time they reach the bench, and reading them in the wrong order sends you chasing the symptom instead of the disease. Damage cascades: a debris dent (plastic deformation) becomes a stress raiser that seeds a spall (fatigue); a lost lubricant film (adhesive wear) drives up heat until the ring discolors and eventually cracks (thermal fracture). The last, loudest damage is rarely the first.

The discipline is to find the primary mode, the one that started the chain, because that is the only one connected to a preventable cause. Two clues help: primary damage usually sits in the load zone and is more advanced, while secondary damage is scattered and shallower; and the mode with a clean root-cause story (a grounding fault, a mounting shock, a contamination source) is almost always the initiator. Fix the primary cause and the whole cascade stops. Treat the secondary damage as the real problem and you will replace the bearing into the same failing conditions.

How do you read the damage to find the cause?

The location and texture of the mark narrow the cause fast. A few of the most diagnostic patterns are worth committing to memory, because they are frequently confused with ordinary fatigue and lead to the wrong fix.

The general reading rules:

• Location tells you load. Damage confined to a narrow band in the outer-race load zone is normal loading; damage all the way around a rotating ring, or displaced from the load zone, signals misalignment, a rotating load, or a fit problem.
• Spacing tells you a lot. Marks spaced exactly at the rolling-element pitch (brinelling, false brinelling, debris dents) implicate an event or condition tied to element position, not continuous running.
• Color and texture tell you heat and lubrication. Blue/brown discoloration means overheating; a dull frosted band means the oil film thinned and metal touched metal; polished mirror bands can mean smearing or fretting.
• Fluting is its own tell. A regular washboard across the raceway is electrical, full stop, chase the grounding and the VFD, not the bearing supplier.

Which failure modes are most common, and which are preventable?

Here is the counterintuitive part. Rolling contact fatigue is the only mode a bearing is actually designed to eventually experience, the L10 rating life is a fatigue number. Yet most bearings that fail early do not fail from fatigue. Bearing-maker analysis of premature failures attributes roughly a third to lubrication problems, about a third to fatigue (much of it surface-initiated by poor lubrication), and most of the remainder to mounting/handling and contamination. In other words, the majority of early bearing deaths are decided at the loading dock, the lube route, and the installation bench, long before the raceway ever gets tired.

A 5-step bearing failure post-mortem

1. Bag the evidence before you clean it. Note orientation, mark the load zone, photograph the raceways and rolling elements as removed. Wiping the bearing down destroys the frosting, discoloration, and debris that carry the diagnosis.
2. Classify the primary mode. Match the dominant damage to one of the six ISO 15243 classes. If several are present, find the one that came first, secondary damage always follows the primary failure.
3. Trace the mode to a root cause. Fluting to grounding, brinelling to a mounting shock, frosted wear to contamination. Use 5 Whys or a fishbone diagram to get past the mode to the system that produced it.
4. Fix the process, not just the part. Replacing the bearing without changing the cause buys you the same failure on the same clock. If contamination is the mode, the fix is sealing and lube cleanliness; if brinelling, the fix is installation method.
5. Feed it back into reliability data. Log the mode and cause against the asset so patterns emerge across bearings and machines. Repeated fluting across a plant's VFD motors is a design finding, not six unlucky bearings; that pattern only shows up if failures are recorded consistently and land on your maintenance KPIs.

What the standards and numbers say

• The six-mode classification, subcategories, appearances, and causes are defined in ISO 15243:2017, Rolling bearings, Damage and failures, Terms, characteristics and causes (ISO 15243:2017). It classifies damage occurring in service, not manufacturing defects.
• The premature-failure cause split most often cited, roughly lubrication 36%, fatigue 34%, mounting/handling 16%, contamination 14% comes from bearing-manufacturer failure-analysis training and should be read as a directional estimate, not a controlled study (Reliability Solutions, contamination and bearing life; SKF bearing-damage analysis). Independent studies put solid-particle contamination alone behind roughly 14–25% of premature failures. The through-line: most early failures are preventable.

Reading failure modes only pays off if the readings accumulate somewhere the whole team can see. Harmony pulls maintenance history, downtime reasons, and machine signals into one operational data layer, so a recurring failure mode surfaces as a pattern across assets instead of dying in a technician's notebook, and it can draft the corrective work order for a human to approve. It layers onto the CMMS and machines you already run, with no rip-and-replace; see how it works or the CLS case study. For where this fits in a maintenance strategy, see equipment reliability predictive maintenance and the role of lubrication management in preventing the two biggest failure modes on the list.