Wind Turbine Main Shaft Bearings: Micropitting, Roller Slip and the Limits of Upsizing

The Scale of the Main Bearing Problem

The main bearing supports the rotor and hub and reacts the non-torque loads of the wind before they reach the gearbox, and Hart et al. (2020) noted that its replacement typically requires the complete removal of the rotor. The failure record is substantial. Hart et al. (2019), drawing on operational data from more than 1,000 turbines of each type in the 1.5 to 2.5 MW class, reported 20-year main bearing failure rates of roughly 30 percent for single main bearing drivetrains and roughly 15 percent for double main bearing drivetrains, and observed that most main bearings do not reach their design lives, some failing in under six years. The largest field study to date sharpens the figure. Hart et al. (2023), analyzing 689 main bearing replacements across 7,707 turbines and about 15 GW of capacity, mostly land-based, geared, three-point machines with spherical roller main bearings, found that 10 percent of a fixed population would be expected to have been replaced by 10.5 years, close to half the 20-year design value, and that fitted distributions placed replacements by year 20 at 22 to 25 percent, with no statistically significant relationship between turbine rating and replacement time. The best-fitting log-normal distribution showed the monthly hazard of replacement rising through about year 8.5 and leveling thereafter, and the authors noted that the Weibull distribution, the usual default in reliability work, gave the poorest fit of the four tested. Gbashi et al. (2024), reproducing an earlier breakdown of powertrain failures, reported bearings as 84 percent of the failure rate of wind turbine powertrain components and the main bearing alone as 30 percent, a secondary figure best read as an indication of scale.

The economic consequence follows the mechanical one. Pulikollu et al. (2024), in an EPRI white paper, placed the total cost of a main bearing replacement at $225,000 to $400,000, depending on turbine model and site, and listed lost production from extended delays as a separate consequence; NSK (2025) noted that because main bearings are large and frequently made to order, a failure without a replacement in hand can extend downtime to a year or more. Remaining useful life estimates for these bearings (Rezamand et al., 2020) therefore carry weight in decisions to retire, extend or repower a fleet, a choice the repower guide takes up in detail.

Drivetrain Configuration and Its Effect on the Bearing

Hart et al. (2020) described the three-point and four-point suspensions as the most common configurations for geared turbines, and Gbashi et al. (2024) reported that many main bearings in the three-point arrangement fail earlier than their rated life. Hart et al. (2019) cautioned that contextual factors such as control strategy and operating environment may contribute to the difference between the two configurations, so the twofold gap is best read as strong evidence of a configuration effect and not as a fixed ratio. Kotzalas and Doll (2010) identified the 230/600 series spherical roller bearing as common in the 1.5 to 2 MW class, which places the analysis below squarely within the fleet Malloy Wind supports.

Failure Mechanisms

The Thin Film at Main Shaft Speed

A rolling bearing achieves its rated life when a lubricant film separates the rolling surfaces. Kotzalas and Doll (2010) explained that micropitting requires a film too thin to separate the residual surface texture of roller and raceway, a condition expressed as a lambda value, the ratio of film thickness to composite surface roughness, below 1. Because film thickness depends on entrainment velocity, main shaft bearings, which NSK (2026) describes as running at around 10 to 20 rpm, operate routinely in this low-lambda regime under heavy load, particularly when grease replenishment falls short, and the asperities on roller and raceway make contact through the film.

Load Concentration on the Downwind Row

The same authors, citing Ionescu and Pontius (2009), showed that the combination of radial and thrust loads with the radial clearance inherent in the bearing places essentially the entire load on the downwind row of a typical main shaft spherical roller bearing, leaving the upwind row nearly unloaded (Kotzalas & Doll, 2010). One row of rollers carries a load that the rating calculation distributes across two, which helps explain why the downwind raceway is where damage first appears in the field. The field data agree: Hart et al. (2023) reported that the vast majority of main bearing failures in their data occurred on the downwind row of the locating bearing, and that the inner raceway was the damage location most often listed, followed by the outer raceway, the rollers, the cage and the housing.

Differential Slip Across the Roller

Rollers in a spherical roller bearing slide on the raceway because the roller and raceway curvatures do not match exactly, an effect Kotzalas and Doll (2010) termed Heathcote slip. Only two points across each roller roll without slip; elsewhere the surfaces slide at differing speeds (Malloy Wind, 2023b). Sliding under heavy load in a thin film produces adhesive wear, the micro-welding and tearing that initiates peeling, and the debris it releases drives abrasive wear as it recirculates in the grease, first as two-body and then as three-body wear (Malloy Wind, 2023a, 2023b). At loads near 10 to 20 percent of the static load rating, the wear develops a two-peak pattern on the raceway, with the worst damage between the true rolling points (Malloy Wind, 2023b). Kotzalas and Doll (2010) traced the evolution from low-cycle micropitting at the center of the inner raceway to spalling initiated at the edge of the wear track, where the worn geometry concentrates stress.

Contaminated and Degraded Grease

Peng et al. (2023) documented main bearing grease contamination as a distinct lubrication failure, and Gbashi et al. (2024) listed fatigue spalling, flaking, abrasive wear, fretting, smearing and scuffing among the main bearing failure modes reported in the literature. Hart et al. (2020) noted that abrasive particles in the lubricant produce wear and a gradual increase in bearing clearance, which in turn alters load distribution across the rollers. Contamination compounds the thin-film problem: dents from hard debris raise local stress and roughen the surfaces that the film must already struggle to separate.

What the Damage Reports Show

Field damage reports add an important qualification to the mechanism story. Across the damage reports collected by Hart et al. (2023), spalling was the damage type most often listed, alongside indents, microspalling, pitting, chipping, wear, fracture, debris in the grease, axial cracking and heat damage. Spalling is the end state of both surface-initiated and subsurface-initiated rolling contact fatigue, and the authors read the accompanying surface damage as evidence that at least part of the spalling began at the surface, a pattern associated with surface distress and possibly insufficient lubrication and made more likely when slip combines with abrasive particles; wear, fracture and heat damage pointed to further mechanisms such as abrasion, smearing and impact or excessive load. They were careful, however, to state that the data cannot yet separate wear-induced, surface-initiated and subsurface-initiated fatigue, and that subsurface fatigue should not be discounted as a contributor to early main bearing damage.

What the Lubrication Research Shows

The thin-film problem has now been modeled directly for the bearing type that dominates the three-point fleet. Hart et al. (2022) estimated film thickness across a dataset of contact conditions for the double-row spherical roller main bearing of a 1.5 MW turbine, using main bearing loads derived from aeroelastic simulation and the properties of a commercial grease marketed for main bearings, with a base oil of 460 mm²/s at 40 °C. Under the best case of fully flooded lubrication, a contact inlet temperature of 35 °C marked the transition from elastohydrodynamic to mixed lubrication, and by 40 °C close to 80 percent of operating points fell in the mixed regime. Because a grease-lubricated main bearing is expected to run starved for most of its life, the authors also applied an order-of-magnitude starvation estimate, a 30 percent reduction in film height, under which close to 90 percent of operating points at 35 °C and all points at 40 °C fell into mixed lubrication; they cautioned that real starvation could be more severe. Their sensitivity analysis found temperature, starvation level and the pressure-viscosity coefficient of the base oil to dominate the result, while roller load had a small effect, and maximum contact pressures stayed near 1.6 GPa.

Wu et al. (2023) reached a parallel conclusion by a probabilistic route for the spherical roller main bearing of a 5 MW reference turbine, finding that lubrication failure probability rose steeply as mean wind speed fell, from nearly reliable near 13 m/s to nearly certain failure near 7 m/s. Read together, the two studies suggest that the main bearing spends a meaningful share of ordinary operation, and especially its low-wind hours, in mixed lubrication, and that grease temperature and replenishment, more than the size of the bearing, govern how large that share is. Both are simulation studies on specific bearings and greases, so the figures describe risk for those cases more than a measured fleet condition, a limitation the authors themselves emphasize.

Lubrication practice carries the same weight in the review literature. Peng et al. (2022), surveying tribological failure and lubrication research across wind turbine bearings, reported estimates that about 70 percent of bearing failures are linked to lubrication, and noted that the timed, fixed-quantity relubrication most turbines use risks both under- and over-lubrication, with wear, chemical reaction and grease spillage as the consequences. The same review summarized grease work aimed specifically at the main shaft, including SKF's LGWM series and a synthetic hydrocarbon, complex lithium grease developed for main shaft bearings, and cited a condition-based criterion of replacing lubricant whose dropping point has fallen by more than 30 °C or whose iron content exceeds 5,000 ppm.

The Limits of Upsizing

A common field response to main bearing failures has been to replace the 230/600 series bearing with a 240/600 series bearing, whose longer rollers spread the load along the raceway and raise the calculated fatigue life (Kotzalas & Doll, 2010). Two comparisons appear in the literature and should be kept apart: Kotzalas and Doll (2010) discussed the 230/600 against the 240/600 on the same 600 mm bore, while Kenworthy et al. (2024) compared a 230/600 with a 240/630, which changes the bore and capacity as well as the series, so their results cannot isolate the effect of the series alone. The difficulty is that the life-limiting mechanism appears to be only partly the subsurface fatigue the rating measures. Kotzalas and Doll (2010) argued that the 240 series is probably no less prone to micropitting than the 230 series, since the longer roller increases the sliding across its length, and they characterized the upsizing response as one adopted where tribological expertise was not engaged in turbine design and maintenance. Kenworthy et al. (2024) reached a complementary conclusion from the other direction. Rating two spherical roller main bearings for a three-point 1.5 MW turbine under IEC 61400-1 and ISO 281, a 240/630 with a fatigue load limit of 1,141 kN and a 230/600 with a limit of 750 kN, they found basic rating lives of 126 to 174 years for the larger bearing and 34 to 44 years for the smaller, and modified rating lives that stayed well above the 20-year design life under expected conditions of slight to typical contamination and mean temperatures below 40 °C. The 240/630 operated below its fatigue load limit for the vast majority of its life while the 230/600 operated above its limit most of the time, and temperature and contamination moved the results far more than wind shear or turbulence did; only a combination the authors judged unrealistic, severe contamination with halved lubricant viscosity on the smaller bearing, approached the 22 to 30 percent year-20 failure rates reported in the field. Their conclusion was that ISO 281 rating life assessment, as used in IEC 61400-1, does not account for the reported failure rates in 1 to 3 MW turbines, and they recommended a parallel analysis under ISO/TS 16281, since replaced by ISO 16281:2025, further work on root causes and possibly an application standard specific to main bearings. The bearing science guide sets this result beside the rating gaps found at the other drivetrain positions.

A larger fatigue rating therefore addresses part of the failure record at most. The field damage record (Hart et al., 2023) keeps subsurface fatigue in view, which is a reason to treat a higher rating as one element of a replacement specification instead of the whole of it. The lubrication modeling above sharpens the point, since Hart et al. (2022) found a 1.5 MW main bearing likely to run in mixed lubrication at contact inlet temperatures near 35 to 40 °C, a range Kenworthy et al. (2024) treated as comfortably safe for fatigue, which suggests that the bearing can satisfy its rating while its surfaces wear. The design response that the evidence supports targets the surfaces themselves. Kotzalas and Doll (2010) reported that a durable WC/aC:H coating combined with a superfinished surface on the rollers can eliminate or significantly delay the onset of micropitting, and Gbashi et al. (2024), coding the main bearing developments described in 12 publications from five bearing manufacturers between 2014 and 2023, found that 35 percent of those developments centered on coatings, 25 percent on cage improvements, 25 percent on configuration and 15 percent on novel bearing designs, a count of published developments and not of bearings sold or installed, while calling for optimization studies that account for micropitting, smearing and scuffing in addition to fatigue.

The Upgrade Malloy Wind Stocks: NSK Super-TF Main Bearings

NSK's Super-TF spherical roller main bearings combine the material, coating and cage responses the literature describes. NSK (n.d.-a, n.d.-b) specifies carbonitrided Super-TF ring steel containing fine carbides and a controlled level of retained austenite, which relieves stress concentration at the shoulders of dents formed by contaminant particles; rollers treated with a diamond-like carbon (DLC) coating to raise surface toughness and wear resistance; a controlled raceway surface finish and roller-to-raceway profiles that govern roller motion; custom radial internal clearance; and dimension series 230 and 241 for shaft diameters of 500 mm and larger. The extra-capacity ECA variant adds a machined brass, roller-guided cage that removes the center guide ring and the sliding friction it introduces, developed in 2024 for North American operators (NSK, 2025).

NSK's published figures, which reflect the manufacturer's own test conditions, include a basic dynamic load rating 23 percent above standard, an L10 life more than ten times that of ordinary carburized steel under contaminated lubrication, white structure flaking life twice that of carburized steel and four times that of standard bearing steel, and raceway wear reduced by 90 percent or more when Super-TF steel, DLC rollers and the ECA cage are combined (NSK, n.d.-b, 2025). These are manufacturer claims and are best weighed against the field record for a given platform. NSK's March 2026 technical review of the same bearing adds engineering detail on each element, again under the manufacturer's own tests: the roller-guided ECA cage gives about 1.1 times the load capacity and 1.3 times the bearing life of the conventional CA cage it replaces, a comparison on a different basis from the 23 percent load rating figure above; Super-TF steel showed better raceway wear resistance than standard bearing steel in a Sawin-type wear test; and the DLC coating, several micrometers thick, uses a layered structure intended to combine adhesion to the roller with coating hardness, with surface wear expected to fall below one-tenth of a standard uncoated bearing (NSK, 2026). The review does not give load, speed or lubrication conditions for these tests in its text, and it does not discuss electrical current or delamination.

Malloy Wind's own field record adds a necessary qualification to the coating story. DLC coated rollers became a common main bearing upgrade and performed with few roller failures on 1 MW and smaller units, but on 2 MW platforms Malloy Wind has documented the coating delaminating from the rollers, which then spall and in some cases split during operation (Malloy Wind, 2023c). The cause of the delamination remains debated, and Malloy Wind's response at the time was to favor a steel whose wear, debris and white etching crack resistance is built into the material and cannot delaminate, together with improved lubrication through manual greasing, grease flushes or an automatic lubricator (Malloy Wind, 2023c). NSK's 2025 configuration, which pairs Super-TF ring steel with DLC rollers and the roller-guided ECA cage, postdates that analysis, and its field record on 2 MW platforms is still accumulating, so the two positions are best read together: the steel carries the wear resistance that cannot delaminate, and the coating is one layer of protection on top of it rather than the sole defense. NSK's 2026 work on coating adhesion is the manufacturer's engineering response to that concern, and platform-specific field validation is what will settle it. Malloy Wind also stocks select main bearing seals and will consider adding inventory for a specific fleet on request.

A 2026 conversation between Cory Mittleider, then of Malloy Wind, and Loren Walton of NSK moved the question further. Mittleider reported that operators had begun asking for main bearings supplied without DLC on the rollers, a reversal after more than a decade in which the coating was the standard response to peeling, and he described its original purpose: a layer 1 to 2 micrometers thick of a very hard material dissimilar to the steel, which resists the welding and tearing of asperities that begins peeling once the film breaks into mixed and boundary lubrication (Hall, 2026). Walton credited the coating with raising life on the affected platforms from 1 to 5 years to 10 years or more without any change to the bearing construction, and he was careful to say that DLC remains sound in lighter duty while becoming an accelerant to failure once other factors are added. The factor discussed most was electrical, and it is best held as an industry hypothesis. The coating is nonconductive, and the host, Allen Hall, cited measurements attributed to Howard Penrose of well over 100 amperes circulating through nacelle ground paths, along with a shift in converter switching toward frequencies of 10 kHz and above that are harder to ground; on that account, current breaking down the coating would leave hard fragments in the bearing that wear it like sand. Walton added that an owner operator NSK worked with had found the current present for years and across models, which suggests that current alone does not explain the recent failures (Hall, 2026). These are interview statements. Current in a nacelle grounding path is not the same as current through an individual roller contact, and no published measurement yet links the two or ties either to coating delamination, so the relationship needs further investigation; the bearing current evidence in the generator bearing guide is the better reference for how such currents behave. Both guests located the remedy in the steel itself, its alloy chemistry and heat treatment, so that removing the coating does not return the bearing to the peeling it was meant to prevent, while Mittleider urged operators who can identify and remove the stray current to do so (Hall, 2026). Read beside NSK's 2025 configuration, which keeps DLC on the rollers, the position amounts to a judgment about where the coating belongs, and that judgment is made platform by platform.

Supported Turbine Platforms

Malloy Wind supports main bearing applications on GE 1.5, 1.6, 1.7, 1.85, 2.x, 2.3, 2.5, 2.7 and 2.8; Siemens 2.3; Suzlon S64 and S88; Vestas V80, V82, V90, V100, V110 and V112; Gamesa G80 and G90; NEG Micon 750 and 900 kW; and MHI 1000A and 2.4 MW turbines (Malloy Wind, 2026). The list reflects applications Malloy Wind currently stocks or supports, and matching is done from the existing bearing designation and the platform's failure history, so operators with a platform not listed can send the turbine model and that designation for review.

Inspection and Maintenance Practice

Purged grease offers the most accessible evidence of main bearing condition, read against a fresh-grease baseline and earlier samples, since dark color alone is not necessarily a sign of abnormal wear; metallic particles, water or a change from the previous result are the more useful signals. Temperature is a comparatively late indicator and means most when corrected for operating conditions: Hart et al. (2020) described normal-behavior models that predict main bearing temperature from rotational speed, active power and ambient temperature, and reported that industry practice combining vibration and temperature has detected main bearing faults fairly consistently five to six months before failure. On three-point drivetrains the downwind row deserves first attention during borescope inspection, since the load and the damage concentrate there (Kotzalas & Doll, 2010). Relubrication should follow the OEM interval and grease specification, with care to avoid both over-purging and contamination during service, and purged grease sampling for iron content and dropping point offers a condition-based check on whether that interval suits the site (Peng et al., 2022). Because film thickness falls quickly as grease temperature rises (Hart et al., 2022), main bearing temperature trends are worth reading as a lubrication indicator as well as a sign of damage. Hart et al. (2023) identified site conditions, including wind class, turbulence intensity and annual temperature range, and maintenance records of relubrication and grease flushing as the variables most worth adding to future failure analysis, which suggests that owners who log both are better placed to explain their own fleet's replacement pattern. Walton and Mittleider added several checks for newer and larger turbines (Hall, 2026). Original fits and clearances deserve verification, since NSK has seen four-point drivetrains in which the floating bearing was installed so that it could not float, leaving both bearings fixed. Automatic lubricators, now common on newer main bearings, are machines of their own that can fail, stop or overdeliver, and they need the same inspection as the bearing. A grease sampling baseline taken at about six months and again at one year makes later results interpretable; a laboratory report typically lists 10 to 15 elements in parts per million, with iron, copper alloys from the cage, silicon and phosphorus among the readings that matter, and water measured by a separate test, and NSK shares guideline levels at which a result warrants watching or a purge to see whether it stabilizes. Given rotor-off replacements and long lead times, the practical difference between a planned replacement and an extended outage is often whether the bearing is already on the shelf.

In Finality

The literature on wind turbine main bearings converges on a reading that the conventional replacement decision has been slow to absorb. The bearings are reaching spalling far earlier than their ratings predict, and the most consistent explanation is surface distress under thin films, concentrated load on a single row and differential slip across each roller, with subsurface fatigue still a possible contributor (Hart et al., 2023; Kenworthy et al., 2024; Kotzalas & Doll, 2010). In the largest field data set available, one main bearing in ten had been replaced by 10.5 years (Hart et al., 2023). Single main bearing drivetrains carry roughly twice the failure rate of double main bearing designs in the available field data, a gap to which control strategy and site conditions may also contribute (Hart et al., 2019), and that makes the three-point fleet, including the GE 1.5 family, the population where this distinction matters most. Lubrication modeling of spherical roller main bearings indicates that mixed lubrication is likely for much of ordinary operation once grease temperature, starvation and low wind speeds are considered (Hart et al., 2022; Wu et al., 2023). A replacement strategy that answers the documented mechanism, through wear-resistant ring steel, sound lubrication practice and a cage that reduces sliding, appears more likely to change the outcome than one that only raises the rating.

Tech Notes on Main Shaft Bearings

• What is Heathcote Slip?
• Main Bearings: What is peeling damage?
• Are DLC coatings no longer cutting it?
• What are Super Tough bearings?
• Through Hardened vs Case Hardened Bearings
• Forthcoming: Micropitting in main bearings; The 230 and 240 series compared; Main bearing grease and relubrication practice; Super-TF, DLC and the ECA cage; Planning a main bearing replacement

Frequently Asked Questions

Requesting a Main Bearing

Call 605.336.3693, email wind@malloyelectric.com or request a quote with the turbine model, the bearing designation and a description of the condition observed. Photographs of the raceway and the purged grease allow the team to match the documented mechanism to the right replacement.

References