Wind turbine gearboxes are designed to last 20 years. However, wind operators often experience gearbox failure rates of up to 50% across the lifetime of their fleets. If component fatigue, fretting or wear is caught early enough, an operator can mitigate consequential damage and prevent catastrophic gearbox failure. If the window of prevention is missed, however, the operator can incur costly repairs between $200,000-$500,000 for just the gearbox system. This includes a minimum of two weeks of lost revenue from asset downtime, the cost of the gearbox replacement, and additional crane and labor costs.[adright zone=’190′]
As the wind industry grows and prospers, operators are investing in software solutions, in addition to their condition-based monitoring systems, to predict component failures early and prevent major gearbox failures. These investments help them move from corrective maintenance practices to predictive health maintenance.
Bearings cause approximately 64% of all gearbox failures, according to the National Renewable Energy Laboratory’s gearbox failure database. Of the 16-22 bearings, depending on gearbox type and configuration, operated in a standard 1.5 MW-rated machine, the highest number of failures are observed in the high-speed shaft (HSS) and high-speed intermediate shaft (HSIS) bearings due to axial cracking and spalling.
If an operator has the data to know which bearing assembly is running with damage, life extension of the gearbox can be achieved through early component replacement, a physics-based derate to reduce the load on the gearbox, a lubrication exchange, or the use of lubricant additives.
For example, the replacement cost of an HSS bearing is approximately $25,000. Action taken to save or replace the HSS assembly can result in a $60,000 savings by preventing secondary damage in the HSIS pinion assembly. If there is damage in the HSIS bearing assembly, secondary component risk is placed on the low-speed shaft (LSS), HSIS gear and the planetary system. Damage to the LSS and planetary systems can cost over $100,000 each. At this stage, the operator needs to evaluate if an up-tower component replacement could save the gearbox or if a physics-based derating of the turbine could slow down the damage, thereby optimizing the timing of the gearbox replacement.
White etch cracking (WEC) is another cause of premature failures in rolling element bearings, occurring as early as 1% to 50% of the calculated L10 life. Unfortunately, WEC has been found in bearing applications, with operating conditions having no common denominator. The tribological drivers for WEC formation are also contested.
Microstructural alterations in the vicinity of inclusions are often observed, which serve as crack nuclei that initiate microstructural alterations commonly referred to as butterfly wings. WEC formation is not always accompanied by butterfly wing formations, though. In these instances, they are called irregular WECs, which makes finding the root cause on them challenging.[adleft zone=’190′]
In one study with a major wind operator, low-speed intermediate shaft (LSIS) bearings and HSIS bearings in the same gearbox configuration were analyzed. According to the results, the LSIS bearing, which rotates at a speed of ~100 rpm, experienced wear and debris dents. The HSIS bearings, which rotate at ~450 rpm, showed axial cracking, wear and debris dents in the surface topography.
HSIS bearings experience large torque reversals at the time of generator-grid engagement and high-torque peaks during sudden breaks, which induce excessive stresses beyond design loads. It’s possible that excessive loading from transient events, combined with tensile stresses, initiates cracks. The cracks, once initiated, can lead to the formation of WEC.
In another study, rolling contact fatigue tests with artificially induced cracks in the materials showed WEC was the consequence and not the cause of the damage. It is also believed that dynamic loads experienced by high-speed stages can cause localized plasticity and form WECs in the HSIS bearing, the paper says.
It is also well known that nascent hydrogen is detrimental to WEC initiation and progression, which can be generated from lubricant decomposition, stray currents and water ingress. Therefore, it’s possible that transient loading and other tribo-mechanical aspects, such as high sliding, can result in increased friction energy, driving WEC formation.
WEC on the surface and subsurface is driven by different mechanisms comprising material, mechanical, thermal and chemical phenomena, so a clear understanding of these damage mechanisms would assist in developing life prediction models that could be used to quantify the bearings at the design stage and reduce costs through predictive health maintenance.
Bearing suppliers have developed coating solutions designed to increase the resistance of WEC formation. Engineers use materials that reduce the stress placed on the bearing, including certain coatings claimed to minimize the risk of damage caused by slippage. The coatings are designed to prevent corrosion and hydrogen diffusion and improve protection against WEC formation.
In a different customer case, four bearing supplier offerings were evaluated to measure WEC resistance in the HSS bearing assembly. Multiple trade-off studies and simulation cycles were conducted to compare the life impact of each of the offerings. Simulations were completed under representative field data to predict bearing life. The study concluded that the bearing using a diamond-like coating (DLC) layer could reduce the impact of hydrogen embrittlement caused by overload and/or water contamination in the oil, thus reducing the risk of WEC. Non-DLC-coated bearings showed a higher risk for WEC during standard operating conditions.
To prevent repeat failures in a fleet, it is important to understand how different bearing offerings impact the life of a gearbox. Understanding gearbox configuration and the failure rates of the components within could make a huge difference on an operator’s revenue. An operator certainly wouldn’t want to replace a poor-performing bearing with the same bearing. Instead, it would want to evaluate different options to ensure the replacement will result in life extension of the gearbox.
Harpal Singh is materials scientist and lab manager at Buffalo, N.Y.-based Sentient Science. He can be reached at firstname.lastname@example.org.