Wind turbine drive train failures can be expensive to remediate due to crane fees, replacement component cost, man power costs and lost production. In order for wind turbine owners and operators to better manage the drive train lifecycle a sound understanding of failure modes and corrective actions is required. This article focuses on the techniques to establish the failure mechanism of wind turbine gearbox bearings and the process of root cause analysis (RCA) to determine the origins.
The recently revised wind turbine gearbox design standard ISO 61400-4 requires the gearbox bearings to be sized per ISO 281 and ISO 76, which include calculations for two failure modes: a) sub-surface initiated rolling contact fatigue and b) yielding under maximum stress. The standard also mandates items such as bearing steel quality (to meet ISO 683-17), values for parameters involved in the sizing calculations and how design loads should be applied. The standard also provides good advice to the application from a design perspective, as often the failures seen in the field are not related to the bearing meeting the design calculation for these two failure modes. We see failures related to end loading, bearings races spinning in their fit, surface originated fatigue and retainers coming loose for example. Figure 1 illustrates some of the factors involved in bearing performance and each or several can influence a premature bearing failure, where the life is much less than that found by the design life calculation.
ISO 15243 is a useful standard to classify the damage and Figure 2 provide a few examples of bearing failures in the wind industry and the associated costs. Note that damage classification is not root cause of failure! One or more initiating factors are often behind the observed damage.
The first step for a proper bearing RCA is to remove the bearing from the gearbox, separate the rings and rollers and determine the origin(s) of the failure. Prior to removing the bearing from the housing it must be photographed to document the as found condition. It is helpful to write on the components with a paint pen to indicate part orientation prior to removal from housing. Look for evidence that the outer ring has spun in the housing or that the inner ring has spun on the shaft journal, abrasive wear and the nature of fretting corrosion often provide valuable clues.
The bearing rings must be separated carefully without damaging any evidence. Bearings generally are held together by rivets in the bearing cage, snap-rings, or end covers. Bearing disassembly is not without hazards and a proper risk assessment should be conducted and reviewed by the site safety manager prior to work commencing. Some of the hazards that must be mitigated during bearing disassembly include:
Once the bearing failure mode has been identified the root cause analysis work begins to determine the origin of the failure. Root cause analysis must follow an orderly process to achieve conclusive results. A typical wind turbine bearing failure RCA will incorporate the following 12 steps:
Documentation requirements, which are established in Step 1 must be followed by all parties involved in the RCA.
This article focuses on two key steps in the RCA process, identification of potential failure modes and metallurgical investigation. These steps should be performed by engineers who are experts in bearing failure investigation.
In order to identify all of the potential causes for failure it is essential to assemble all persons with knowledge of the wind turbines operation as well as relevant experts for a brainstorming session. The goal of the meeting is to leave no stone unturned and list out all potential causes so that they can be formally addressed and the output a Ishikawa or fishbone diagram that organizes the potential causes into groups. Once the diagram is created the group then ranks each potential cause by its perceived influence on the failure. To help guide the team to the root cause, all items with Moderate, High or Further Analysis influence factor rankings should be investigated further. The group then defines the documentation or operational data that must be reviewed or testing that must be performed prior to the item being eliminated as a potential root cause. Figure 5 shows a typical fishbone diagram for such an investigation.
In order to verify the items shown in the "Materials" bone of Figure 5 it is often necessary to send the failed bearings to a qualified metallurgical laboratory for testing and review by a metallurgist. At the lab the bearing may be sectioned, for example, in the axial and circumferential directions with samples prepared as shown in Figure 6.
There are many tests that can be performed and an initial round of testing often includes the following:
Much can be learned about the failure mode from these tests, such as evidence of grinding burn under the Nital etch, evidence of sub-surface cracks originating from inclusions, an inadequate material hardness or a unsatisfactory steel microstructure.
If any of the above tests prove that the material was insufficient then a study needs to be conducted of the manufacturing quality documentation to identify the total affected population. If the metallurgical testing does not reveal a root cause then the investigation should focus further on the gearbox design, assembly, and operation.
Romax routinely perform RCA on failed wind turbine bearings and find the critical items in the investigation are generally factory teardown, metallurgical study and gearbox or bearing simulation. With a detailed approach to these tasks, buttressed by the significant application expertise, the root cause can generally be identified. This allows for the OEM or owner to transition quickly to assessment of corrective actions and risk/cost reduction.