Blade Bearing Cage Failures
By Cory Mittleider on 2/10/21
Background
Slewing bearings use either spacers or a cage to separate rolling elements inside. These are important to ensure proper operation, load sharing and long life of the bearing in its intended application. Spacers are common in industrial applications and were used in early versions of blade bearings for Wind. As blades grew longer and deformation of the bearing increased spacers started to bind up leading to some failures and increase in pitch faults. They were then replaced with a steel cage which better masked these effects. Steel cages have been the standard in Wind pitch applications for a decade now. Once again turbine designs are growing and are now pushing the limits of this steel cage. Further introduction to blade bearings can be found here
Fig. 1: Components of a blade bearing
Current Situation
Some of the new generations of turbines with large rotor diameters are experiencing premature failures of the steel cage that separates the balls. Many sites are seeing this failure after about 5 years of operation, but some have seen as early as 2 years.
Replacement of blade bearings often costs over $200,000 in addition to lost revenue from the turbine downtime.
The extremely short bearing life combined with high cost of replacement is not acceptable and a real solution to this problem must be found.
Investigation
We have investigated multiple failures of this type and have cataloged the below failures.
External indirect signs of cage failure:
When cages in these bearings fail it will lead to an increase in operational torque and may even lead to rollers getting stuck on cage debris and sliding. As this failure progresses the three axes will have different pitch response times. This is a parameter turbine controllers monitor, and when this happens operators will start to see an increase in pitch related faults. Additionally, this increase in operational torque will apply additional loading to the entire pitch system such as motors, gearboxes, and electronics.
External direct signs of cage failure:
This is the most obvious external sign of blade bearing failures. This is first recognized with seal failure and grease leaking. If a new seal is installed it will fail quickly. This is because of the cage breaking down internally creating debris that that tears up and pushes out the seal. As this progresses pieces of the cage will be found sticking out of the seal gap. These can come in many shapes and sizes as seen in figures 2 & 3. These pieces are often sharpened and should be handled carefully. Also, these pieces fall out of the bearing to the ground creating a safety hazard for those working at or near the turbine.
In later stages of failure inspection in the seal gap will reveal balls bunching in some areas (Fig 4) and some areas with no balls or seals at all with a clean line of sight through the entire bearing! (Fig 5)
Internal appearance of cage failure:
It is important to look inside of the failed bearing for a full understanding of the cage failure and collect additional information in order to understand the root cause of the problem.
Here are several observations from the teardowns and inspections done by Malloy.
Blade side row vs hub side row:
In every case the blade side row had much further progress cage wear and failure than the hub side row. In one case (fig 6) the amount of cage found inside of the bearing accounted for less than 25% of the cage that the bearing started with. In this case there were 0 remaining close cage pockets on the blade side row, and only 50 cage pockets remained closed of the 111 it started with on the hub side row.
Fig. 6: Cage pieces removed from failed bearing
Size of cage pieces:
The range of size and shapes of remaining cage pieces show the progression of cage failure in this bearing. It is clear that many of the pieces were over-rolled during operation breaking them down resulting in very small steel flakes.
Fig. 7: Small cage debris
Shape of cage pieces:
The cages broke apart in 2 places. First is the cage edge near a ball pocket and second is on the ligament between ball pockets. Additionally, many pieces of the cage were concave from being force into the raceway.
Wear found on cage pieces:
Figures 11 and 12 show remaining segments of cage. There are multiple signs of wear in these 2 pictures
Cage edge wear
Bulging on all 6 edges near pockets
Cracks found on 5 of the 6 bulges
Ball pockets smaller than ball diameter
Displacement of material in ball pockets
The edge wear was found on every segment of cage removed from multiple failed bearings. Bulging of the edges and ball pocket compression was also common. Cage cracking/breaking through the cage edges is very common as many of the broken segments separated at this location.
Raceway inspection:
The raceways of failed bearings were carefully measured and graphed in figure 13. The colored bands displayed are the measured ball wear path on each raceway section, and the red line on upper and lower indicates raceway edges.
The bending moment of the bearing is easily seen on this graph with the wide ball wear path occurring in different directions on opposite halves of the bearing. The red ellipses on this graph identify where the ball wear path spilled over the edge of the raceway. This is called ellipse truncation, which will be covered in greater detail in another article.
Understanding the problem
There is clear evidence that these cages are subjected to much higher forces than intended:
Material movement in ball pocket
Extreme edge wear from rubbing on the raceway shoulders
Deformation from cage being forced in to the raceway
Understanding what happens to this bearing in operation is critical to understanding why these failures are happening.
What does this bearing look like during operation?
As previously covered in this article here the internal design of this bearing and the application allows for deformation of the rings and changing contact angles during operation
Fig. 14: Deformation of blade bearing during operation
The graphed ball wear path from failed bearings confirms the operating contact angle changes based on the applied load; i.e. this bearing type has a load dependent contact angle. Combining this empirical data with FE models exposes the full scope of this problem.
When the bearing is manufactured the raceways create a precisely controlled tubular ball path. In operation however; this ball path is distorted, and the ball is sent on a roller coaster ride. There is stenosis of the ball path in some areas and wide gaps between rings in other areas leading to a change in contact angle.
The change is contact angle in the raceway means that balls all around the raceway of this bearing have a different rolling radius causing them to roll at a different speed.
What causes high forces on the cage?
The job of the cage is to keep large groups of balls moving together as a unit. The deformation occurring in this bearing leads to a force buildup between balls for 2 primary reasons. First is the balls will have difficulty passing through the areas of stenosis causing other balls in the same cage segment to push and pull balls through this area. The second reason is the differential rolling speed of balls around the raceway leads to ball jamming with fast balls in the cage segment pushing against the slower balls.
All the forces described above are applied to the cage. Some of them directly cause the damage seen on the cage others are indirect.
Fig. 15: Change in contact angle under load leads to differential rolling speed of balls
Cage forces causing direct damage.
The displacement of material in the cage pockets, and the compression bulging on the cage edges are a direct result of the forces applied to the cage.
Cage forces causing indirect damage.
When the balls are pushing and pulling on one another in a cage segment this causes the cage to slip either inboard or outboard of the ball centerline. When this happens the cage edge contacts the raceway shoulder of the bearing. This is the reason all cages show heavy wear on the cage edges and also have damage on the raceway shoulder.
Most of these bearings have 3 x 120° cage segments around the circumference. The forces occurring drive the ends of these segments into the raceway. This is the reason for the extreme edge wear found on the segment ends. Also some of the segment ends have a concave shape from this.
Solving the problem
The root cause of this problem is a mismatch between system stiffness and selected bearing type. In order to properly address this problem the solution must address this root cause.
Why will a harder cage not solve the problem?
The cage failure is a symptom of the problem, not a cause of the problem. Making the cage out of a harder material does not address the deformation allowed in the application leading to ball path stenosis and changing contact angles which cause the failure. These problems will still exist and will still cause higher than intended loading to the cage. Additionally, there are other failure modes occurring in these bearings that have the same root cause, and a change in cage material do nothing to address those failures. These will be covered in later articles:
Ring cracking Part 1 --> **READ HERE**
Ring cracking Part 2 --> **READ HERE**
Ellipse truncation and edge failure
Ring cracking
What does a real solution look like?
The investigation found failures stemming from the type of slewing bearing selected; 2 row 4-point contact ball. It is well established that this bearing type requires the mounting structure to be sufficiently stiff.
Good news
There are alternate slewing bearing types that are better suited to handle structures with reduced stiffness and have been known for decades. The good news is this bearing has already been adapted to the special requirements of Wind pitch applications. Additionally it has already proven its suitability in Wind applications.
For more information on this special bearing type, what platforms it's currently available for, and the history of development and testing give me a call at 605-357-1076or email atcmittleider@malloyelectric.com
-Cory Mittleider