Authors: Dr.-Ing. Matthias Stammler, Karsten Behnke
Whenever we have the opportunity to look at damaged bearings, we as scientists love to do so. Be it the bearings we tested in our own lab or bearings directly from the field, it is always exciting to see what story they can tell us. This blog article will guide you through the steps we usually take to determine the condition and possible failure causes of a bearing.
Our team at IWES’ Large Bearing Laboratory (LBL) has tested more than 300 bearings: most of the tests are designed to produce damage on the bearing; some do just this, some do not, and some produce damage we did not expect. Afterwards, we look at the bearings and analyze the damage causes. Every once in a while, we also get to look at a bearing from an actual turbine and give our opinion about it to help partners from the wind industry to better understand what happened. While this involves some tedious work, it is fun to relate the operation to the final condition of the bearings. In this article, we will take a closer look at what exactly we do, using a large four-point blade bearing by way of example.
First, we take a look at the external condition: Are there any signs of relative movements of the nuts or washers on the ring? Did the ring move against its mounting surface? Are the seals in place? We then extract the balls from the fill holes in the rings (see Figure 1). This process involves two full rotations of the bearing. There might be some slip on the ball set, so a good approach is to mark the initial position of the cages so as subsequently to be able to relate them to the rings.
With all the balls out, we separate the bearing rings and the cages (Figure 2). Now comes the super fun part of getting the grease off them. But obviously not before we have taken some final samples. In a 5-meter bearing, you can expect about 30 to 40 kilograms (kg) of grease. It is worth taking a good look at the grease: Did it change its color from the initial state? Does it contain any water or visible particles?
So now we have clean rings, balls, and cages. At this stage, we can look at almost all the components of the bearings and get a lot of the answers we need: What do the raceways and cages look like? In what condition are the balls or rollers? Figure 3 shows an outer ring with a number of wear marks on it. These marks were the result of the first tests in the HAPT project.
Fine tuning our methods
Sometimes, we have to dig a bit deeper, and our beloved belt saw takes over and cuts the rings. With smaller pitch bearings, it is also possible to use a cutting torch, but this is noisy and not so easy, and with larger rings, the melted material often clogs up the cut. Achieving good support for the huge and heavy rings is interesting enough when they are intact, but with the first cut the residual stress is released and the ring becomes a wobbly pool noodle, which is a real headache the first time you try to get to grips with it. Our team has lots of experience with this meanwhile and also has some adjustable support racks and rigging cylinders which enable it to handle the rings safely and securely.
Figure 5 shows a piece of a bearing ring with various cuts to allow for subsequent microscopic analysis. Although the primary conclusion in this case is visible even to the naked eye, microscopic analysis provides additional surface data which are useful for our modeling approaches. It also gives us repeatable visualizations that allows us to compare different bearings.
The section cuts may be used for metallurgical evaluation as well. This gives a good idea of hardness depth, material composition, and sub-surface cracks, see Figure 6.
Taking this together with the external condition of the bearing, the grease samples, and our experience from looking at test and field bearings allows us to draw conclusions about damage modes and causes most of the time. In the case of our own tests, the measurement data stored in our database completes the picture. But let’s stick with the types of damage seen in the figures above: The wear marks on the rings (Figure 3 and Figure 5) stem from tests specifically designed to create such marks. The crack in the hardened layer (Figure 6) is part of the fatigue damage caused by rolling contact. This is just one of the possible types of damage. Most common in wind turbines are ring cracks and cage damages of pitch bearings.
It is particularly enjoyable to look at bearings from turbines and compare the issues and types of damage with those we produce at the test rigs. This allows us to test our understanding and to develop our test programs and calculation models further. It additionally helps our partners to operate their pitch systems safely and reliably.
Most of the images in this article stem from the publicly funded HBDV project. The purpose of the tests was to achieve a better understanding of the factors that affect raceway wear. The results are due to be published soon and can be accessed as a preprint here.
More information here: