Power Gear Box (PGB) advanced planet bearings development

Final Report Summary - ABAG (Power Gear Box (PGB) advanced planet bearings development)

Executive Summary:
The current planetary bearings including the surrounding components are designed according to pragmatic rules which are less and less relevant to the conditions predicted for the next generation of engine concepts and the related power gearbox requirements because of higher torque density, more severe working conditions, higher temperatures and simultaneously reduced weight. These resulting overall gearbox requirements together with the need for reduced weight lead finally to drastically increased load and speed conditions for the planetary bearings which can not be compensated by increasing the bearing dimensions. With current bearing designs, materials and analytical tools the requirements on life capability, reliability and contamination resistance cannot be sufficiently met and are way above existing applications and experiences. Therefore it is necessary to improve the surface and near surface robustness in the contact areas of planet bearing systems to take full advantage of the material inherent life potential without taken the risk of surface initiated premature system failures.

To overcome these problems, the following main RTD activities are a necessity:

1. New, improved and more durable materials and material processing technologies in order to increase
• component life under high load and speed
• wear resistance under contamination and boundary lubrication

2. New and verified stress and life analysis methods to fully consider the improved material capabilities

Also by various steps to improve the melting procedure for the base material Ferrium C61 and various improvements for the applied heat treatment the test results gained still showed not sufficient results for the high demanding application. A fishbone diagram was used to investigate the failure root cause and all potential root causes were investigated. These investigations let to certain improvements with regard to the achieved material properties but were not satisfying the application. Potentially nitriding heat treatment could be beneficial but performed trials did not reach the required depth.

The M50NiL-DH material showed a very good performance at the endurance, contamination and spall propagation test. No sudden ring failure could be observed. The spall propagation speed is dependent on the initial spall size. In average additional 0,8x to 2,4x of the running time until the initial spall occurred, could be reached until 20% of ring circumference is spalled. Nevertheless, these results serve as a very profitable Baseline for further material and heat treatment studies and developments.

Project Context and Objectives:
Today’s engines and their main components such as reduction gear boxes, rely on traditional rolling element bearings of steel. No alternative technology is currently available off the shelf. Based on the experience gained in the earlier Brite Euram projects ‘Advanced Surface Engineering Techniques for Future Aerospace Transmissions’ (ASETT), this programme proposes the validation of innovative bearing technology under extreme engine conditions utilising advanced materials and surface engineering techniques, in order to increase engine efficiency.

The current planetary bearings including the surrounding components are designed according to pragmatic rules which are less and less relevant to the conditions predicted for the next generation of engine concepts and the related power gearbox requirements because of higher torque density, more severe working conditions, higher temperatures and simultaneously reduced weight. These resulting overall gearbox requirements together with the need for reduced weight lead finally to drastically increased load and speed conditions for the planetary bearings which can not be compensated by increasing the bearing dimensions. With current bearing designs, materials and analytical tools the requirements on life capability, reliability and contamination resistance cannot be sufficiently met and are way above existing applications and experiences. Therefore it is necessary to improve the surface and near surface robustness in the contact areas of planet bearing systems to take full advantage of the material inherent life potential without taken the risk of surface initiated premature system failures.

To overcome these problems, the following main RTD activities are a necessity:

1. New, improved and more durable materials and material processing technologies in order to increase
• component life under high load and speed
• wear resistance under contamination and boundary lubrication

2. New and verified stress and life analysis methods to fully consider the improved material capabilities

Over the last decade FAG AC has developed and verified surface engineering technologies for rolling element bearings and successfully introduced in several new engine programs. These technologies have shown a bearing life improvement of more than 10-times compared to standard technologies, especially under high load, contamination and low lubrication conditions, and are promising candidates for being further developed for use in gearbox planetary bearing applications and integrated gears. In order to adapt and transfer the technologies the following main steps are necessary:

• Definition of a suitable technology adaption and integration plan
• Demonstration of the choosen technologies through experimental testing

Project Results:
Design Study of Advanced Planet Bearing
In the preliminary design study the selection criteria for advanced bearing design / technologies and the required material characteristics was defined to meet the requirements of
• extended bearing life under
o high load
o oil contamination
o wear conditions

A planet bearing design study was performed. The bearings, subject of this study, shall be applied in the planetary stage of future aircraft engine gearboxes. Such gearboxes shall provide a higher power density at minimum weight and improved operational reliability than state-of-the-art gearboxes which in parallel means for the associated bearings higher load, speed and live capabilities.
Therefore the design study was considering the following aspects:

- Suitable types of bearings
- Detailed analytical investigation and calculation
- Potential bearing and gear materials
The specific operating conditions with regard to speed, load, centrifugal forces and mesh forces were analysed and calculated.
Technology Development
Two different materials were identified for the test campaign. Ferrium C61 is an innovative high strength steel and M50NiL (high speed tool steel) as Baseline, because it is the state of the art material for aerospace bearings.

The material characteristics of Ferrium C61 as lined out in the material data sheet are very promising to fulfil the requirements of planetary bearings, because of the following aspects:
- very high fracture toughness of > 140 MPa?m
- high tensile and yield strength
- high core hardness of 47 – 50 HRC
- good surface hardness of 60 – 62 HRC
- possibility of plasma nitriding to increase the surface hardness
- good thermal stability

There are two different requirements on the properties regarding the case hardness profiles of gears and bearings.

The heat treat processes for M50NiL and Ferrium C61 were investigated and the process parameters were optimised to achieve the desired surface hardness and case depth. Many heat treat tests using gas carburizing, plasma nitriding and gas nitriding procedures were processed and analysed to find the optimal process parameters.

For the hardware tested in this project a special heat treating procedure was developed. This procedure consists of low pressure carburizing, vacuum hardening and tempering. Figure 1 shows typical hardness profiles and a micrograph of the carburized case of Ferrium C61 heat treated with this process. The process details are AVIO proprietary.

Figure 1: Hardness Profiles and microstructure of two specimens (discs), Ferrium C61
carburized and hardened

The microstructure looks good, no carbide network, the carburized layer is uniform, no marked variation in case depth. Figure 2 shows a residual stress profile of one disc. The stress profile shows residual compressive stress which is typical for a carburized layer.

Figure 2: Residual stress profile of a specimen (disc) Ferrium C61
carburized and hardened

A part of the test hardware should be nitrided in order to increase fatigue life under severe conditions. Typical nitriding processes are performed at temperatures above 500°C, Ferrium C61 is sensitive against tempering above 500°C, therefore a study was performed to show the limit of the nitriding temperature of Ferrium C61. For this, specimen of Ferrium C61 were carburized, hardened and exposed to annealing cycles in protective atmosphere at several temperatures, each for a duration of t1. This should simulate a nitriding process at the respective temperature. Figure 3 shows hardness profiles of the carburized case after different annealing cycles. Only after annealing temperatures of T1 and T2 no significant reduction in hardness is visible. This means, that the nitriding temperature for Ferrium C61 should not exceed T2. As the diffusion rate of nitrogen is comparatively low at this temperatures, either the nitriding depth is low or a long cycle time is necessary.

Figure 3: Hardness profiles of specimens Ferrium C61, carburized and hardened
after annealing at different temperatures (T1

Figure 4 shows hardness profiles of Ferrium C61 after plasmanitriding and gasnitriding. In general the nitriding depth is low due to the low nitriding temperature. The hardness measurement after nitriding was done with an indenter load of 300gf (HV 0.3) in order to get a tighter spacing of the indentations and a lower distance of the first indentation to the surface.

Figure 4: Hardness profiles of specimens, Ferrium C61, carburized and hardened
after different nitriding cycles, the content of nitrogen in the process gas
was different in the plasma nitriding cycles (t1
Figure 5 to 7 show the microstructure of the nitrided layer of all three nitriding variants. In case of the gas nitrided specimen there are marked precipitates at the grain boundaries and slight formations of a white layer on the surface.

Figure 5: Microstructure of Ferrium C61, carburized, hardened and plasmanitrided (T2/t3)

Figure 6: Microstructure of Ferrium C61, carburized, hardened and plasmanitrided (T2/t1)

Figure 7: Microstructure of Ferrium C61, carburized, hardened and gas nitrided
(T2/t2) IGNs (black arrows), white layer (red arrows)

For nitriding of the test hardware (small scale test bearing IRs) the plasma nitriding process with a cycle time t3 was applied (t3>t2>t1).
The evaluation results of these test bearing IRs are shown in Figure 8 (hardness profiles) and Figure 9 (microstructure).

Figure 8: Hardness profiles of test bearing IRs, Ferrium C61, carburized, hardened and
plasmanitrided (T2/t3)

Figure 9: Microstructure of test bearing IRs, Ferrium C61, carburized, hardened and
plasmanitrided (T2/t3)
slight white layer but no IGNs

Test Plan
A test plan (consisting of 3 different tests: Endurance, contamination and spall propagation) was developed to substantiate bearing life improvement and applicability to gears. The endurance tests were performed under low lambda conditions, which can occur during operation. The contamination tests are running with pre-damaged inner rings caused by a HRC indenter. For the spall propagation tests the spalled bearings of the contamination tests are used under a reduced pressure. The spall propagation is monitored several times (depending on the propagation speed) until 20% of the inner ring circumference is spalled.

M50NiL plasma nitrided rings were used as the Baseline, since numerous previous test series had shown that plasma nitrided M50NiL bearing components provided the highest fatigue limit for bearings compared to the standard materials in use for aerospace applications.

Endurance Testing
In the following table the test results of Endurance test under mixed lubrication are shown:
Material Bearings tested IR failures B10 life* Weibull slope
M50NiL-DH 12 6 746x 3,88
Ferrium C61 13 8 9x 0,51
Ferrium C61-DH 13 6 1x 1,10
* referenced to the calculated life according to DINISO 281

The slope of the partial regression line is very smooth and reaches only a ?-value of 0,51 for Ferrium C61. If the ?-value is smaller than 1 it is usually an indication for an unsteady process leading to infant mortality. In the case of wear and fatigue a ?-value bigger than 1 was expected, as it was determined in the M50Nil-DH testing (? = 3,88). The plasma nitriding of Ferrium C61 was resulting even in a deterioration of the life time capability of Ferrium C61.
Some of the Ferrium C61 bearings showed a long running time without a failure, indications that Ferrium C61 has basically potential for good performance under mixed lubrication conditions.

For remaining parts, investigation performed indicated an infant mortality issue related to an early melting practice and a possible improvement of the carburising process.

Therefore further tests with improved material (of the latest melting practise) and improved heat treatment process of Ferrium C61 were performed.

In the following table the test results of Endurance test under mixed lubrication are shown:
Material Bearings tested IR failures B10 life* Weibull slope
Ferrium C61 (old) 13 8 9x 0,51
Ferrium C61 (new) 12 7 7x 0,51
* referenced to the calculated life according to DINISO 281

The repetition test of Ferrium C61 with the optimised heat treatment and material of the latest melting practise showed no life time improvement at all. The Weibull curve is a duplicate of the previous test. The implemented modifications were not appropriate to improve the capability of Ferrium C61.

Due to the fact of the infant mortality of Ferrium C61 a second endurance test under full lubrication conditions should clarify whether we have surface initiated or subsurface initiated failures. The test bearings were from the same manufacturing lot as used for the endurance test under mixed lubrication.

In the following table the test results of Endurance test under mixed and full lubrication are shown:
Material Bearings tested IR failures B10 life* Weibull slope
Ferrium C61 (mixed lubrication) 12 7 7x 0,51
Ferrium C61 (full lubrication) 6 3 1x 0,42
* referenced to the calculated life according to DINISO 281

Both Weibull curves are almost identical and are characterized by a very smooth slope. The use of full lubrication conditions did not show any improvement compared to the mixed lubrication conditions. This is a strong indication that we are faced with subsurface fatigue instead of surface initiated fatigue.

In the literature a hind could be found that high cobalt based high strength steels show a poor rolling contact fatigue behavior. In Erwin Zaretsky’s book “Tribology for Aerospace Applications, page 343f “ only this statement can be found, but an explanation of the reason is missing.

The contamination test was only performed with the Baseline Material M50Nil-DH because of the not yet satisfying results of Ferrium C61 and Ferrium C61-DH in the endurance testing. These tests will be performed when Ferrium C61 with enhanced heat treatment would be available. The bearing test campaign was performed with 12 bearings. The maximum test time was fixed until an inner ring spall occurs (apart high running time tests which were suspended).

Contamination Testing
The contamination test was only performed with the Baseline Material M50Nil-DH because of the poor results of Ferrium C61 and Ferrium C61-DH in the endurance testing. The bearing test campaign was performed with 12 bearings. The maximum test time was fixed until an inner ring spall occurs (apart high running time tests which were suspended).

The pre-damaging of the inner rings was performed by a modified Rockwell indenter. In total 8 HRC indents with different angles (19,2° to 41,6°) distributed over the whole running track circumference were inserted in the inner ring (see figure 10). The indent size was fixed to a diameter of 160 µm.

Figure 10: Indentation pattern on inner ring raceway

In the following table the test results of contamination test are shown:
Material Bearings tested IR failures B10 life* Weibull slope
M50NiL-DH 12 11 10x 1,23
* referenced to the calculated life without considering the influence of contamination

All failed parts showed a typical inner ring pitting at position 1 (19,2°) to position 4 (28,8°), as it was expected. All other components (balls, outer ring and cage) showed no failures. This contamination test will be used as the Baseline for further high strength material development for bearings.

Spall Propagation Testing
The failed bearings of the contamination test were used to show the spall propagation behaviour of M50NiL-DH. The initial spall size after the Contamination Test was documented and determined by microscope (figure 11).

Figure 11: Example of Initial spall size after contamination test

The spall propagation was documented after several hours depending on the propagation speed of the inner ring. The shut down of the test rig was regulated by the vibration signal. The threshold was set to a 10% vibration increase based on vibration level starting the test. The goal was to generate at least 3 to 4 data point for each test bearing.

Each spall propagation test was continued until 20% of the raceway circumference was spalled. In figure 12 an overview of the spall propagation results of the material M50NiL-DH is shown.

Figure 12: Spall Propagation Behaviour of M50NiL-DH material
The limit of “20% of the circumference is spalled” is reached between additional 0,8x to 2,4x of the running time the initial spall occurred. It is obvious that the initial spall size influence the spall propagation speed. If the initial spall is below a certain size (in this case smaller than 1% of the circumference) no spall propagation can be observed. The dwell point lies between 6% to 10% spalled area of the circumference. If this failure size is reached, the propagation speed is accelerated immediately, because 2 rolling elements are in contact with the failed raceway at the same time. This spall propagation test will be used as the Baseline for further high strength material development for bearings.

Potential Impact:
The results of the project should initially be exploited at an early stage by application to current development projects as a demonstration and test of the new design and materials technologies. It was planned that the technology acquired would be fully utilised in new commercial aero-engines (e.g. CROR). Due to the non satisfying performance of Ferrium C61 as a new high strength material for planetary bearings, it was decided not to disseminate the results in publications, conferences or flyers. The very promising results of M50NiL-DH and the improvement on the characterisation process and on the knowledge of the material properties are considered highly sensitive information and can not therefore be published. All other information regarding the material treatment and properties were already disseminated in various publications, conferences and flyers over the past years.

The plasma nitriding and gas nitriding of M50NiL was patented several years ago. Because the carburizing and the gas and plasma nitriding processes of Ferrium C61 did not lead to the expected results regarding rolling contact fatigue, it was decided not to apply for a patent for these technologies.

Documents connexes

final1-d3-2-final-report-abag.pdf

Final Report Summary - ABAG (Power Gear Box (PGB) advanced planet bearings development)

Documents connexes

Partager cette page

Télécharger