Introduction With its higher efficiency than solar panels, wind power is a fast-growing worldwide industry, with over 650GW of capacity installed all over the world now. 1.9 GW of freshly installed offshore wind capacity brought the total German fleet to 56.1 GW in 2021. This makes up over 28.000 turbines for now with number rising. Aus Wissenschaft und Forschung 31 Tribologie + Schmierungstechnik · 70. Jahrgang · 2/ 2023 DOI 10.24053/ TuS-2023-0010 Simplified tribological approach for predesign of wind turbine bearing cases, combined with model test investigation Ivan Grozev, Sagar Dalal, Nazlim Bagcivan, Serhan Bastuerk, Christian Lueffe, Thomas Stahl* Dieser Beitrag wurde im Rahmen der 63. Tribologie-Fachtagung 2022 der Gesellschaft für Tribologie (GfT) eingereicht. Mit dem Ziel, unsere Vorgehensweise des Produktentwurfs zu optimieren und durch den Übergang von verschwenderischen Feldtests zu schlanken tribologischen Modelversuchen die Nachhaltigkeit des Endprodukts zu erhöhen sowie ein wirkungsvolles und umsetzbares Systemranking für Lagergehäuse von Windkraftanlagen zu erreichen, befassten wir uns näher mit den folgenden gängigen Berechnungsansätzen: • Plint’s and Alliston-Greiner’s Energy Pulse (EP) • Matveesky’s friction power intensity (FPI) • Transmitted Energy (TE) Mit dem Konzept der übertragenen Reibenergie (Ef) erreichten wir zufriedenstellende Ergebnisse. Des Weiteren stellten wir fest, dass Zwei-Scheiben- Tribometer verwendet werden können, um einen Gleitvorgang von 100 % zu simulieren. Dies erleichtert den qualitativen und quantitativen tribologischen Vergleich von Systemen mit unterschiedlichen Schlupfraten durch die Verwendung von nur einer Maschinentype, was die Messsicherheit erhöht. Schlüsselwörter Produktentwurf, Nachhaltigkeit, Reibenergie, Reibintensität, Energiepuls, Zwei-Scheiben Tribometer, Tribometer, Tribometrie, Tribolabor, Tribologie With the purpose of improving our predesign approach and increasing the end-product sustainability by going from resource wasting field testing to slender tribological model testing, we looked closer to commonly relevant calculation approaches as • Plint’s and Alliston-Greiner’s Energy Pulse (EP) • Matveesky’s friction power intensity (FPI) • Transmitted Energy (TE) to achieve fast and feasible tribo-system ranking for wind turbine bearing cases. With the concept of transmitted friction energy (Ef), satisfactory results were reached. We learned that 2-Disc tribometer may be used to simulate 100 % sliding, making comparison of different slippage rate easier by only using one tribometer, thus reducing measurement inaccuracies caused by different type of machines (e.g., pin on disc) or operator’s influence. Keywords Product predesign, sustainability, transmitted friction energy, friction power intensity, energy pulse, twodisc machine, tribometer, tribometry Kurzfassung Abstract * Ivan Grozev 1 (federführender Autor) Sagar Dalal 2 Dr. Serhan Bastuerk 1 Dr. Nazlim Bagcivan 1 Christian Lueffe 2 Dr. Thomas Stahl 2 1 Schaeffler Technologies & Co.KG, 91074 Herzogenaurach 2 Schaeffler Technologies & Co.KG, 97421 Schweinfurt • Pin type cages that consist of two side washers and a bolt • Stees Segment Cages (SSCs) optimized for clean and easy production. Within our investigation, we looked closer at the Archard’s wear law, the Matveesky’s friction power intensity, Plint’s and Alliston-Greiner’s energy pulse approach and considered that the concept of transmitted friction energy (Ef) appears to rank the investigated cage systems in a clear and relatively precise way. Compared to a system description, mainly based on the product of load and relative velocity (p*v), the transmitted energy approach is much more accurate, quickly pinpointing possible system performance challenges and thus directing a pre-design process more precise. Additionally, we learned that the two-disc tribometer may be used to technically simulate complete sliding, making comparison of different slippage rate easier by only using one tribometer, thus reducing measurement inaccuracies caused by different type of machines (e.g., pin on disc) or operator’s influence. Simplified tribological approach for predesign of wind turbine bearing cases, combined with model test investigation At the beginning of the 20 th century, the American essayist and naturalist John Burroughs said: “The fuel in Earth will be exhausted a thousand or more years, and its mineral wealth, but man will find substitutes for these in the winds, the waves, the sun’s heat, and so forth.” Aus Wissenschaft und Forschung 32 Tribologie + Schmierungstechnik · 70. Jahrgang · 2/ 2023 DOI 10.24053/ TuS-2023-0010 With wind energy pros as a clean and renewable source of energy, relatively low operating costs, effectiveness and decreasing prices due to technological advancements and increased demand, there is little surprise that a global player as the Schaeffler Group concentrates in various solutions specially designed for wind energy turbines. For the blade adjustments, the Schaeffler Group offers plain bushes, spherical plain bearings, deep grove ball bearings, planet gears and output shafts. For the rotor shaft - the single bearing concept and sensor monitoring. In the gearbox there is the planet carrier, the planet gear, the hollow shaft, the intermediate output shaft. There are solutions for the generator and wind tracking. With such a product variety, there appears naturally the need of clear design vision and the possibility of prompt and effective model testing. In this work we try to answer the following questions: • Is it possible to predict, due to design qualities which bearing cage is more vulnerable to wear, this reducing the tendency of overengineering, and on the other hand, pile up on robustness for bearings that tend to need it more? • Can a two-disc model test deliver plausible answers for pure sliding contacts with the purpose of substituting materials in order to optimize and concurrently reduce costs of the production process? We concentrated on four different types of bearing cages for the latter investigation: • Inner ring guided FBP thrust ball bearing cages • JP3 type roller-guided sheet metal cages Figure 1: Overview of Schaeffler Group product applications for wind turbines This statement is as valid today as more than a hundred years ago. 1.9GW of freshly installed offshore wind capacity brought the total German fleet to 56.1GW in 2021 [1]. This makes up over 28.000 turbines for now with numbers raising. The latter message is taken from the study commissioned by the country’s wind association BWE. With wind energy pros as a clean and renewable source of power, low operating cost (1-2c per KWh), effectiveness and decreasing prices due to technological advancements and increased demand, there is little surprise that a global technology player as the Schaeffler Group concentrates on various solutions specially designed for wind energy turbines (see Figure 1 below): For the blade adjustments we offer plain bushes, spherical plain bearings, deep groove ball bearings, planet gears and output shafts. For the rotor shaft, we have the single bearing concept and sensor monitoring. In the gearbox there is the planet carrier, the planet gear, the hollow shaft, the intermediate output shaft. There are solutions for the generator and wind tracking. Additionally, the Schaeffler Group offers condition monitoring, simulation, calculation and testing. With such a product variety, there appears naturally the need of clear design vision and the possibility of prompt and effective model testing. We tried to answer, via tribological means the following questions: • Is it possible to predict, due to design qualities which bearing cage is more vulnerable to wear, thus reducing the tendency of overengineering, and on the other hand, pile up on robustness for bearings that tend to need it more? • Can a two-disc model test deliver plausible answers for pure sliding contacts that correlate with our field experience with the purpose of substituting materials in order to optimize and concurrently reduce costs of our production process? We concentrated on four different types of bearing cages for our investigation: • Inner ring guided FPB thrust ball bearing cages, which are used successfully for many years. • JP3 cages, which are roller-guided sheet metal cages for large size tapered roller bearings. Generally, the validation of the JP3-design logic is based on track record with over 150 different JP type cages and over 6.000 cages in operation. • Pin type cages that consist of two side washers and a bolt. To achieve a higher wear resistance the bolt is heat treated and additionally coated. • Steel Segment Cages, or SSCs, that are optimized for clean and easy production. The design allows a flexible scaling from midrange wind bearings up to the largest of applications. The SSC cage design is currently available for double row tapered roller bearings (TRBs). During the first part of our tribological study, we were interested if a simplified calculation, based on compressed theory may correlate our field experience. Our purpose was to give more decisiveness and clearness of a pre-design process, based on facts and numbers, simultaneously saving the time and resources needed for simulation. We looked closely at the work done, due to friction as a measure of wear, based on four different approaches: • Archard’s wear law [2] - there are some limitations, including the fact that the law takes into consideration only the softer of materials and doesn’t contain any properties characterizing adhesion: • Matveesky’s friction power intensity (FPI) [3] - here we had the problem that it defines only the rate of energy generation and doesn’t take contact time into consideration. The normal load and the apparent area of contact in the FPI-equation can be alternatively substituted with the surface pressure p[N/ mm 2 ], resulting in: Within the Schaeffler Group we have gathered empirical knowledge for functional area FPI-values of certain coating types, and we use the latter equation for model test designing purposes. The Q f value may be considered a cross-contact energy and derived from our experience correlates well with the Archard’s wear coeff. k [mm 3 / N] at the end of the model tests. • Plint’s and Alliston-Greiner’s energy pulse approach (EP) [4] - this is a revised version of the FPI that can be “regarded as an incremental contribution to wear or surface damage in contact. Sum of Eps can be used as a measure of total wear” [5]. The energy pulse is only applicable to the type of contact in which the contact point moves relative to the two surfaces, and has obtained encouraging values, which can be used to describe the teeth gear [x]. V = Q f = . = μ [W/ mm²] Q f = = μ [ / ² ] = [mm³] . Aus Wissenschaft und Forschung 33 Tribologie + Schmierungstechnik · 70. Jahrgang · 2/ 2023 DOI 10.24053/ TuS-2023-0010 well. Taking the compressed energy that can be converted to temperature into consideration, we have set a critical border prospectively at 1000J per minute. Still this assumption needs to be further validated. Based on discussions with the application engineering and after looking at different contributing factors, the 100 % ranking of various cage designs is relatively difficult. Still the above results and the approximate classification are plausible, and this simple ranking shows good potential in pre-design approach, pinpointing optimization possibilities and where to look for them. The next important question of our study was if a twodisc model test can deliver plausible answers for pure sliding contacts that correspond to our field experience. The reason for not choosing a pin on disc test was that Schaeffler Group has gathered a considerable experience with two-disc tribometer in the past, even accomplishing to reproduce a very similar wear mechanism close to what we are used to see in field conditions. Additionally, we had a better representation of the contact area geometry, due to the flexibility a two-disc system can offer. On top, a rolling / sliding wear tester is one of the most popular machines for investigating wear and friction of systems under rolling, sliding or a combination of both conditions. We chose a two-disc tribometer by Optimol Instruments Muenchen. The machine operates tow discs, fixed to two parallel shafts, pressed against each other. Driven by a motor through a train of gear, the specimens are rotating along with the shafts. The rotating speed can be controlled, so that when the linear speeds of the two wheels are equal at the contact point (V 1 = V 2 ), a pure Aus Wissenschaft und Forschung 34 Tribologie + Schmierungstechnik · 70. Jahrgang · 2/ 2023 DOI 10.24053/ TuS-2023-0010 The transit times for the contact areas are: What we understood from the above approaches is that roughly kinematics, surface pressure, contact geometry and thereof derived contact ratios, velocity, temperature inand out-flow, lubrication and its regime, running time, material and surface properties play a decisive role in the tribological performance. We assumed, that a friction coefficient is closely related to the contact energy, which is a product of the above listed initial tribological system properties and can be regarded as a resulting resistance value needed to propagate a kinematic motion. Based on the latter recognition, we decided to use, we called it, the transmitted friction energy (E f ) as a comparison guidance in order to evaluate the studied cage types (see Table 1 below) The transmitted energy approach appeared to describe and thus rank the cage-systems more precisely, based on and correlating with our field experience, and give a clearer idea about detecting possible difficulties in a predesign stage than the generally assumed and largely used (p*v) [kg/ s 3 ]. Additionally, the E f - values give a hint if there is a trigger possibility for surface tribochemical processes as Ep = . = μ [J/ mm²] = 2 and = 2 . . ( ) E f = [ ] = μ Cage type FBP JP3 Pin type SSC Load (sta c force) in N 2466.66 384.36 2991.40 1.09 1.5 Contact pressure (static load) in N/ mm² 2.9 132.7 129.3 0.1 12.4 Rela ve velocity in m/ s at n=min -1 0.05 0.046 0.017 0.05 0.05 Cage assembly Rollers inserted directly Rollers inserted via hea ng procedure Pins inserted through rollers, welded, and threaded at both ends Combined segments Circumferen al usage in % 85 92 95 92 Size of cage in mm 500 - 2250 1000 - 3500 1500 - 3500 1000 - 2500 μ 0.4 0.2 0.3 0.2 E f in J at n=min -1 2960 212 915 0,65 0,90 Table 1: Cage system ranking based on E f and rolling contact is achieved. When V 1 and V 2 are different and both wheels are rotating, a combined rolling-sliding takes place [6]. We chose discs with elliptical contact areas as follows R 1X = 40 mm, R 2X = 40 mm; R 1Y = 320 mm, R 2Y = ∞, or simply disc with crowned radius versa a cylindrical disc (see the contact simulation below). Aus Wissenschaft und Forschung 35 Tribologie + Schmierungstechnik · 70. Jahrgang · 2/ 2023 DOI 10.24053/ TuS-2023-0010 Table 2: Average weight loss of diff. material systems after 2-disc model test Figure 2: Hertzian pressure contact simulation We decided to apply minimum of 2 N on the machine, resulting in 150.8 N/ mm 2 surface pressure, which fits quite well with contact pressures in large size real applications [7]. The precise testing conditions and the specific materials cannot be named for apparent reasons, but the result of the study is to be seen in the graph below. logical model test approach for comparison of systems with different slippage easier, by using only one type of machine and not a combination of e.g., pin on disc and two-disc tribometers. References [1] German Wind Energy Association Report 2021 [2] Archard J.F. “Contact and Rubbing of Flat Surface”. Journal of Applied Physics 1953 24 (8): 981-988. doi: 10.1063/ 1.1721448 [2] Archard J.F., Hirst W. “The Wear of Metals under Unlubricated Conditions” 1956-08-02 Proceedings of the Royal Society. A-236 (1206): 297-410. doi: 10.1098/ rspa.1956.0144 [3] Matveesky R.M. “The critical temperature of oil with point and line contact machines”. Trans. ASME. 1965; 87: 754. doi: 10.1115/ 1.3650672 [4] Plint M.A., Alliston-Greiner A.F. “The energy pulse: A new wear criterion and its relevance to wear in gear teeth and automotive engine valve trains”. 1996 Lubrication Science / Volume 8, Issue 3: 223-251. doi: 10.1002/ ls.3010080303 [5] Plint G. Cambridge University Tribology Course 2015, Slides [6] Dalal S. “Wear Investigation of Large Bearing Cages for Wind Turbines”. 2021 M.Sc. Thesis TU Bergakademie Freiberg / Schaeffler Group [7] Kock S., Jacobs G., Bosse D. “Determination of Wind Turbine Main Bearing Load Distribution”. Journal of Physics: Conference Series - 1222 (2019) 012030: 2-6. doi: 10.1088/ 1742-6596/ 1222/ 1/ 012030 Aus Wissenschaft und Forschung 36 Tribologie + Schmierungstechnik · 70. Jahrgang · 2/ 2023 DOI 10.24053/ TuS-2023-0010 The latter represents a number of tested systems with measured weight loss after the test. At least four test pro System were performed. The bar shows the average value of the single measurements. The orange bar represents always one and the same counterpart material. The results correlate to a larger extent with our field experience and observations thus proving that via careful test design and basic system understanding, a complete sliding can be plausibly simulated via a two-disc tribometer. This is true for a roller vs. cage contact but must be still validated for a shaft vs. bearing bushing combination. This realization opens new possibilities in tribological model testing methodology, as a complete sliding is usually simulated on a pin on disc machine. For systems where a simulated slippage ratio varies, performing model testing on only one machine may prove beneficial. The rough but essential lessons learned from this study is that system tribological properties can be promptly and efficiently described via a transmitted friction energy, taking into consideration mainly the friction coefficient, the static load and the relative velocity. Compared to a system, description, mainly based on the product of load and relative velocity (p*v), the transmitted energy approach (E f ) is much more accurate, quickly pinpointing possible system performance challenges and thus directing the pre-design process more precise. Additionally, we learned that two-disc tribometer may be used to technically simulate complete sliding, making tribo-