Introduction The friction behavior of wet clutches is notably impacted by how the friction interfaces interact with the lubricant. A degradation or performance loss of the latter can cause adverse NVH effects, perceptible as shudder noise to the driver. In a worst-case scenario, this can damage the entire powertrain unit. The decisive factor for this adversity is the course of the coefficient of friction (CoF) versus the sliding speed. When a negative slope or a static CoF considerably surpassing the dynamic CoF is present, the control behavior deteriorates, leading to an increased tendency to shudder [3]. This behavior can be obtained from Figure 1. To avoid these adverse effects, oil additives, such as friction modifiers (FM), serve to shape the friction behavior of wet clutches [4-6]. However, in addition to the gradual aging of the lubricants due to continuous energy input over their lifetime, external factors, such as water contamination, can, due to physical adsorption between water molecules and additive elements, accelerate this performance loss [7-9]. Numerous studies have indicated that water impacts the friction behavior of wet clutches, leading to a higher static CoF and a negative gradient of the CoF versus the slip speed [7,10 -14]. Besides that, Yokomizo et al. [15] detected an increase in the static CoF due to iron contamination by comparing the friction behavior at the be- Science and Research 23 Tribologie + Schmierungstechnik · volume 71 · issue 5-6/ 2024 DOI 10.24053/ TuS-2024-0035 Impact of Water Contamination and Iron Particles on the Performance Loss of e-Drive Transmission Fluids in Wet Clutches Johannes Wirkner, Mirjam Baese, Astrid Lebel, Charlotte Besser, Hermann Pflaum, Katharina Voelkel, Thomas Schneider, Karsten Stahl* submitted 20.12.2023 accepted: 20.12.2024 (peer review) Presented at GfT Conference 2023 The present study delves into the connection between the driving comfort and safety of electric vehicles and the e-drive transmission fluid (ETF), examining how water influx arising from environmental factors and iron particles resulting from wear of various transmission components can lead to performance degradation in the lubricant. The performance loss is evaluated based on the changes in the friction behavior of wet multi-plate clutches. Previous studies on automatic transmission fluids (ATF) have shown that water and iron contamination contribute to the earlier failure of transmission components and tend to lead to NVH behavior (noise, vibration, harshness), particularly in wet clutch systems [1,2]. Hence, it is essential to expand this knowledge on e-drive lubricants to prevent premature damage to various transmission components by exchanging the lubricant. This paper provides insights into component test rigs and associated test methods to determine the performance loss of the e-drive transmission fluid in an application-related test environment. Exemplary results are shown from investigations with wet clutch parts from a transfer case used in serial automotive applications, considering the influence of water contamination and iron particles as well as their mutual interactions. Keywords NVH, Shudder, Water Contamination, Iron Particles, Lubrication, Additives, ICP-OES, Friction Behavior, Wet Clutch, ETF, E-axle Abstract * Johannes Wirkner, M.Sc. 1 (corresponding author) Dr.-Ing. Mirjam Baese 2 Dipl.-Ing. (FH) Astrid Lebel, M.Sc. 3 Mag. Dr. Charlotte Besser 3 Dr.-Ing. Hermann Pflaum 1 Dr.-Ing. Katharina Voelkel 1 Thomas Schneider, M.Sc. 1 Prof. Dr.-Ing. Karsten Stahl 1 1 Technical University of Munich (TUM), School of Engineering & Design, Department of Mechanical Engineering, Gear Research Center (FZG), Boltzmannstrasse 15, D-85748 Garching, Germany 2 Magna Powertrain GmbH & Co. KG, Industriestrasse 35, A-8502 Lannach, Austria 3 AC2T research GmbH, Viktor-Kaplan-Strasse 2/ C, A-2700 Wiener Neustadt, Austria which was operated in slip mode. In contrast to other testing methods, component test rigs combine the advantages of the application reference of field tests with a simultaneous reduction in ongoing testing costs. Figure 2 shows an overview of the test rig setup with its modules. The clutch disks used in transfer cases from serial automotive production are mounted within the inspection chamber onto the respective carriers. While the inner carrier, which is connected to a specially designed shudder unit to reduce the torsional stiffness of the shaft, rotates, the outer carrier is fixed to the housing. The lubricant is injected in a radial direction from the inside via an oil nozzle. A force-controlled piston applies the pressure hydraulically. The creep drive accelerates the friction disks in the engaged clutch state to a defined rotational speed, which is determined by an incremental encoder at the end of the shaft. A complete clutch pack comprises six steel and five friction disks, leading to ten friction interfaces. Both disks are mainly defined by their mean diameter of 116 mm and the radial width of 18 mm of the friction surface, respectively. A detailed description of the test parts can be found in [2]. Oil samples (0.8 L per test) of the lubricant L-401 that were previously damaged in test rigand vehicle-based endurance runs in E-axles are evaluated by the friction behavior of wet clutches. General technical data of the lubricant are listed in Table 1. Untreated rainwater is injected to replicate genuine environmental conditions. The respective element content of the oil samples is determined by inductively coupled plasma optical emission spectrometry (ICP-OES) after microwave digestion Science and Research 24 Tribologie + Schmierungstechnik · volume 71 · issue 5-6/ 2024 DOI 10.24053/ TuS-2024-0035 ginning and the end of a durability test. Moreover, a correlation between water contamination and wear debris, specifically iron particles, has been discovered, eventually resulting in a reduction of the CoF [16,17]. The awareness of these effects obtained from conventional ATFs can now be extended to E-axles and ETFs within the scope of the present investigations. Due to the high tribological requirements, wet clutches are the component for evaluating failure criteria. Experimental Setup and Methodology The experimental investigations in the presented study were carried out on the KLP-260 component test rig, Figure 1: Course of the CoF versus the slid-ing speed based on Naunheimer [3] Figure 2: Schematic setup of the KLP-260 component test rig based on Meingassner [18] Lubricant Kinematic viscosity at 40 °C Kinematic viscosity at 100 °C Density at 15 °C L-401 21.7 mm²/ s 5.0 mm²/ s 0.853 g/ ml Table 1: Technical data of the tested lubricant with nitric acid (iCAP 7400 ICP- OES Duo, Thermo Fisher, Waltham, Massachusetts, USA) [19]. The experimental testing aims to determine a critical water amount that can be added to the lubricant without leading to adverse shudder behavior of the clutch. Therefore, an iterative test procedure is developed. After a run-in phase of 100 cycles, a pre-defined amount of water is added to the test rig. The lubricant is circulated for 15 minutes within the test rig to establish a homogeneous distribution of the added water. Then, a sequence with pressure p = 1.0 N mm -2 and rotational speed Δn = 100 rpm consisting of three steady and transient slip shifts in each configuration is run, respectively. The added water is entirely evaporated by an additional 90 cycles of steady slip, and if the system does not show any adverse NVH effects, an increased amount of water is added. This iterative process is repeatedly executed until irreversible shudder behavior occurs. The details of this test procedure can be obtained from Figure 3. To quantify the performance of the respective oil samples with their respective pre-testing background, the parameter m µ that represents the gradient of the CoF versus the sliding speed is used. Due to its initially described relation to the shudder tendency of wet clutches (Figure 1), this value enables the monitoring of the lubricant status at the addition of different water amounts. m µ is determined by calculating the average gradient of the CoF versus the sliding speed in the range of a rotational speed ∆n = 5 … 25 %. The value is again averaged for statistical validation over the three transient slip cycles. The calculated energy input corresponds to the work performed by various machine elements during previously conducted test rig and vehicle-based endurance runs. It should be noted that the efficiency of the overall application must be subtracted from this energy input to ultimately deduce the energy input into the lubricant. However, this is of secondary relevance for the investigations presented below. Results and Discussion Figure 4 shows the relationship between the gradient of the CoF m µ , which was evaluated at different levels of water contamination. The individually colored data points refer to a specific energy input brought into the lubricant by the previously performed endurance runs, ranging from dark blue for low energy input to dark red for high values. Generally, a gradient reduction can be obtained by further increasing the added water amount. The corresponding values of 100 % energy input and iron content were chosen according to the specifications of the manufacturer, where exceeding these limits would result in an oil change. Furthermore, a higher energy input leads to a decrease of m µ , consequently contributing to earlier performance loss of the lubricant. At energies exceeding 420 % and water contamination of more than 350 %, a negative m µ prevails, resulting in an audible shudder of the wet clutch system. Considering the oil sample with the highest energy input, this behavior emerges even at 0 % of water addition. To trace the arc from the energy input to the iron content, which results from the wear of various machine elements in the E-axle, Figure 5 shows the characteristic value m µ again depicted versus the water contamination. In contrast to Figure 4, the individual graphs are ordered and plotted according to the iron content ‘Fe’ determined by ICP-OES at the end of the test-rig or vehicle-based endu- Science and Research 25 Tribologie + Schmierungstechnik · volume 71 · issue 5-6/ 2024 DOI 10.24053/ TuS-2024-0035 Figure 3: Schematic test procedure to investigate the influences of water contamination on pre-deteriorated lubricant samples wards the metal surface of the steel disk through the forces of adsorption [20]. Given their substantial polarity, water molecules assume a position at the steel disk, thereby obstructing the dedicated additives from performing their intended task. Iron particles are subject to similar mechanisms in the lubricant and thus contribute to impeding additives from forming a tribo-layer between the respective clutch disks through catalytic effects [21]. This is underlined by the observed effects, as shown in Figure 4 and Figure 5. Since the course of the respective graphs does not seem to vary greatly regardless of the listing of energy input or iron content, a correlation of the lubricant-specific values (Energy and Fe - referring to the sample state prior to the tests in KLP-260) is shown in Figure 6. An approximately linear relation (R 2 = 0.7270) between energy input and iron content prevails, particularly in the range of low pre-damage of the lubricant. With increasing test duration, the scatter increases, which can be attributed to several reasons. Firstly, measurement uncertainties arise since the method of sample collection, and the period between sampling and ICP-OES analysis Science and Research 26 Tribologie + Schmierungstechnik · volume 71 · issue 5-6/ 2024 DOI 10.24053/ TuS-2024-0035 rance run before evaluation in the KLP-260. Essentially, a similar finding to the previous illustration emerges: an increased iron content exceeding 200 % in the lubricant leads to an earlier performance loss of the latter. Furthermore, a second finding can be derived from the two figures. The addition of water only seems to have an influence on the performance of the lubricant in the case of slight prior damage. In contrast, higher conditions of pre-damage (Energy > 300 %, Fe > 150 %) lead to an earlier lubricant deterioration and thus, shudder behavior of the wet clutch system even at low levels of water contamination. Therefore, it can be concluded that in an already pre-damaged state of the lubricant m µ does not further deteriorate through the addition of water. The changes in the friction behavior arise from the deterioration of the lubricant caused by the presence of water contamination in correlation with iron particles. Typically, the molecular makeup of additives such as friction modifiers, dispersants, or detergents comprises a polar head and a hydrocarbon tail. Under uncontaminated conditions, the polar head of friction modifiers is drawn to- Figure 4: Gradient of the CoF versus water contamination with respect to energy input Figure 5: Gradient of the CoF versus water contamination with respect to iron content significantly influence the measurement outcome as wear particles are deposited rapidly. However, the sampling within the scope of these investigations was kept constant allowing for improved comparability of the executed tests. Furthermore, differences e.g., in temperature levels of the test rig or vehicle-based endurance runs, have not yet been considered. Nevertheless, this regression enables abstracting and reducing the complexity of the mechanisms for future studies. Additionally, this visualization reveals a gap in the investigated oil samples between 200 % and 450 % energy input and highlights the potential for further investigations. Conclusion The present study focuses on the influences of energy input and iron particles with respect to water contamination of the lubricant by evaluating the friction behavior of wet clutches. Therefore, oil samples from previously performed test rig and vehicle-based endurance runs are tested in the KLP-260 component test rig. To quantify the performance loss of the lubricant, the characteristic value m µ , representing the gradient of the CoF versus the sliding speed, which serves as a performance indicator of the lubricant. The following conclusions can be drawn: • Due to physical adsorption and catalytic effects, water, and iron contamination significantly impair the performance of the lubricant. • Iron contents over 150 % and energy input exceeding 300 % lead to an earlier lubricant deterioration and, thus, shudder behavior of the wet clutch system. Above these values, additional water contamination does not further downgrade the performance of the lubricant. • A linear regression between energy input and iron content can be drawn to reduce the complexity of the systems approach. Acknowledgement This work was funded by the “Austrian COMET-Program” (project InTribology1, no. 872176) via the Austrian Research Promotion Agency (FFG) and the federal states of Niederösterreich and Vorarlberg. References [1] S. Stehle, M. Baese, K. Voelkel, K. Stahl, Determination of Friction and Shudder Behaviour of Slip-Controlled Wet Multi-Plate Clutches, Graz/ Spielberg, 2022. [2] J. Wirkner, M. Baese, A. Lebel, H. Pflaum, K. Voelkel, L. Pointner-Gabriel, C. Besser, T. Schneider, K. Stahl, Influence of Water Contamination, Iron Particles, and Energy Input on the NVH Behavior of Wet Clutches, Lubricants 11 (2023) 459. https: / / doi.org/ 10.3390/ lubricants11110459. [3] H. Naunheimer, B. Bertsche, J. Ryborz, Fahrzeuggetriebe: Grundlagen, Auswahl, Auslegung und Konstruktion, third. Auflage, 2019. [4] R. Maeki, B. Ganemi, E. Hoeglund, R. Olsson, Wet Clutch Transmission Fluid for AWD Differentials: Influence of Lubricant Additives on Friction characteristics, Lubr. Sci. 19 (2007) 87-99. https: / / doi.org/ 10.1002/ ls.33. [5] H. Spikes, The History and Mechanisms of ZDDP, Tribol Lett 17 (2004) 469-489. https: / / doi.org/ 10.1023/ B: TRIL.0000044495.26882.b5. [6] H. Ohtani, R.J. Hartley, D.W. Stinnett, Prediction of Anti- Shudder Properties of Automatic Transmission Fluids using a Modified SAE No. 2 Machine, SAE Ttransactions 103 (1994) 456-467. [7] N. Fatima, P. Marklund, R. Larsson, Water Contamination Effect in Wet Clutch System, Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering 227 (2013) 376-389. https: / / doi.org/ 10.1177/ 0954407012455145. [8] N. Fatima, Impact of Water Contamination and System Design on Wet Clutch Tribological Performance. Dissertation, Lulea, 2014. Science and Research 27 Tribologie + Schmierungstechnik · volume 71 · issue 5-6/ 2024 DOI 10.24053/ TuS-2024-0035 Figure 6: Linear regression of iron content and energy input [16] Y. Akita, K. Abe, Y. Osawa, Y. Goto, Y. Nagasawa, N. Sugiura, S. Wakamatsu, K. Kosaka, Analysis of Friction Coefficient Variation with Moisture between Friction Surfaces, SAE Technical Papers (2016). https: / / doi.org/ 10.4271/ 2016-01-0411. [17] A. Dorgham, A. Azam, A. Morina, A. Neville, On the Transient Decomposition and Reaction Kinetics of Zinc Dialkyldithiophosphate, ACS Applied Materials & Interfaces 10 (2018) 44803-44814. https: / / doi.org/ 10.1021/ acsami.8b08293. [18] G.J. Meingaßner, H. Pflaum, K. Stahl, Test-Rig Based Evaluation of Performance Data of Wet Disk Clutches, 14th International CTI Symposium (2015). [19] A. Agocs, A.L. Nagy, Z. Tabakov, J. Perger, J. Rohde- Brandenburger, M. Schandl, C. Besser, N. Dörr, Comprehensive Assessment of Oil Degradation Patterns in Petrol and Diesel Engines observed in a Field Test with Passenger Cars - Conventional Oil Analysis and Fuel Dilution, Tribology International 161 (2021) 107079. https: / / doi.org/ 10.1016/ j.triboint.2021.107079. [20] J. Crawford, A. Psaila, Miscellaneous Additives, in: R.M. Mortier, S.T. Orszulik (Eds.), Chemistry and Technology of Lubricants, first ed., Springer Science+Business Media, 1994, pp. 160 -173. [21] P. Nieuwland, T.A. Droste, Automatic Transmission Hydraulic System Cleanliness - The Effects of Operating Conditions, Measurement Techniques and High-Efficiency Filters, in: SAE Technical Paper Series, SAE International400 Commonwealth Drive, Warrendale, PA, United States, 2001. Science and Research 28 Tribologie + Schmierungstechnik · volume 71 · issue 5-6/ 2024 DOI 10.24053/ TuS-2024-0035 [9] N. Fatima, P. Marklund, A.P. Mathew, R. Larsson, Wet Clutch Friction Interfaces under Water-Contaminated Lubricant Conditions, Tribology Transactions 59 (2016) 441-450. https: / / doi.org/ 10.1080/ 10402004.2015.1081998. [10] H. Wang, Effects of Water in All-Wheel-Drive Clutch Oil on All-Wheel-Drive Cornering Noise, SAE Technical Papers (2022). https: / / doi.org/ 10.4271/ 2022-01-5103. [11] M. Baese, A. Lebel, R. Franz, J. Prost, Analysis of Wet Vehicle Clutch Damage Mechanism caused by Water, Tribologie und Schmierungstechnik 69 (2022) 7-13. https: / / doi.org/ 10.24053/ TuS-2022-0008. [12] W. Williamson, B. Rhodillegible, The Effects of Water on Cellulose-based Frictional Surfaces in Automatic Transmission Clutch Plates, SAE Technical Papers (1996). https: / / doi.org/ 10.4271/ 961917. [13] K. Berglund, P. Marklund, R. Larsson, M. Pach, R. Olsson, Wet Clutch Degradation Monitored by Lubricant Analysis, in: SAE Technical Paper Series, SAE International400 Commonwealth Drive, Warrendale, PA, United States, 2010. [14] A. Dorgham, A. Azam, P. Parsaeian, T. Khan, M. Sleiman, C. Wang, A. Morina, A. Neville, Understanding the Effect of Water on the Transient Decomposition of Zinc Dialkyldithiophosphate (ZDDP), Tribology International 157 (2021). https: / / doi.org/ 10.1016/ j.triboint.2021.106855. [15] M. Yokomizo, T. Iwai, K. Narita, M. Kudo, Lubricants Formulation Technology for Fuel Saving Performance in Automatic Transmissions, SAE International, 2015.