Introduction Axial piston pumps are hydraulic displacement units that are widely used in both industrial and mobile applications due to their high power density and compact design / Mat14/ . In an axial piston pump, there are many tribological contacts. This paper concentrates on the tribological contact between cylinder block and valve plate. The contact has several similarities with axial thrust bearings but due to the valve geometries there are also other challenges to be dealt with. During operation, the cylinder block is preloaded against the valve plate to reduce leakage. The pressurization of the pistons and load shocks from the engine generate additional axial loads. A widely used type of axial piston pump is the swash plate machine. A detailed view of the components of this sub-type is shown in Figure 1. There is a combination of hydrodynamic and hydrostatic effects in the contact between the cylinder block and the valve plate. The valve plate is hydrostatically balanced by pressurized oil, creating a thin oil film between the surfaces. However, the resulting axial load is dependent on the piston position, which leads to a periodic pulsation of the gap height. In addition, a hydrodynamic lubricating film builds up during operation. Due to changes in load and rotational speeds, the balancing of the contact will usually not be ideal in all states of operation, resulting in the hydrodynamic lubricating film failing at low speeds and mixed friction occurring. / Sei08/ . In this pa- Science and Research 5 Tribologie + Schmierungstechnik · volume 71 · issue 1/ 2024 DOI 10.24053/ TuS-2024-0002 Experimental development and validation of tribological run-in strategies to reduce friction in hydraulic applications Zita Tappeiner, Achill Holzer, Katharina Schmitz * submitted: 20.09.2023 accepted: 22.01.2024 (peer-review) Presented at the GfT Conference 2023 Das Einlaufen eines tribologischen Kontakts hängt von der aufgebrachten Normallast, der Gleitgeschwindigkeit und der Einlaufdauer ab. Um Axialkolbenpumpen effektiv einzulaufen, suchen die Hersteller nach optimalen Einlaufparametern. In diesem Paper wird eine Methodik entwickelt, um geeignete Einlaufparameter zu bestimmen. Die Parameter wurden an einem Scheibe-Scheibe-Tribometer validiert und bestätigt. Ziel der Untersuchung ist die Minderung von Reibung und Verschleiß im späteren Pumpenbetrieb. Schlüsselwörter Axialkolbenpumpe, Hydraulik, Einlaufen, Stribeck- Kurve, Tribometer, Scheibenkontakt A particularly demanding contact in axial piston pumps is the one between cylinder block and valve plate. The tribological behavior of the contact can be changed by a run-in process. Publications on fast and efficient run-in are rare as this knowledge is often considered confidential. For this paper, tests have been carried out on a disc-on-disc tribometer to examine the run-in behavior of a material pairing and to identify suitable parameters for run-in. A methodology has been developed which can be used to find the optimal normal load and rotational speed for the runin process to shorten the run-in time significantly. Keywords Axial Piston Pump, Disc-On-Disc, Hydraulics, Run- In, Stribeck Curve, Tribometer Kurzfassung Abstract * Zita Tappeiner, M.Sc. Orcid-ID: https: / / orcid.org/ 0009-0008-7442-5809 Achill Holzer, M.Sc. Orcid-ID: https: / / orcid.org/ 0000-0003-1190-1819 Univ.-Prof. Dr.-Ing. Katharina Schmitz Orcid-ID: https: / / orcid.org/ 0000-0002-1454-8267 Institut für fluidtechnische Antriebe und Systeme der RWTH Aachen University Campus-Boulevard 30, 52074 Aachen on the influence of the normal force and the rotational speed occurring during operation. In addition, it is to be determined how the friction and wear during operation can be reduced using different run-in parameters. In tribology, the term “run-in” is used to describe the processes that occur during the initial encounter of two friction partners until a steady-state friction condition is reached. These processes are accompanied by a change in the coefficient of friction and/ or the wear rate as a function of the run-in time, the number of friction cycles or the friction distance. Run-in is used specifically for surface conditioning in tribosystems / Bla91/ . Surface conditioning is achieved by changing the mechanical, chemical and physical properties as well as the contact geometry of the friction partners. This process takes place on the surface and in the microstructure near the surface / Fes14/ . The three most important characteristics of friction transitions are the change in the coefficient of friction, the time required for a tribosystem to reach a steady state (or other distinctive state such as galling), and the characteristics during short-term fluctuations in the frictional load. Blau identified common shapes of run-in curves / Bla08/ . Figure 2 shows the variation of the frictional force typically occurring in pure metal contacts. The shape shown is called form (f) by Blau. The decrease in frictional load occurs due to changes in the near-surface layers and the smoothing of the surface. Overall, four processes can be identified that constitute run-in: Material transfer, film Science and Research 6 Tribologie + Schmierungstechnik · volume 71 · issue 1/ 2024 DOI 10.24053/ TuS-2024-0002 per the tribological behavior of the contact between cylinder block and valve plate is examined. Its focus lies Figure 1: Axial piston pump in swash plate design Figure 2: Friction behavior during the run-in of metal contacts according to / Bla08/ Figure 3: Combined effect of load and speed on the surface roughness variation / Akb13/ formation or removal, deposition, and cyclic surface deterioration. The curve shape shown cannot be clearly assigned to mechanisms, and similar curve shapes can be caused by different boundary processes / Bla08/ . The parameters selected during run-in affect the performance of the tribological system in operation. Accordingly, the operating conditions during the run-in phase should be carefully selected. Optimizing the run-in process can lead to an extension of a tribo system’s service life and stable operation. Design, surface mechanics, chemistry and materials play important roles in optimization / Kho21/ . Akbarzadeh and Khonsari determined the operating conditions that result in the minimum value of the surface roughness R a at the end of the run-in period. Figure 3 shows the effect of the combination of load and speed on the change in surface roughness / Akb13/ . It is noticeable that the operating conditions for the maximum change in surface roughness are very close to operating conditions that cause a significantly smaller change in surface roughness. In addition to load and speed, the initial surface roughness and the initial friction coefficient, which in turn depends on load and speed, are also decisive / Kho21/ . Methods Tribometer Model Test Rig Figure 4 shows the setup of the disc-on-disc tribometer test rig (tribometer) also known as Siebel-Kehl tribometer / Mur10/ , used for the experimental part of this work. The two ring-shaped test specimens “stator” and “rotor”, which are pressed onto each other with a defined load by the hydraulic contact cylinder, are the central elements of the tribometer. A load cell between the cylinder and the moment support is used to measure the contact pressure. A hydraulic motor generates the necessary relative movement of the rotor. The frictional torque between the rotor and stator is determined on the stationary, upper disc via a force sensor connected to a lever. The test specimens are selected according to the material pairing in axial piston pumps. The stator is made of heat-treated steel (1.7225) and the rotor of the special brass (2.0550). A mineral oil (Renolin B 15 VG 46) containing zinc and ash of viscosity class 46 was used for the tests / Fuc24/ . The fluid contains a zinc dithiophosphate as extreme pressure and antiwear additive. All test runs were repeated with three specimens of the same material combination. After each test run, the surface of the discs was renewed using wet sandpaper with a final grit of 1200. The test bench can reach rotational speeds of up to 1600 min -1 . With an average diameter of the contact surface of 27 mm, an average speed of 4.5 m/ s is achieved. Loads of up to 3500 N were used, inducing surface pressures of up to 4 MPa. In a load analysis on an axial piston pump, Wegner determined the surface pressure at the control surface to be 4 MPa. / Weg21/ Thus, the parameters used are within the range of values that occur in the hydraulic application. Experimental approach In a first approach, the tribological contact was examined under constant conditions. In a test run lasting one hour, temperature and frictional torque in the contact between the discs were measured. The test parameters are listed in Table 1. Science and Research 7 Tribologie + Schmierungstechnik · volume 71 · issue 1/ 2024 DOI 10.24053/ TuS-2024-0002 Figure 4: Tribometer test rig Oil temperature 40 °C Test duration 60 min Rotational speed 1600 rpm Normal load 1000 N Table 1: Test parameters for constant tests Subsequently, three tests were carried out in which either the normal load or the rotational speed were changed step by step. This procedure was chosen in order to be able to compare the measured friction coefficients of At 1600 min -1 , the normal load was adjusted to 1000 N. Finally, the Stribeck curve was measured following a speed ramp (duration: t Stribeck . The following values for the run-in time t run-in were tested: 0 min, 1 min, 3 min, 5 min, 10 min and 30 min. These tests were carried out for the normal loads F N,run-in 1000 N, 2000 N and 2500 N. Results Figure 6 shows the results for 1000 N normal load (corresponding to a surface pressure of about 1.2 MPa). The coefficient of friction drops at the beginning of the measurement and remains constant. The coefficient of friction of pairing no. 3 is slightly lower overall and shows hardly any fluctuations over the entire period. The mean value for the coefficient of friction is 0.0322 at the beginning. Within one hour at constant test conditions, the mean coefficient of friction drops to 0.0300. This corresponds to a 7 % decrease in the coefficient of friction. A comparison of the different normal forces shows that the measured coefficient of friction decreases as the contact pressure increases. This contrasts with the findings of Popov / Pop15/ . In Bollók’s investigations with a steelbronze pairing, however, the same dependence can be Science and Research 8 Tribologie + Schmierungstechnik · volume 71 · issue 1/ 2024 DOI 10.24053/ TuS-2024-0002 two stages with the same test parameters. In these three tests, the influence of normal load and speed was investigated iteratively. An overview of these tests can be found in Table 2. Test parameters Test 1 Test 2 Test 3 Oil temperature 40 °C 40 °C 40 °C Holding time per step 5 min 5 min 5 min Rotational speed 1600 rpm variable 250 rpm Normal load variable 2000 N variable Step-size 250 N 125 rpm 250 N Initial value 250 N 1500 rpm 250 N Maximum value 3500 N 1500 rpm 3500 N Table 2: Test parameters for variable tests In a final series of tests, specific run-in conditions were compared. To compare the quality of the run-in process, the same Stribeck curve was repeated each time. The test load and speed procedure are shown in Figure 5. In the beginning of the measurement, the speed was increased to 250 min -1 with a defined slope. After that, the normal load was set according to the run-in test. The run-in then took place at constant conditions for the time t run-in . Hereafter, the speed was increased to 1600 min -1 . Figure 6: Characterization of the tribological contact Figure 5: Test conditions for Stribeck tests observed. As the normal force increases, the coefficient of friction decreases / Bol06/ . Figure 7 shows the change in the disc surface due to the run-in process. Abrasive running marks occur in the contact area with the other disc. In some cases, even deep ridges are formed, which are either due to roughness peaks of the counter body or particle contamination in the fluid. No adhesion is recognizable in the surface after run-in. A detailed examination of the surface roughness was not part of this work but could provide further information on the run-in condition of the components. The measurement results in Figure 8 show that the pairings tolerate a normal load of 3500 N, as no fretting occurs in pairings 1 and 3 under these test conditions. In pairing 2, on the other hand, a sharp increase in the coefficient of friction and, analogously, in the temperature already occurs at 1750 N. The reason for this is unclear, since in a preliminary study pairing 2 withstood a normal load up to 3000 N without an increase in the coefficient of friction. Thus, pairing 2 is considered as an outlier here. The other two pairings exhibit a different, very similar behavior to each other. In both pairings, the coefficient of friction decreases as the normal load increases. At the beginning of each new load level, the coefficient of friction increases slightly and then decreases again shortly afterwards. This is an indication that run-in takes place at each step. After the first half of the gradational progress has been run through and the load decreases again, these increases no longer occur. Overall, the values of the coefficient of friction as well as the temperature on the right-hand side of the gradational progress are lower than on the left-hand side. In order to obtain a quantitative measure for the run-in, a run-in quotient μ after / μ before is introduced. The run-in quotient is formed according to the illustration in Figure 8. For each load level, the friction coefficient is averaged. Then, for each load step, the mean value μ after from the right side of the stair is divided by the corresponding mean value μ before from the left side. Thus, a quotient of the improvement of the friction coefficient can be assigned to each value of the normal load. The graphical representation shows that the coefficient of friction decreases for almost all load levels over the duration of the test. At this point, it should be noted that the time difference between the left and right sides of the gradational progress is not constant for the individual load steps. Due to the stepwise test methodology, the time difference and thus the run-in time is significantly larger for the lower values of F N . Nevertheless, an improvement in the coefficient of friction can be observed, especially for the Science and Research 9 Tribologie + Schmierungstechnik · volume 71 · issue 1/ 2024 DOI 10.24053/ TuS-2024-0002 Figure 8: Influence of the normal force on the run-in behavior Figure 7: Change in the rotor surface due to run-in difficult to evaluate. Nevertheless, it is noticeable that despite the fretting, there was an improvement in the coefficient of friction. At this point, it is important to note that the improvement is only evaluated against the same normal load. The following experimental investigations are evaluated analogously to the procedure in Figure 8. The method can be applied not only to the tests with varying normal loads, but also with varying speeds. The aim here is to find the optimal parameters for the run-in of the contact. Since the run-in process depends on a combination of the parameters speed and load, this investigation must be carried out iteratively. The plots of the run-in quotients of experiment 2 and 3 in Figure 9 show clear minima at 250 rpm (0,71 m/ s) and between 2500 and 3000 N respectively. To verify an Science and Research 10 Tribologie + Schmierungstechnik · volume 71 · issue 1/ 2024 DOI 10.24053/ TuS-2024-0002 higher normal loads. Since the quotient is formed from μ after / μ before , the greater the change, the smaller the value of the quotient. Thus, a smaller value of the run-in quotient is an indication of a greater improvement in the coefficient of friction. Values higher than 1 indicate an increase in the coefficient of friction and thus a worsening of the friction behavior. This is the case with the lower load values of pairing 3. Why a deterioration of the friction condition occurred in this case, cannot be explained at this point. For other normal loads and pairings, a general decrease in the coefficient of friction can be observed. It is interesting to note that the progressions of the curves of pairings 1 and 2 are very similar. Both have a global minimum at about 2500 N. Apparently, run-in under a normal load from this range of values leads to a stronger improvement of the frictional properties than a run-in at values around 1000 N. Since fretting occurred in pairing 2, the values of the run-in quotients are more Figure 10: Validation of run-in parameters Figure 9: Run-in quotients of Test 2 (left) and Test 3 (right) improvement in the run-in, validation was carried out with Stribeck tests. The results of the validation are shown in Figure 10 by the normalized friction coefficient. For better comparison the friction coefficient is normalized with the maximum friction coefficient μ ref . Evidently, the parameters for major changes in the coefficient of friction determined by means of the run-in quotient do indeed lead to better run-in results. In particular, the extreme reduction of the run-in time from run-in at unfavorable conditions (1000 N) to run-in at more favorable conditions (2000 N and 2500 N) poses many economic and ecological advantages. The final validation is shown in Figure 11. The Stribeck test was carried out at different normal loads after a runin time of three minutes in each case. A further increase in the normal force does not lead to an improvement in the run-in process, but to higher coefficients of friction than the running-in at 2000 N or 2500 N. This means that an optimum can indeed be determined by the developed methodology. Discussion A general methodology for identifying suitable run-in parameters for a tribological contact can be derived from the procedure determined in this work empirically. Since a tribological system depends on many factors, such as geometry, material, lubricant and temperature, the runin parameters developed during the experimental investigation cannot simply be transferred to any tribological system. However, the findings developed here can be used to characterize a new tribological system. An illustration of the procedure followed in this investigation is shown in Figure 12. Starting from a variation of the normal load at an initially randomly selected speed (Test 1), a normal load suitable for the run-in was determined. With the determined load, the speed achieving the greatest run-in effect was sought (Test 2). The initial test was then repeated at the new speed (Test 3). Thus, three trials with two transmissions (I and II) took place. Following this procedure, suitable run-in parameters can already be identified. However, it is advisable to supple- Science and Research 11 Tribologie + Schmierungstechnik · volume 71 · issue 1/ 2024 DOI 10.24053/ TuS-2024-0002 Figure 12: Method for creating a run-in map Figure 11: Validation of the influence of the normal load on the friction coefficient and rotational speed can be determined for any tribological contact. This is particularly useful for new material pairings in order to improve the friction coefficient and reduce the run-in time. References / Akb13/ Akbarzadeh, S., Khonsari, M. M. On the optimization of running-in operating conditions in applications involving EHL line contact, Wear, Vol. 303, 1-2, S. 130-137, 2013. / Bau20/ Bauer, G., Niebergall, M. Ölhydraulik - Grundlagen, Bauelemente, Anwendungen, Lehrbuch, 12., neu bearbeitete Auflage, Springer Vieweg, Wiesbaden, Heidelberg, 2020. / Bla91/ Blau, P. J. Running-in: Art or engineering? Journal of Materials Engineering, Vol. 13, Nr. 1, S. 47-53, 1991. / Bla08/ Blau, P. J. Friction science and technology - From concepts to application, Dekker Mechanical Engineering, 2nd ed., CRC Press, Boca Raton, London, 2008. / Bol06/ Bollók, P., Kozma, M. Comparison of surface layers developed during sliding friction of metal pairs, International Colloquium Tribology, 2006. / Fes14/ Feser, T. Untersuchungen zum Einlaufverhalten binärer alpha-Messinglegierungen unter Ölschmierung in Abhängigkeit des Zinkgehaltes, Schriftenreihe des Instituts für Angewandte Materialien, Karlsruher Institut für Technologie, Print on demand, KIT Scientific Publishing, Karlsruhe, 2014. / Fuc24/ RENOLIN B 15 VG 46, https: / / www.fuchs.com / de/ de/ produkt/ product/ 149041-RENOLIN-B-15- VG-46/ (abgerufen am 22.01.2024), 2024. / Kho21/ Khonsari, M. M., Ghatrehsamani, S., Akbarzadeh, S. On the running-in nature of metallic tribo-components: A review, Wear, 474-475, S. 203871, 2021. / Mat14/ Matthies, H. J., Renius, K. T. Einführung in die Ölhydraulik - Für Studium und Praxis, 8. Auflage, Springer Vieweg, Wiesbaden, 2014. / Mur10/ Murrenhoff, H. Umweltverträgliche Tribosysteme - Die Vision einer umweltfreundlichen Werkzeugmaschine, Springer-Verlag Berlin Heidelberg, Berlin, Heidelberg, 2010. / Pau17/ Paulus, A. Reaktionsschichtbildung auf bleifreien Bronze- und Messingwerkstoffen im Kontakt von Zylinder und Steuerscheibe einer Axialkolbenpumpe, 2017. / Pop15/ Popov, V. L. Kontaktmechanik und Reibung - Von der Nanotribologie bis zur Erdbebendynamik, 3., aktualisierte Auflage, Springer Vieweg, Berlin, Heidelberg, 2015. / Sei08/ Seifert, V., Alaze, N. et al. HYDRANO - Leistungssteigerung hydraulischer Verdrängereinheiten durch Nanocomposites - Schlussbericht. Projektlaufzeit: 01.06.2005 bis 30.11.2008, S. 1-236, 2008. / Weg21/ Wegner, S. Experimental and Simulative Investigation of the Cylinder Block/ Valve Plate Contact in Axial Piston Machines, Reihe Fluidtechnik D, Shaker, Aachen, 2021. Science and Research 12 Tribologie + Schmierungstechnik · volume 71 · issue 1/ 2024 DOI 10.24053/ TuS-2024-0002 ment these tests with a further test. This is shown in black in Figure 12 and forms the fourth edge, which is necessary to create a run-in map. In this way, a graphical representation of the inlet quotient could be developed analogous to the change in surface roughness in Figure 3. This map could be supplemented as required in further tests, resulting in an increasingly accurate characterization of the run-in behavior of the tribological system. The method developed here can be applied to discs of different materials using different oils at any oil temperature on the tribometer used in this work. With the aid of the method developed here, it is then possible to determine run-in parameters that can be used for all similar applications. Thus, the often very time-consuming and energy-intensive run-in processes in industry could be significantly optimized. Conclusion In this work, run-in examinations have been carried out to improve run-in behavior of tribological contacts by identifying optimal parameters for normal load and rotational speed. With the identified parameters both the friction coefficient could be decreased and the run-in period shortened. The influence of speed and normal load on the run-in process was analyzed using a gradational progress on a tribometer model test rig. In order to measure the run-in progress, a run-in quotient was introduced, as a measure of the change in the coefficient of friction before and after the run-in process. A comparison of the parameters rotational speed and normal load shows the influence of normal load on the run-in behavior being significantly higher than the influence of rotational speed. The developed methodology was validated in further experiments. Run-in tests at various normal loads were carried out at the optimum run-in speed. These confirmed that run-in at the optimum normal load determined following the methodology leads to a more significant improvement in friction behavior than run-in at other normal loads. A 7 % decrease in the coefficient of friction at the low point of the Stribeck curve was achieved. In addition, with the correct choice of run-in parameters, the duration of run-in could be significantly reduced. Using the optimal parameters, the run-in can be completed within 3 min. With this finding, significant time savings are possible compared to the usual run-in period in practice. Following the scheme of the methodology developed in this work, a suitable set of parameters for normal load \ Gesundheit \ schaft \ Linguisti schaft \ Slawisti \ Sport \ Gesun wissenschaft \ L wissenschaft \ philologie \ Spo Fremdsprachend \ VWL \ Maschi schaften \ Sozi Bauwesen \ Fre