Tribologie + Schmierungstechnik 63. Jahrgang 4/ 2016 loads will be repeated until the resulting friction moments will have mostly stabilised at all speeds. The biggest change due to running in is seen at shaft speeds less than 400 rpm. The speed of transition in the Stribeck diagram is also shifted towards 400 rpm, with respective changes depending on the pairing of materials in mixed friction. Within the first few hours of the test, the running-in process starts very quickly (figure 6) and getting to a nearly steady state after about 100 h. After the running in process is finished a Stribeck diagram is recorded for speeds from 200 rpm to 1600 rpm. 2.3 Results Three different friction pairings will be used to show the influence of roughness, hardness and friction coefficient of the test shafts on the running in process and friction losses over time. Using the same bearing pedestals from cast AlSi9Cu3 material with a bore finished to Rz 3.2 the following set of steel E355 shafts have been tested: • ground to Rz 2 (“reference”), ca. 150HBW • ground and super-finished to Rz 0.5 • ground to Rz 2, DLC coated with Graphit-IC™, 355 HV0,5 Figure 7 shows the Stribeck-diagram of the three shafts at begin of the runningin process. Both, the fine finish and particularly the DLC coated shaft show a distinctive advantage compared to the reference model. The reduced roughness on the super-finished shaft reduces elastic/ plastic and abrasive friction and allows a smaller effective oil film height, improving the change from mixed friction to hydrodynamic running conditions. The advantage of the coated shaft can be assumed to be a result from the lower friction coefficient of the pairing DLC-AlSi9Cu3 in new condition, as the roughness itself is hardly changed by the coating. Due to wear during the running-in time the amount of mixed friction changes significantly, as can be seen in figure 8 for a shaft speed of 250 rpm. Both tests with the uncoated shafts show a much stronger reduction in overall friction compared to using the coated shaft within the first 24 hours. After this comparatively short time the initial advantage of the DLC coated shaft is already lost, and the difference between ground and superfinished shaft is strongly reduced. For each material pairing two sets of samples have been examined. Figure 9 shows Stribeck diagrams of the finished tests, averaging over results from identical configurations. The results for each set of tests had a deviation between 0.001and 0.007 Nm, showing the excellent reproducibility of the experiments. The differences between the friction moments after running in are close to being insignificant in some areas. But the conclusion that both noncoated shafts result in a virtually identical amount of friction, with the indication of being better than the DLC coated shaft, is slightly surprising. 2.4 Analysis of Test Samples Dimensions, roughness and tolerances in shape and position of all test samples have been exactly measured before and after the experiments. This allows to compare the mechanical effects of the runningin wear on the dif- 17 Aus Wissenschaft und Forschung Figure 7: Stribeck diagram taken at the start of the running-in process Figure 8: Friction moment at 250 rpm during runningin Figure 9: Stribeck diagram after completion of running-in process T+S_4_16 02.06.16 12: 26 Seite 17 18 Tribologie + Schmierungstechnik 63. Jahrgang 4/ 2016 ferent surfaces with the initial condition. Figure 10 shows confocal microscopy scans of the surfaces in new condition. Superfinishing changes several surface characteristics compared to the original ground condition: All bearing pedestals have the same surface finish with Rz 3.2 before the experiments (figure 11). Studying the surfaces of the shafts after completion of the tests, there are more or less distinctive smoothed areas showing at the bearing positions. Due to the constant load of the simulated camshaft drive the wear in bearing 1 is the most pronounced. The transient cam forces result in smooth patches only in certain parts of the circumference. A close analysis reveals that the ground reference shaft is significantly smoothed, while the super-finished one retains the already low roughness with occasional gauging due to particles. The DLC coated shaft shows significantly less smoothing (s. figure 12). This is a good indication for the wear resistance of the coating, but hinders smoothing effects, and as such effectively the reduced friction that was expected. Creating a finer finished surface before applying Aus Wissenschaft und Forschung Figure 10: Shaft surfaces before testing Figure 11: Exemplary bearing surface of pedestal before test Figure 12: Shaft surfaces at position of bearing 1 after test Reference Rz 2 Superfinish Rz 0,5 DLC-Coated Rz 2 Reference Superfinish DLC-Coated ground super-finished Rpk 0.1141 µm 0.0505 µm Rvk 0.3931 µm 0.1271 µm Rk 0.5270 µm 0.1600 µm Rmr (0.5µm) 82.31% 99.94% T+S_4_16 02.06.16 12: 26 Seite 18 Tribologie + Schmierungstechnik 63. Jahrgang 4/ 2016 the coating would be a possible option, but cause a further increase in costs. The analysis of the bearing surfaces in the pedestals shows an overall similar result, where in addition to opened porosities and gauges the smoothing in some areas was so significant, that the original structure of the machined surface is not visible any more (s. figure13). These effects are characteristic and have been found in all three material pairings to a similar extent. Primary profiles across a large selection of worn surfaces show a typical crowning in the scale of 1 µm near the edges. Results from further experiments with different approaches to friction reduction, e. g. nitrated shafts, PEEKcoated shafts or PTFE bushings, all support the conclusion that for similar initial roughness smoothing and a modified micro-contour through running-in is most important to reach a low friction induced driving moment. If the wear through the running-in process is limited to one body, e. g. by using very hard surface layers (DLC, nitrated shaft) or particularly distinct hard/ soft combinations (PTFE/ Steel or PEEK/ AlSi9Cu3), the initially better friction characteristics cannot be kept up compared to Steel/ AlSi9Cu3. In conclusion the wear through running-in is at least as significant in search for friction-optimised camshaft bearings, as is the initial friction coefficient and individual tribo-chemical processes, which currently are under investigation. 3 Validation of Numerical Simulation using the Experimental Results As has been seen from the results of the experiments, a suitable wear simulation model will be essential for the application and validity of EHD-simulations for camshafts and similar configurations with significant mixed friction and edge loads, validity of computational results in cases less dominated by wear has been shown before, e. g. in [3]. The currently implemented wear model allows wear to appear only on one of the surfaces in a bearing. For this reason the test with super-finished steel shaft was chosen as reference for the numerical simulation. As was shown above, the wear was mostly limited to the bearings in this case. The simulation model consists of an elastic shaft and the four elastic bearing pedestals. The stiff load transmission including the roller bearings can safely be ignored and the loads added directly onto the camshaft. All four bearings are modelled as EHD-bearings, but only the bearings No. 1, 2 and 3 are subjected to wear. No mixed friction was detected in bearing 4 during simulations at any speed from 200 to 1600 rpm. To keep the amount of necessary simulations at a minimum, the following sequence was defined to simulate the full running-in process: 1. Wear simulation at a speed of 500 rpm until the wear algorithm converges. The approximately linear section of the Stribeck diagram for speeds above 500 rpm after running-in suggests that mixed friction is insignificant at these speeds. 2. Further wear simulations will be run at the lower speed limit of 200 rpm while monitoring the Stribeck diagram. The measured data for this speed indicate that wear has not fully completed, as the moment is still falling when the tests were finished (compare figure 8) Further simplifications have been introduced into hydrodynamic bearings and the mixed friction. The flow factors for the hydrodynamics have been calculated using the machined surface of the bearings when new, the surface of the shaft was considered smooth due to the very low roughness in comparison. The contact pressure characteristics are calculated using the statistical model by Greenwood Tripp and the average roughness extracted from measured data. Depending in the average wear depth on a bearing, the roughness is reduced accordingly, so that the gap height in the bearing can be reduced before mixed friction starts. In figure 14 the calculated total friction moment for the speed of 500 rpm is plotted in comparison to the experimental data. As the wear simulation does not calculate 19 Aus Wissenschaft und Forschung Figure 13: Smoothed area in bearing 1 after test, right: primary profile across full width T+S_4_16 02.06.16 12: 26 Seite 19 20 Tribologie + Schmierungstechnik 63. Jahrgang 4/ 2016 Aus Wissenschaft und Forschung Figure 15: Stribeck diagram of the experiment after running in (black crosses) and different steps of the wear simulation between reaching convergence for 500 rpm (step 25, green line) and total stop of simulated wear (step 50, orange line). The arrow indicates the progression of wear at 200 rpm Figure 16: Comparison of calculated wear area in bearing 1 compared to a photograph showing the smoothed region Figure 17: Comparison of profile across wear in bearing 2 between simulation (top) and experiment (bottom) fixed time steps, but determines the time that is necessary to achieve a defined increase of wear at each step, the steps at the begin of the simulation appear in quick order. Even though the apparent frictional energy density e R* was not known exactly for the materials involved, and the regime in the experiment involved changing rotational speeds, it can be seen that the calculated running-in time is close to reality. The slightly lower overall moment compared to the experiment is partly due to being limited to a constant roughness per bearing. To slow down wear the roughness needs to be reduced, but this also affects friction losses in the remaining area of the bearing. The idealised conditions of the simulation should generally deliver less friction compared to reality anyway. After switching to the lower speed of 200 rpm after 25 steps with 500 rpm the wear simulation was continued. Stribeck diagrams have been plotted at each step, to compare these with the experimental data. Like before, the reduction of friction at speeds between 200 and 500 rpm is very quick initially. figure 15 shows a good agreement between the measured data at the end of the test and the simulation step 29 (light blue line). Continuing the wear simulation will eventually result in a state where mixed friction is not significant any more and no further wear occurs. The Stribeck diagram is showing a nearly proportional increase of friction moment with shaft speed at this point at simulation step 50. No direct measurement of the wear volume was taken, but a comparison of the additional contour due to wear at simulation step 29 with a photograph of the smoothed area in bearing 1 is in very good agreement in both size and position (s. figure 16). Similarly the primary roughness plot of bearing 2 has a good analogy to the calculated wear depth along the same line (s. figure 17). Figure 14: Friction moment from wear simulation at 500 rpm compared to data at 500 rpm extracted from running-in test T+S_4_16 02.06.16 12: 26 Seite 20 Tribologie + Schmierungstechnik 63. Jahrgang 4/ 2016 4 Conclusion The experimental study of surface treatments and coatings in camshaft journal bearings has confirmed, that it is not enough to improve single factors like coefficient of friction, hardness or surface roughness of newly manufactured parts. Rather, a good pairing and interplay of the various tribological properties of both shaft and journal is required for optimal friction reduction during running-in. Complex tribo-chemical processes in the boundary layers and their effect on boundary friction won’t eliminate the need for experimental tests in the near future. Nevertheless, the implementation of a wear model is a very important measure for increasing the conclusiveness of EHD-simulations for similar applications. For the numerical simulation of wear using the software FIRST the experiments show that realistic calculations of friction reduction due to running-in can be achieved. Wear depth, location and running-in times also show favourable agreements. The results also give hints for future developments to increase robustness of application and results of the simulation process, i. e. for more efficient simulation of wear at different engine speeds. References [1] Patir, N. und Cheng, H.S.: An Average Flow Model for Determening Effects of Threedimensional Roughness on Partial Hydrodynamic Lubrication. Transactions of the ASME, Series F, Journal of Lubrication Technology, Vol.100, 1978, S. 12-17 [2] Knoll, G., Lang, J., Winijsart, A., Umbach, S., Wolf, C.: EHD/ MKS-Simulationstechnik mit integriertem energetischen Verschleißmodell - Fallstudie dynamisch belastete Motorengleitlager [3] Tuzcu, S. ; Knoll, G. ; Meusel, J. ; Muller, T.: Tribologische Optimierung einer Leichtbaunockenwelle durch EHD-Simulation und deren experimentelle Validierung; VDI BE- RICHTE; 2115; 37-48; VDI-Verlag, Düsseldorf, 2010 21 Aus Wissenschaft und Forschung Bestellcoupon Tribologie und Schmierungstechnik „Richtungsweisende Informationen aus Forschung und Entwicklung“ Getriebeschmierung - Motorenschmierung - Schmierfette und Schmierstoffe - Kühlschmierstoffe - Schmierung in der Umformtechnik - Tribologisches Verhalten von Werkstoffen - Minimalmengenschmierung - Gebrauchtölanalyse - Mikro- und Nanotribologie - Ökologische Aspekte der Schmierstoffe - Tribologische Prüfverfahren Bestellcoupon Ich möchte Tribologie und Schmierungstechnik näher kennen lernen. Bitte liefern Sie mir ein Probeabonnement (2 Ausgaben), zum Vorzugspreis von 7 39,-. So kann ich die Zeitschrift in Ruhe prüfen. Wenn Sie dann nichts von mir hören, möchte ich Tribologie und Schmierungstechnik weiter beziehen. Zum jährlichen Abo-Preis incl. Versand von 7 189,- Inland (incl. MwSt.) bzw. 7 198,- Ausland. (In der EU bei fehlender UID-Nr. zzgl. MwSt.). Die Rechnungsstellung erfolgt dann jährlich. Das Jahresabonnement ist für ein Jahr gültig; die Kündigungsfrist beträgt sechs Wochen zum Jahresende. Firma, Abteilung Straße, Nr. Name, Vorname PLZ, Ort Ort/ Datum, Unterschrift: (ggf. Firmenstempel) Coupon an: expert verlag, Abonnenten-Service, Postfach 2020, 71268 Renningen oder per Fax an: (0 71 59) 92 65-20 T+S_4_16 02.06.16 12: 26 Seite 21