1 Introduction Fretting wear can be an ambiguous case to describe in general. According to Heinz [1,3] fretting wear occurs in a contact with an oscillational movement with amplitudes below 200-300 µm. The movement amplitude is smaller than the contact length in the direction of movement. This results in an engagement or contact ratio close to or equal to 1 for both friction partners. Analogous consideration also applies for rotationally symmetrical systems. Vingsbo and Söderberg [7] include the contact pressure into fretting wear maps to better describe the conditions within the tribological system. The terms of stick, partial slip and gross slip do link to wear mechanisms like surface cracks, abrasion and adhesion in applications. The systematic approach of Czichos and Habig [4] can also help to define fretting to a topic where contact size, oscillating motion amplitudes and elastic deformation around the contact area matter. The time scale can be quite wide for thermal expansion cycles on the low end to vibrating machinery harmonics e.g. from piston engines in the kHz range. Aus Wissenschaft und Forschung 15 Tribologie + Schmierungstechnik · 68. Jahrgang · 1/ 2021 DOI 10.24053/ TuS-2021-0003 Fretting Wear Observation of Flat Contacts Henrik Buse, Erika Hodúlová* Mit Hilfe eines speziell für flächige Kontakte entwickelten Modell-Prüfstands sollen neue und einfache Methoden der Material- und Schmierstoffprüfung auf Grundlagen- und Anwendungsebene unter Schwingverschleißbedingungen aufgezeigt werden. Auf die übliche Abstraktion der Kontaktgeometrien mit einfach auszurichtenden Punkt- oder Linienkontakten wird verzichtet, um den häufig flächigen Kontakten in angewandten Tribosystemen gerecht zu werden. Im Speziellen des Beitrags wird eine Methode untersucht eine gezielte Beobachtung der Oberflächen zu evaluieren. Die Methode nutzt eine verfahrbare Oberprobe um den Kontakt zu öffnen und einen Zwischenstand des Tests per Bildaufnahme zu dokumentieren um u.a. Zeitraffer-Aufnahmen des Schwingverschleißfortschritts zu erhalten. Gerade das Öffnen des Kontaktes kann bereits das Tribosystem und das Ergebnis eines Tests beeinflussen. Es wird gezeigt, ob und wie dieser Öffnungsvorgang einen Einfluss gegenüber Tests mit durchgehendem (nicht unterbrochenem) Kontakt hat. Schlüsselwörter Schwingverschleiß, Fretting, Tribometrie, Schmierstoffprüfung, Materialprüfung, Kontaktmechanik With the instrumentality of a newly developed fretting test bench for planar contacts, a state-of-the-art method of inter-mediate imaging helps to understand fretting wear mechanisms of different materials and lubricants. The test bench uses application like planar surfaces unlike the usual point or elliptical contact in model testing (with the tribological test chain). Applications considered prone to fretting wear have large planar contacts - like bearing seats and shaft hub connections - and contact pressures normally perceived as low or uncritical. This article examines a method to evaluate a targeted observation of the surfaces. The method uses a movable upper sample to open the contact and to document an interim status of the test by image recording. Among other things, this is to obtain time-lapse recordings of the progressive wear and tear. Just opening the contact can already influence the tribological system and the result of a test. It is shown whether and how this opening process has an impact on tests with continuous contact. Keywords Fretting, Tribometry, Lubrication Testing, Material Testing, Contact Mechanics Kurzfassung Abstract * Henrik Buse, MSc., Corresponding author Workplace: Institute of Production Technologies, Faculty of Materials Science and Technology, Slovak University of Technology, 917 24 Trnava, Slovak Republic Hochschule Mannheim - Kompetenzzentrum Tribologie 68163 Mannheim doc. Ing. Erika Hodúlová, PhD. Slovak University of Technology in Bratislava, Faculty of Materials Science and Technology, 917 24 Trnava, Slovak Republic ti r A m c e B \ u \ \ TuS_1_2021.qxp_TuS_Muster_2020 12.03.21 16: 23 Seite 15 2 State of the art for experiment with plane contacts Fretting wear testing can be done on different scales and abstraction levels. Typically, easy tests tend to simplify or avoid contact alignment issues by using point contacts. The compromise is accordingly an overdrawn local pressure and a very small contact area. The small contact area further requires correspondingly small amplitudes of motion in order to achieve the fretting-typical large contact ratio for both contact partners. Using coplanar contacts - flat or with any curvature - on the other hand introduces at least unavoidable shape deviations from manufacturing or preloading deformations. Furthermore, alignment becomes most important to avoid unintended edge bearing situations. However, edge bearing may be considered realistic for some applications, e.g. for splined couplings with moderate alignment tolerances. Considering literature about fretting wear, both contact types can be found [1, 2, 3, 6]. For point contacts, prominent works [8, 9] use the effect of partial slip to provoke high contact ratios and good repeatability, while contact pressure tends to be extreme and many applications do not rely on point contacts. The simplicity of point contacts appeals for standardizations. And these are conducted at least by the ASTM [10], e.g. D7594 - 19. But at a Aus Wissenschaft und Forschung 16 Tribologie + Schmierungstechnik · 68. Jahrgang · 1/ 2021 DOI 10.24053/ TuS-2021-0003 As a one-dimensional system schematic, figure 1 gives the respecting contact lengths of both bodies L C1 and L C2 and the displacement Δx in a typical ratio. Together with distributed normal force FN on the contact area the tangential force T will act against the motion. Also, the term fretting is used for a wider variety of special cases. A true special case are standstill marks in roller bearings which are exposed to vibrations and very small rotational oscillations with a contact ratio around one. The term fretting is also used in the roller bearing SNR-FEB2 and Fafnir ASTM D4170 standard test, but the contact conditions may have contact ratios well below one and open the contact completely [10]. So, it can be, that the fretting wear term is uses for different tests or phenomena under any oscillating motion. Within this article, the term fretting is used to the definitions of Heinz [1,3]. Fretting wear still occurs in many technical and everyday applications. Examples of fretting can be taken from earlier times described by Bartel [5] that are still of relevance today, as shown in figure 2. Especially the tooth shaft hub coupling of figure 2b) is a very present topic of early functional failure. Often the collective load would not be de-scribed as critical. Low loads for example may lead to less friction fit and more displacement. Larger contacting surfaces then are exposed to self or externally induced vibrations and fretting conditions can be met if the displacement does not exceed the contact length. It should be the objective to avoid or to mitigate the reason of vibration or relative motion, but this is often not possible by economical design or the usage of a machine. Some best practices define the necessary precision of mating parts (e.g. bearing seats) to avoid relative motion, most elements however require a clearance in order to allow a certain freedom of motion or for assembly reasons. Hence, a reasonable treatment is often needed for the avoidance or mitigation of fretting wear as symptom of small oscillating relative motion in large contacts. Typical fretting wear symptoms are abrasion, adhesion, metal debris forming, oxidation of the particles and surfaces if oxygen is present. And it is not clearly known if abrasion or adhesion happens first or if the causal mechanics can also depend on the tribological system. The later symptoms with debris and oxidation dynamics are also called tribo-oxidation [3]. Especially under partial slip under higher local pressures, surface cracks occur in the slip regions because the local contacts may interact intensively and fatigue with every cycle [3]. Figure 1: fretting wear system schematic Figure 2: Examples for fretting wear in applications 1960 s [5], most are still relevant today TuS_1_2021.qxp_TuS_Muster_2020 12.03.21 16: 23 Seite 16 first glance, the experimental setup just tests lubricant reflow with 150 µm oscillation amplitudes equal to the contact length, this gives a contact ratio of 0.5 one contacting surface therefore is completely exposed for lubricant access. Only with excessive initial wear, the contact gets large enough to suffice fretting contact ratios. So extreme pressure additives (about 2.2 GPa initial pressure) are needed to pass the test. But point contacts can easily have very high contact ratios under partial slip [8, 9], Experiments are quite simple to achieve because only an oscillating force is needed to overcome contact edge slip conditions. Works for planar contacts were mainly used to test for fretting fatigue and to understand the fretting wear mechanics more deeply. 3 Experimental Development With the help of a model test bench developed for plane contacts, new and simple methods of material and lubricant testing on basic and application research with fretting conditions can be demonstrated. The usual abstraction of the contact geometries with easily aligned point or line contacts is deliberately omitted. This is done to meet the usually flat and conformal contacts, which are prone to fretting in applied tribosystems (shaft-hub connection, bearing seats, mating surfaces, etc.). This also gives the lubricant its minimal volume in between asperities of the mating surfaces. To get a useful planar contact, the alignment as well as the shape deviations of the specimens must be considered. Figure 3a) shows the schematic of the dual axis alignment which is needed to adjust two plane specimens to each other. Figure 3b) shows the typical shape deviation of the specimens used for the described test series. It can be seen here that the surface is slightly convex with 1.5 µm across the distance from edge to centre of the contact. The complete test bench prototype is shown in figure 4a) and every single test records the initial specimen status, 4b), and contact alignment 4c). The specimens are oriented according to figure 4d). Aus Wissenschaft und Forschung 17 Tribologie + Schmierungstechnik · 68. Jahrgang · 1/ 2021 DOI 10.24053/ TuS-2021-0003 Figure 3: Contact alignment and specimen shape deviation Figure 4: Test bench prototype and contact documentation TuS_1_2021.qxp_TuS_Muster_2020 12.03.21 16: 23 Seite 17 An automated opening of the contact in defined intervals is intended in order to observe the wearing surfaces during a test. The schematic of this procedure is shown in figure 5. Opening the contact is to be considered as relevant in the tribological system analysis [3, 4]. The opening could provide an escape for trapped particles or a new flow of a lubricant to more central contact areas. A different approach would be to let the contact closed with a transparent specimen [9], but this is limiting the tribological system validity because of the lack of use of transparent materials in such tribosystems. The drivetrain for the oscillating motion is an adjustable eccentric drive with a servo motor as a main power source and a phase shift mechanism to control the displacement eccentricity. The crankshaft drive is much stiffer than voice coil drive solutions and does not require hydraulic power. Also a simple PID-control of the phase shift is sufficient to sustain the desired displacement amplitude. The displacement measurement is a single measurement consisting of a commercial laser triangulation sensor mounted on a passive mass path near the fixed specimen, the resolution of the device is not better than 0.4 µm and the linearity not better than 0.1 % at a sample cycle of 110 µs. The quasi-static normal force measurement is done via a strain measurement load cell (~2 mV/ V, 0.5 % linearity) in the force path on the motion axis which is also used for opening and closing the contact. The axis supplies up to 2 kN of normal force. The axis control system for the normal force and contact opening axis are based on “GRBL” open-source software which uses g-code for axis commands like CNC-Machines. Aus Wissenschaft und Forschung 18 Tribologie + Schmierungstechnik · 68. Jahrgang · 1/ 2021 DOI 10.24053/ TuS-2021-0003 The article focus is on the description and evaluation of an observation method of the contacting surfaces during the test run. The method uses a defined and repeatable, traversable upper sample to automatically open the contact and captures an intermediate image state of the test. This is to obtain time-lapsable images of the fretting wear initiation and progress mechanics. However, opening the contact can already affect the tribosystem and the result of a test. It is shown if and how this opening process has an influence on tests compared to continuous closed contact. The procedure schematic in figure 5a) also shows a movable mirror for the upper specimen image acquisition which was developed in a later stage. Exemplary images of the lower specimen in figure 5b) to k) show the potential outcome of a lubricated test. Beginning with contact alignment iterations b) to d), initial specimen status e), lubricant distribution before f) and after single contact g) as well as fretting damage progression h) to k). Figure 5k) shows a little spot where adhesion as cold weld is visible. 4 Methods, experiment equipment and setup The test setup consists of two flat specimens with an ideal contact area of 64 mm 2 . The test bench allows a rigid spherical alignment of the specimens for a well distributed plane contact which plane is also parallel to the axis of motion. Pressure sensitive foil is used as an visual alignment check. The foil used in the experiment has its maximum intensity colour at or above 2 MPa. Figure 5: Experimental contact inspection setup TuS_1_2021.qxp_TuS_Muster_2020 12.03.21 16: 23 Seite 18 The friction measurement is done on the structural path from the fixed specimen holder to the drivetrain via two piezo-electric strain sensors in a structure symmetric configuration for the cancelation of bending strain charges. This allows only the resulting push/ pull strain charge to be amplified by the charge amplifier. The resulting calibration charge per force is quite low with ~ 0.5 pC/ N and requires a sophisticated amplifier. Data acquisition hardware uses typical 16to 24-bit analogue interfaces for the measurements. The observation setup uses an industrial camera with a telecentric lens fitted to the specimen dimensions for object size. The control software is custom programmed for the test bench measurements and functionalities. The gravimetric measurement is done with specimen samples before and after a test with analytic balance with a repeatability of 0.015 mg and a linearity of 0.03 mg. 5 Discussion of the observation by intermittent image acquisition A characteristic behaviour of steel contact suffering of fretting conditions is the formation of debris particles. These mostly consist of Fe 2 O 3 , with bright red to brown colour and Fe 3 O 4 , with black colour [3]. A first observation method by a camera inspection on an earlier prototype is displayed in the image of figure 6. The closed specimen setup in contact and motion shows the amount of debris and the moment when the formation of debris starts. Besides the low resolution of the camera used then, the observation from outside does not give information on the mechanisms involved to produce the particles. But the images up close the contact revealed the steady flow of particles to the outside of the contact and therefore an interesting route to take a step further. An actual prototype therefore is further developed to observe the wearing surfaces during a test by automated opening of the contact in defined intervals. The concept schematic is shown in figure 5a). Opening the contact can surely be counted to be relevant in the tribological system analysis [4]. The opening provides at least an escape for trapped particles or a reflow of lubricant to more central contact regions. It must be considered that opening the contact should be done if the related tribological system does not have an interruption of contact. A fretting wear case on a casually open contact can be a shaft / hub spline coupling in an automotive drivetrain which can change sides between acceleration and deceleration and coasting. Cases where opening never occurs are bearing seats or friction fitted parts. Choosing to open the contact for additional information should be done carefully or by testing both possibilities with the same measures to be sure. 6 Experiment for contact opening influences To determine the influence on a tribo-system with planar contacts a test series is set up. Data is listed in table 1. The test without lubricant “dry test” have a fixed run time of around 4 hours for comparison. The time which takes to do the contact opening procedure is not compensated, therefore the tests with opening run slightly shorter. The opening, image acquisition and the closing take about 6 seconds to accomplish. The dry tests deliver measure values for friction levels and the gravimetric measurement of the specimens before and after the test to quantify the lost mass due to wear. The assembly paste “lubricated test” has a time to damage criterion, where the friction force is monitored, and the test is stopped above 600 N of friction force. The time to damage is then a criterion for comparison. The assembly paste is a typical white paste consisting mostly of calcium di-hydroxide Ca(OH) 2 , which is often used for applications where fretting wear can occur. The cause and effects for the general positive behaviour of this solid lubricant is not understood yet. Aus Wissenschaft und Forschung 19 Tribologie + Schmierungstechnik · 68. Jahrgang · 1/ 2021 DOI 10.24053/ TuS-2021-0003 Figure 6: Early prototype imaging of outside observation showing oxide particles Contact opening Lubricant Sample size Every 120 s. Dry 5 No opening Dry 5 Every 120 s. Assembly paste 3 No opening Assembly paste 5 Similar on all tests: pair of 100Cr6 / 1.3505 Steel specimens, 64 mm² contact area, 15.6 MPa / 1 kN normal load, 200 μm stroke at 49,5 Hz frequency, 120 °C Table 1: Test Series Data TuS_1_2021.qxp_TuS_Muster_2020 12.03.21 16: 23 Seite 19 without contact opening start at around 800 N whereas the tests with contact opening start with scattering from 800 to 1000 N directly. The sample number of dry tests allow a reasonable use of boxplots, but obviously the number of trials could have been higher. Data is shown in the upper figure 9. The friction force levels on top of figure 9 do show only small differences. But it can be seen in the lower plot, that the wear in the case of cyclic opened contacts is slightly different than in the case of a completely closed contact; scattering is increased slightly on contact opening. Considering the slightly shorter tests due to the opening cycles, the non-opening tests show less wear. The amount of wear after 4 hours of test (713k Cycles, 285 m of sliding distance) in the single digit milligram range is well measurable with gravimetric measure equipment. More distinct and unexpected is the wear ratio between upper and lower specimen, the lower ones having around 50 % less wear without any dependency on contact opening. It is unclear why this is that pronounced between the specimens, but an assumption is, that there is a nodding or rolling kinematic on very high friction forces caused by the lower specimen holder. The lower specimen holder stiffness is much lower than the upper specimen holder due to cavities for heating elements and softer thermal isolation plates in between. This may lead to a nodding together with a smaller more pressured area on the motion following edge. Another Aus Wissenschaft und Forschung 20 Tribologie + Schmierungstechnik · 68. Jahrgang · 1/ 2021 DOI 10.24053/ TuS-2021-0003 7 Results and discussion dry test The dry test results show a comparable picture after the test. Figure 7 shows the acquired images of the dry tests. Columns are different test samples, and the top three rows are results with contact opening. In the first row the alignment documentation is imaged, the second row has the first image after 120 s and the third row the last image after 4 hours. The bottom 2 rows are images of the tests without contact opening showing alignment and last image after 4 hours. There are 2 images missing because of failed recalls of the manual image acquisition after the test. In the second row in figure 7, after 120 seconds every sample shows red oxide particles as well as a fresh adhesion or cold weld spot. Larger adhesion damage is no more visible at the end of the test. Together with the measured friction forces shown in figure 8 there is hardly any difference comparing the opened and closed contact tests. The shown diagrams plot the absolute mean friction force calculated over time. In every condition, the gross slip regime [7] is present. It is expected that the closed contact may lack of oxygen in the contact centre and therefore show more bare metal surface. But this is not the case when comparing the macro images. The measured friction forces are all in the region of 800 to 1000 N therefore the coefficient of friction is around 0.8 to 1. Most noticeable is the difference of friction forces in the first half hour of the test. All tests Figure 7: Image acquisition results on dry tests TuS_1_2021.qxp_TuS_Muster_2020 12.03.21 16: 23 Seite 20 Aus Wissenschaft und Forschung 21 Tribologie + Schmierungstechnik · 68. Jahrgang · 1/ 2021 DOI 10.24053/ TuS-2021-0003 Figure 8: Measured friction force on dry tests Figure 9: data comparison boxplot on dry tests TuS_1_2021.qxp_TuS_Muster_2020 12.03.21 16: 23 Seite 21 the friction force shows spikes towards lower forces because the data filtering does not remove every measure sample while the contact opening procedure is ongoing. 9 Conclusion and outlook on the benefits of contact opening Opening a contact in intervals does affect results like wear amount less than other effects. But the shown lubrication tests issue a need for an increased number of tests for better statistics and other lubricants to determine a clear influence of contact opening. Opening a contact for additional status information of the tribological system in test may be very useful but can also influence the result compared to a test with an always closed contact. A bit harder to explain in short text and figures is the amount of detailed data generated by image acquisition while the contact is opened. Image data of the test series show the initial damage type on the lubricated test and the initial and progressive damage mechanisms in dry testing. The images are well suited for generation of time lapse movies for a better understanding of progression. Figure 12 shows some successive images of a test and figure 11 uses a combined view of online-data together with synchronized images. Event based image acquisition can also be used to only take an image when friction levels or other measures Aus Wissenschaft und Forschung 22 Tribologie + Schmierungstechnik · 68. Jahrgang · 1/ 2021 DOI 10.24053/ TuS-2021-0003 difference between upper and lower specimen is the mass temperature since the lower specimen is heated directly by the holder and the upper specimen gets heated passively through the contact and the friction losses. A difference of around 40 K is normal on the temperature measurements (120 °C on lower and 80 °C on upper specimen) but this is also depending on sensor placement. If tests with high friction forces (~ 1000 N on the shown device) are conducted in the future, this behaviour must be addressed as it may induce an oscillating unknown dynamic contact pressure distribution. 8 Results lubricated tests The white assembly paste lubricated test show the resource conflict of long test runs and enough repetitions for statistical evidence. With the given results shown in figure 10, the time of failure can be seen in a quick rise of the friction force at a certain time. The tests with and without opened contact show nearly the same time to damage median. Both can scatter by 100 %. Also, one test with closed contact showed some measurement trouble (the longest run). A conclusion could be that contact opening does not increase time to damage by more lubricant flow into the contact regions or other effects. But this may differ with the type and consistency of lubricant. Values for the friction force are comparable. On the contact opening tests, Figure 10: Measured friction force on lubricated tests TuS_1_2021.qxp_TuS_Muster_2020 12.03.21 16: 23 Seite 22 raise or drop by a certain amount. The topic is looking out to online publications with additional media usage to better show the general benefit of the tests done with contact opening and inspection. Especially with internal studies of polymer coatings the procedure seems to have no effect on the test outcome. This provides the opportunity to understand transfer layer built up of different coatings and the effects of additional influences like temperature or other physical changes. The experiment setup can also include additional analytics like UV/ VIS-spectroscopy, photometry or fluorometry if the surrounding can be shielded well enough. Fretting can induce many subsequent surface contacts in a very small area and therefore some analytical methods can resolve fresh reactions of the surface materials. References [1] R. Heinz, Schwingverschleiß-Erscheinungsformen - Prüfmethoden und Abhilfemaßnahmen, Journal, Materialwissenschaft und Werkstofftechnik 20(1): 14-20, 1989 [2] R. Simon, G. Schmitt-Thomas. Reibkorrosion - Literaturrecherche. FVA-Number 313, Forschungsvereinigung Antriebstechnik e.V., 1990 [3] K. Sommer, R. Heinz, J. Schöfer. Verschleiß metallischer Werkstoffe, ISBN 978-3-8351-0126-5, 2010 [4] H. Czichos, K.-H. Habig, Tribologie-Handbuch, ISBN 978-3-8348-1810-2, 2015 [5] Albert A. Bartel. Passungsrost bzw. Reiboxydation, besondere Verschleißprobleme. Erfahrungsberichte. Allianz Versicherungs-AG, 1965. [6] R. Simon, G. Schmitt-Thomas. Reibkorrosion - Literaturrecherche. FVA-Number 313, Forschungsvereinigung Antriebstechnik e.V., 1990 [7] Olof Vingsbo, Staffan Söderberg, On fretting maps, Wear, Volume 126, Issue 2, 1988, Pages 131-147, ISSN 0043- 1648, https: / / doi.org/ 10.1016/ 0043-1648(88)90134-2. [8] Johnson, K. L., Contact Mechanics. Cambridge: Cambridge University Press. 1985, doi: 10.1017/ CBO9781139171731 [9] Popov, V. L. Kontaktmechanik und Reibung, Springer Verlag, 2010, ISBN 978-3-642133-02-2 [10] Neu, R., Progress in standardization of fretting fatigue terminology and testing. Tribology International - TRI- BOL INT. 44. 2011, https: / / .doi.org/ 10.1016/ j.triboint. 2010.12.001 Aus Wissenschaft und Forschung 23 Tribologie + Schmierungstechnik · 68. Jahrgang · 1/ 2021 DOI 10.24053/ TuS-2021-0003 Figure 12: successive images from a lubricated test Figure 11: Two frames of a time lapse video with online data TuS_1_2021.qxp_TuS_Muster_2020 12.03.21 16: 23 Seite 23