1 Introduction Reliable radial shaft seals (RSS) depend on a mechanism, in which the combination of sealand shaft material, geometry, surface (micro-)structure and manufacturing method, creates an active reverse pumping effect in the sealing contact. The latter can, on the one hand transport liquid or gaseous fluids under the sealing lip in Aus Wissenschaft und Forschung 22 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 Online determination of reverse pumping values of radial shaft seals and their tribologically equivalent system Christoph Burkhart, Stefan Thielen, Bernd Sauer* Eingereicht: 11. 9. 2020 Nach Begutachtung angenommen: 26. 11. 2020 In diesem Beitrag wird eine digitale Förderwertmesseinrichtung auf Basis der Zweikammermethode vorgestellt. Dabei wird der Ölfüllstand mithilfe des hydrostatischen Drucks in beiden Kammern mit zwei hochpräzisen Drucksensoren überwacht. Über den Druckabfall kann der Förderwert bestimmt werden. Mit dieser Methode lassen sich auch zeitabhängige Aussagen über den Förderwert treffen. Dazu wurden Förderwerte eines repräsentativen Dichtringsystems bei verschiedenen Temperaturen, Drehzahlen und Schmierstoffen bestimmt. Weiterhin wurden unterschiedlichen Wellenoberflächen hinsichtlich des Förderverhaltens untersucht. Das Portfolio umfasst dabei geschliffene, einfach und mehrfachgedrehte Wellen aus 16MnCr5 (1.7131, AISI 5115). Die gleiche Prüfeinheit wurde auch modifiziert, um den Fördermechanismus im tribologischen Ersatzsystem des RWDR, dem Ringflächentribometer (RFT) nachzuweisen. Dabei gleitet eine einfache Gummischeibe auf einem Wellenkegel. Der Kegelwinkel ist dabei so gewählt, dass die Kontaktverhältnisse zwischen Scheibe und Welle denen des RWDR-Dichtkontaktes nahe kommen. Auch hier wurden Förderwerte bei unterschiedlichen Drehzahlen bestimmt und mit den Ergebnisse der Förderwertmessung beim RWDR verglichen. Die Ergebnisse deuten darauf hin, dass der Fördereffekt im tribologischen Ersatzsystem (RFT) existiert. Schlüsselwörter Radialwellendichtringe; Förderwert; Tribologisches Ersatzsystem; Online-Messung In this contribution, an online pumping rate measurement device for radial shaft seals according to the two-chamber principle is presented. The oil level in both oil-filled chambers is monitored based on the hydrostatic pressure by pressure sensors and related to a pumping value. With this method also time depended statements of the pumping value can be formulated. Pumping rates of a representative radial shaft seal system under variation of temperature, speed and lubricant were determined. In addition to the operation parameters the influence of the shaft micro structure is discussed. The shaft portfolio includes ground and simpleand multiple turned shaft made out of AISI 5115. The same device was modified to measure and prove the pumping mechanism in the tribologically equivalent system of radial shaft seals, the ring cone tribometer (RFT). In this system, a simple elastomer ring is sliding against a conical shaft under lubricated conditions. The contact situation of a shaft seal can be replicated by the cone angle. Pumping rates of the same material combination under variation of speed were determined and compared to the corresponding results obtained with the RSS. The results indicate that the pumping mechanism is also present in the tribologically equivalent system of the RSS. Keywords Radial Shaft Seal; Reverse Pumping Value; Tribologically Equivalent System; Online-Measurement Kurzfassung Abstract * Dipl.-Ing. Christoph Burkhart Orcid-ID: https: / / orcid.org/ 0000-0002-5485-893X Jun.-Prof. Dr.-Ing. Stefan Thielen Orcid-ID: https: / / orcid.org/ 0000-0003-3310-7659 Prof. Dr.-Ing. Bernd Sauer Orcid-ID: https: / / orcid.org/ 0000-0002-3489-5805 Technische Universität Kaiserslautern Lehrstuhl für Maschinenelemente und Getriebetechnik (MEGT), 67663 Kaiserslautern TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 08 Seite 22 order to lubricate the contact zone and transfer heat away from the contact and into the oil [1, 2]. On the other hand, the same effect can prevent leakage, seal the gap and even pump lubricant back to the oil side of the seal [3]. Driving force of the mechanism is an asymmetric distortion of axial oriented wear structures [4, 5]. The asymmetric distortion is thereby mainly controlled by an asymmetric contact pressure distribution, initiated by a larger oil side angle compared to the air side angle of a seal as well as the location of the spring centerline (figure 1). Following H ORVE [2] the mechanism is already starting up within the first revolutions of a sealing system. The rough shaft removes material from the sealing lip and initiates the formation of micro-asperities on the seal contact interface by smoothing of the shaft in the seal wear track. The roughness of the seal is directly linked to the pumping ability [6]. Since the ability of reverse pumping oil is crucial for the seal operation, specifically to analyze the dynamic tightness of shaft seals, the reverse pumping value was established as a characteristic number for shaft seals [4]. Many methods for the experimental determination of the reverse pumping values were developed and established. Early, J AGGER found evidence that a RSS is able to reverse pump a fluid by an investigation adding red color to a clear liquid lubricant which he applied on the air side of the seal. He was able to find the colored lubricant on the oil side of the seal, proving a transport mechanism from air to oil side [7]. According to him, the at that time unknown effect, is based on diffusion, since he had different objectives with his tests. In 1973 S YMONS established a verification method for the sealing mechanism of shaft seals by injecting a small oil droplet on the air side [8]. This method was standardized in SAE J1002 (withdrawn) [9]. Based on their observation of leaky seals, H ERMANN and S EFFLER used a similar injection method for the verification of the existence of the reverse pumping effect. A small droplet of oil was set off close to the sealing contact at the air side. The release of the drop was leading to a sharp decline in the friction torque, which normalized after the oil droplet was consumed [10]. H ORVE examined the pumping ability by a measurement of the time required to transfer a known quantity of oil from the air to the oil side [6]. Also in his experiment a “conventionaly” installed seal [11] was subjected to droplets on the air side, that were pumped to the oil side. Besides the monitoring of the friction torque, an observation of a fluorescent-dyed oil meniscus was also possible (figure 4, right). The mentioned methods are suitable for quick examinations of sealing systems. The major disadvantage is the lack of quantification of the reverse pumping value due to short-term testing and the lack of quantification of the pumped fluid dose, which can lead to systematic errors when determining absolute values. Even though the injected volume can be more or less quantified, the quantity of the lubricant that enters the sealing contact cannot. B RITZ und F RITZSCHE have further criticized this method for capturing an instationary reverse pumping effect, since surface tension and capillary forces cannot be neglected on the short-term as well as the highly temperature depended fluid behaviour [12] . To investigate the time dependency of the suction volume of the oil a “conversely” installed seal with flooded air side, but unlubricated oil side, was used by K AWAHARA and H IRABAYASHI [11]. The excess lubricant on the oil side was evaluated using a leakage collector. With this method the influence of roughness, viscosity, eccentricity, speed and material provided further information on the “sucking effect”, also known as “pumping effect”. For the first time they have observed a dependency of the pumping coefficient K [11] on the surface orientation and consider tangentially deformed axial microstructures as reason for the reverse pumping effect [13]. V OGT and J OHNSTON mounted the seals with the air side towards the oil side (“converse”) and measured the amount of fluid running out when the shaft is rotated [14]. A comparison between injection and the method by K AWA - HARA and H IRABAYASHI [11] was achieved by S CHULER [15]. Based on the criticism about instationary reverse pumping effects with the so far evaluated injection and converse installation methods, B RITZ [12] and F RITZSCHE [16] developed a procedure for the reliable measurement of the reverse pumping value based on the two individual oil chambers, separated by a central oil seal [12]. The lubricant is reverse pumped from a first to a second chamber on the oil side, while the transported oil volume is displayed in capillary tubes. The method was later enhanced by R UHL [17]. The fully flooded condition allows for a clear determination of the shear induced flow, since surface tensions, induced by pressure differences can be neglected. K UNSTFELD summarizes and presents different ways of operating such a setup [18]. According to him, fully flooded conditions show the best reproducibility, while a wetted air side and fully flooded oil side are more application oriented. He modified the test setup accordingly by implementing a single riser pipe to the fully flooded air side and operated the oil side with oil levels based on the application in focus. Using this established method, a variety of influences on the reverse pumping values of radial shaft seals have been investigated: S CHMUKER analyzed the influence of the shaft surface roughness, viscosity of the oil and sliding speed on reverse pumping value in sealing systems with ACM, NBR and FKM seals [19]. First K LAIBER , and later S CHULER correlated wettability measurements (contact angle, surface tension, wettability factor [20] and adhesion work) of various lubricants (with a selection of additives) on shaft and elastomer surfaces to the reverse pumping values [15] and other characteristic values of a sealing systems [20, 21]. Based on a high amount of measurements with variation of all parameters applicable in a sealing system, R EMPPIS developed an Aus Wissenschaft und Forschung 23 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 08 Seite 23 In general the two chamber method combined with pressure sensors allows for a direct and more accurate determination of this characteristic value, which is the reverse pumping value. A delay between reverse pumping action and it’s detection, which occurs in a collector based measurement setup, is not expected. 2 Experimental The following section gives an overview about the materials and lubricants used for the investigations. A synthetic polyalphaolefin (PAO) lubricant of the SAE class 0W20 was chosen as a representative of a typical modern engine oil. This oil is a reference lubricant, based on a blend of two polyalphaolefines (PAO4, 18.4 wt.-% and PAO6, 65 wt.-%) and an ester (Plastomoll DNA, 10 wt.-%). Only a solution of multifunctional dispersant viscosity-index improver (VII) (Viscoplex ® 6-850: Dispersant Polyalkyl Methacrylate (PAMA), 6.4 wt.-%) is included as an additive. As reference lubricant for an industrial oil, a polyalkylene glycol (PG) lubricant from the ISO VG 68 class was used. Apart from a dispersant and anti-oxidants no additive packages are included. The PG lubricant is soluble in water, but insoluble in hydrocarbons. [30] Radial shaft seal seals of type DIN 3760-A80x100x10- FKM were used [30, 31]. These were moulded form a very common commercial state of the art fluorelastomer (FKM). The cross-linking system is of bisphenolic nature. The seals are equipped with a spring, but have no protection lip. On the oil side the sealing lip was pricked with a Aus Wissenschaft und Forschung 24 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 empirical model for the estimation of a reverse pumping value [22]. B EKGULYAN expanded the application area of the empirical equations towards synthetic lubricants, low temperature and instationary test conditions like shaft runout [23]. T OTZ et al. reviewed the reverse pumping effect of seals, when subjected to a change in rotational direction after they have been running in a certain direction for quite some time (“conditioning”). They found smaller reverse pumping values after the directional change [24]. Regarding the measurement results of all the methods presented above, it must be noted, that the determined reverse pumping values represent the maximum possible amount of fluid transported back to the oil side and not the actual reverse pumping effect that occurs during conventional operation. During conventional operation of a seal, the reverse pumping effect is counteracted by the oil meniscus being ingested into the sealing contact until the reverse pumping effect is ~0 [25]. Nevertheless, the measurement of the reverse pumping values provides valuable information about the resistance to leakage promoting effects such as shaft lead. Following the tradition of the investigation of the reverse pumping value at the institute of machine elements gear and transmissions, TU Kaiserslautern (MEGT), initiated by B RITZ and S TEINHILPER , in this contribution, an online reverse pumping value measurement device (twochamber principle) for radial shaft seals and their tribologically equivalent system is presented. A similar device has been presented by the IMKT in Hannover [26 to 29]. In FVA 432 I additionally the airflow at different pressure gradients was evaluated with a setup similar to F RITZSCHE [16]. Figure 1: Scheme of distortion in the sealing contact leading to the reverse pumping mechanism. TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 08 Seite 24 Parameter Standard Unit PAO PG MIN Base Oil - - Polyalphaolefin Polyalkylene Glycol Mineral (FVA 2) Kin. Viscosity, 40°C DIN 51562 mm 2 / s 38.9 71.23 32 Kin. Viscosity, 60°C mm 2 / s 23.1 36.57 - Kin. Viscosity, 80°C mm 2 / s 14.05 21.93 - Kin. Viscosity, 100°C DIN 51562 mm 2 / s 9.2 14.82 5.35 Viscosity Index ASTM D2270 - 231 220 97 Viscosity Class DIN 51 519 ISO VG 46 68 32 Density, 20°C DIN 51757 kg/ m 3 835 1,036 870 TAN (initial condition) DIN 51558-1 mhKoH/ g 0.02 0.15 0.01 Water Content DIN 51777 % < 0.1 < 0.5 - Pour Point ISO 3016 °C - - 50 -15 Flash Point ISO 51376 °C >120 + 250 +220 Ignition Temperature ISO 51794 °C - + 360 - API Group - - IV V I sharp blade, on the air side the sealing lip surface is defined by the injection moulding process. The seal inner nominal diameter was 80 mm, the bore diameter 100 mm and the width 10 mm [30, 32]. In the relaxed state the inner diameter was measured as 78.664 mm on average. The test shafts with a diameter of 80 mm and with a width of 18.5 mm were manufactured from basic rod material (AISI 5115). During the manufacturing process all test shafts were turned in a first step on a CNC lathe by rough turning them over complete length of 18.5 mm to a diameter of 80.2 mm. During the turning process cooling emulsion was applied as cooling lubricant. The AISI 5115 work pieces were carburized to 0.7 - 1.0 mass-% C in the surface near area at 880 °C - 1,000 °C in carbon-gaseous atmosphere [33] and then case hardened to 60 HRC with a penetration depth (CHD) of 0.8 mm. In a final manufacturing step the shafts were ground to 80.0 +/ - 0.025 mm using a conventional grinding machine (S TUDER S20) with a conventional grinding wheel (CBN, 350 mm, 31 m/ s). During the process a water soluble synthetic cooling lubricant (ZuboraTDD) was applied. Following the currently withdrawn standard DIN 3761 [34] the grinding process was performed as a plunge-cut grinding process to achieve a twist-free surface structure. [30] For the initial verification of the test setup multiple turned shaft surfaces [35] and cryogenic two step turned surfaces (AISI 347) [36] were compared to state of the art infeed ground shaft (AISI 5115) surfaces according to DIN 3761 Part 2 [34] (figure 2). For every shaft surface a fresh seal and shaft were used. Aus Wissenschaft und Forschung 25 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 Table 1: Technical data of the lubricants in this study. The data was taken from the manufacturers data sheets and measurements in the lab. shaft surface a fresh seal and shaft were used. Figure 2: Topography of the shaft surfaces used for the verification of the reverse pumping value measurement in chapter 3. TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 08 Seite 25 turing methods of the elastomer ring samples, the RFT test rig and the tribological properties in terms of comparability can be found in FVA 578 I and II [38 to 40]. 3 Determination of the reverse pumping value in the RSS-system 3.1 Test Setup and Test Parameters of the RSS-System The test setup (figure 3) for radial shaft lip seals consists of a primary and secondary oil chamber of the same volume. The hydrostatic fluid pressure in both oil-filled chambers is monitored by two high resolution pressure sensors. The measurement setup is based on the converse installation of the RSS presented by K AWAHARA and H IRABAYASHI [11] as well as B RITZ and was operated as proposed by K UNSTFELD [18] as follows: Chamber-2 (air side) and the attached riser pipe are initially completely filled with the test lubricant, while chamber-1 (oil side) is just filled up to a different level that represents the actual application (here mid shaft). Measurements using fluorescent dyed oil in chamber-1 with a conversely installed RSS (figure 4, right) in a pre-test have shown the development of a fluorescent oil meniscus on the nominal oil side, thus visualizing the reverse pumping effect. The left side of figure 4 shows a typical pressure development in chamber-2 during a measurement. The pres- Aus Wissenschaft und Forschung 26 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 The coned shafts were also manufactured from case hardened, low carbon steel AISI 5115 and ground according to DIN 3761 Part 2 [37]. A special grinding wheel with a chamfer angle identical to the cone angle was used and parallelism of both workpiece and grinding wheel axis was ensured. The special geometry and parallelism of the grinding wheel is necessary for a lead - free surface. Roughness parameters are in the same range for both ground shaft types(table 2) Parameter Unit Cone (30°) Shaft S k μm 0.488 0.494 S pk μm 0.156 0.1533 S vk μm 0.288 0.300 S mr1 % 8.15 7.88 S mr2 % 87.28 86.87 Table 2: Comparison of the functional parameters of ground cone and ground cylindrical shaft according to DIN EN ISO 25178. In the RFT setup FKM ring samples with an inner diameter of 50 mm and outer diameter of 75 mm were used, which both are commercial materials. They are produced from 2 mm thick test slabs by a defined four step production process. The same FKM material was used for the radial shaft seals. Further details about the manufac- Chamber- 1 Auxiliary Seal Drive Shaft Radial Shaft Seal Chamber - 2 Cover Plate PT-100 Pressure Sensor Housing Bore Test Shaft Riser Pipes Figure 3: Exploded cross sectional view of the two-chamber-principle test setup for the reverse pumping value measurement of radial shaft seals. 0.153 TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 08 Seite 26 The performance of a measurement in one rotational direction first of all leads to a combined reverse pumping value as the sum of the portion created by the RSS and the shaft (figure 5). In this paper, with the exception of the investigation of the three different shaft surfaces, the system reverse pumping value (P SYSI ) were determined. In order to differentiate between the ratio of RSS and shaft, a second measurement in the opposite rotational direction needs to be performed (figure 5). This was always conducted with the same seal in the same running track. According to T OTZ et al. [24] a rotation in a constant direction can cause a conditioning of the seal and lead to a different behaviour, when the same seal is then subjected to a change in rotational direction. However, Aus Wissenschaft und Forschung 27 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 sure and thus the oil level in the riser pipe decreases linearly due to the the entire lubricant transfer through the shear flow during shaft rotation. The duration of a measurement is dependent on the rotational speed. The absolute reverse pumping value (PV) of a combination of shaft and seal can be determined using equation (1), where K is the absolute value of the gradient (compare figure 4) of the pressure signal in chamber-2 (figure 3), d RP the diameter of the riser pipe, ρ the density of the lubricant at test temperature, g the gravity constant (g = 9.81m/ s 2 ) and v the relative speed in m/ s in the contact zone of the radial shaft seal. The reverse pumping value is normalized to the sliding distance and therefore set to µl/ m. Depending on the intention of the measurement also gravimetric specification in g/ h is possible [41]. (1) Prior to the measurement, the radial shaft seals performed a 100 km run-in at 5 m/ s. To prevent thermal effects on the results, a thermal equilibrium between the two chambers was established during the run-in before every test. This was achieved using a PID-feedbackcontrolled heating circuit is available for the individual temperature control of both test chambers. The test was ended, when the riser pipe in the secondary chamber was emptied by the transfer of lubricant from the second to the primary chamber. The lubricant temperature is kept constant by two heating circuits acting individually on each chamber. = | | ∙ ∙ ∙ ∙ 4 ∙ Figure 4: Development of the hydrostatic pressure in chamber-2 of the RSS test setup under variation of the relative speed of the rotating ground shaft in the combination: FKM / PAO / oil sump temperature 60 °C. The pressure decreases faster with increasing relative speed. The red line indicates the gradient obtained with a linear fit. Right: Visualization of the reverse pumped lubricant film on a conversely installed shaft seal using fluorescent dyed oil on the two-chamber-test rig (figure 3). The light blue area in the white box is the oil meniscus visualized by the purple UV-light. PV-RSS PV-RSS PV-Shaft PV-Shaft PV-System I PV-System II PV_SHAFT PV_SYSTEM I -II Figure 5: Measurement technique according to R AAB [42] to determine the individual reverse pumping value of a shaft seal and shaft surface. TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 08 Seite 27 3.2 Results in the RSS-System Figure 6 depicts the system reverse pumping values for different RSS-shaft-lubricant combinations at two different oil sump temperatures. The system reverse pumping value is nearly independent of the relative speed, when normalized in respect to the sliding distance. A slight decrease in the values can be observed for an increase in the oil sump temperature, which is probably caused by a reduced lubricant viscosity. A clear difference is observed for the influence of the base oil. While PAO and MIN (table 1) generate almost similar pumping values in all operation points, the PG lubricant has a three times higher system reverse pumping value at 60 °C, respectively up to two times higher at 80 °C. In agreement with [22], increased viscosity increases also the reverse pumping values of the system. As S CHULER evaluated, that apart from viscosity other influence regarding tie base oil, like wettability contribute to the pumping ability [15]. Apart from the general trends, also the absolute values are in agreement with very comparable systems, evaluated with similar methods for example in [15, 22, 23]. In a second step the pumping values of four different shaft surfaces (figure 2, two ground surfaces, besides the multiple turned and cryogenic turned surfaces, were used) PV shaft was evaluated in the same test configuration (figure 3) using PAO as lubricant at 70 °C oil sump temperature. The same speed range from 2.5 m/ s to 10 m/ s in 2.5 m/ s steps as before was covered. This time both rotational directions were evaluated, in order to allow for the separation between shaft and seal using equation (2) and equation (3). At first the shafts were rotated counter-clockwise, the clockwise. In the first row of figure 7, classified by the rotational direction (left: counter clockwise (PV sys I ,(+)), right: clockwise(PV sys I I (-))), all values during this measurement series for a single RSS are summarized in one box plot to visualize their overall spread across different Aus Wissenschaft und Forschung 28 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 the replacement of a seal or the running track also comes with deviations of the pumping value, for example due to geometry and radial force tolerances of the seal and a slightly different position of the sealing lip, why the former method was chosen in this study. All other parameters were kept constant. According to R AAB et al. [42] the reverse pumping value of the seal is the mean value of the sum of the two system pumping values (equation 2). The reverse pumping value of the shaft can be calculated by half the difference between the reverse pumping value in system I and II (equation 3). (2) (3) It must be noted that this method is only valid for PV shaft < PV seal , which is the case in the investigations presented in this paper. However, if this were not the case, PV shaft would be underestimated. Using the method described above, pumping rates of a representative radial shaft seal system under variation of oil sump temperature (60 °C, 70 °C, 80 °C), speed (2.5 m/ s - 10 m/ s), and three different base lubricants (table 1) were determined. Significant results were obtained by a repetition of every parameter combination by means of three measurements. This also included the change in rotational direction. The influence of speed was evaluated with the same seal in the same wear track on the same shaft. When a test lubricant or shaft were changed, the shaft and seals were replaced by fresh samples. A lubricant exchange was accompanied by a thorough clean-up of the test setup. The results of the measurements in the RSS-system are featured in the next chapter. = 1 2 ∙ ( + ) = 1 2 ∙ ( − ) Figure 6: System reverse pumping values (PV sys I ) for RSS type DIN 3760-A80x100x10-FKM running at ground shaft surfaces at different speed stages for three different base lubricants without additives at a constant oil sump temperature of 60 °C (left) and 80 °C (right) TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 08 Seite 28 speeds and repetitions. Although different shaft surfaces were combined to different RSS of the same type, the reverse pumping values are very similar through the selection. It is noticeable, that the initial rotation in counterclockwise direction led to a slightly higher average reverse pumping value compared to the clockwise direction, that has been evaluated with the same seals afterwards. This is in agreement with the conditioning by [24]. Although since some shafts were turned, lead could superimpose with the latter effect. In figure 8, the same radial shaft seals (figure 7) are summarized per speed stage, without distinction to the surface structure, to depict the spread of different RSS on the system reverse pumping value. As already indicated before, the reverse pumping values are nearly independent of the rotational speed, even across a certain RSS population (due to the normalization based on the sliding distance, bottom left , PV sys I (+)), bottom right: clockwise, (PV sys I I (-)). The system reverse pumping values of the four RSS in both rotational directions were separated by the speed. The Same shaft type was combined to the same RSS (figure 7) and used throughout the whole measurement series. Through a difference of system I to system II according to equation (3) the pumping value of the shaft was calculated (figure 9). Both, multiple and cryogenic turned surfaces are within the range of the state of the art ground shaft surfaces. The ground shaft surfaces seem to have very small lead angles since “Ground I” generated negative pumping values, while “Ground II” had positive values across the speed stages. No explanation for the outlier at 5 m/ s of the cryogenic turned surface could be found. All absolute values are a magnitude smaller than the RSS pumping value. The here tested FKM shaft seals, generate a lot of additional potential to achieve dynamic tightness in this sea- Aus Wissenschaft und Forschung 29 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 Figure 7: Boxplot of the reverse pumping values of the four radial shaft seals corresponding to the shaft surfaces shown in figure 2 (left: counter clockwise (PV sys I ,(+))),right: clockwise, (PV sys I I (-)) . The same seal on the same wear track were used for clockwise and counter clockwise rotation. RSS-1: Multiple Turned; RSS-2: Cryogenic Turned; RSS-3: Ground; RSS-4: Ground; Figure 8: Boxplot of the reverse pumping values of the four speed stages (left: counter clockwise (PV sys I ,(+))), right: clockwise, (PV sys I I (-)). All values of the from figure 7 were summarized, without distinction to the surface structure. TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 08 Seite 29 tuation of a shaft seal is either replicated by the cone angle of the shaft or the chamfer of the elastomer ring sample. Despite its use for many years by researchers in comparable configurations [38, 44, 45] there is still a lack of verification, that this system can reliably reverse pump and hence be used for comparative investigations. A first proof of the existence of a reverse pumping behaviour was achieved by using the vertical test configuration of the RFT and by an injection of a defined volume of lubricant below the ring sample air side (figure 10). Similar to H ORVE [2] a very characteristic behaviour (figure 10) was displayed in the friction torque after injection of the oil by a small syringe. The slight tendency of decreasing friction torque in the time period without injection is due to thermal effects, that were arising since the oil drop test was not conducted in a thermal equilibrium. The vertical configuration however also complicated the accessibility, while the majority of the lubricant was hurled away by the conical shaft. Doubts about the reliability of the presented method lead to an enhancement of the initial test configuration (figure 3) as described in the next section. For a quantitative determination of reverse pumping values of ring samples, the pumping rate measuring setup for RSS (figure 3) was modified to allow for measurements in RFT configuration (figure 11). The principle of the test rig is identical to the setup shown in figure 3, with the exception that an elastomer ring sample is sliding on a conical shaft, replacing radial shaft seal and cylindrical shaft. The radial force is replaced by an equivalent line force (0.14 N/ mm [46]), applied by a spring unit, that presses the elastomer sample against the coni- Aus Wissenschaft und Forschung 30 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 ling system. In a similar investigation with shot peened and particle structured shaft surfaces the pumping values has been [43] Since the shaft surface has a very low impact on the system reverse pumping value, the explanation for the similar behaviour of the seals (figure 7) can be given. 4 Determination of the reverse pumping value in the RFT-system 4.1 Test Setup and Test Parameters of the RFT-System At the institute of machine elements, gear and transmissions, a tribologically equivalent system for radial shaft seals (RFT) was developed. In this system, a simple elastomer ring sample is sliding against a conical shaft (cone angle 30°) under lubricated conditions. The contact si- Figure 9: Pumping values for four different shaft surfaces at four different relative speeds. 0 2 4 6 8 time / min 0 0.05 0.1 0.15 0.2 0.25 0.3 friction torque / Nm Figure 10: Left: Friction torque during an oil drop test according to H ORVE [2] on the original vertical RFT test setup, which indicated the existence of a reverse pumping value in the tribologically equivalent system for RSS. Right: Principle of the ring-cone tribometer. TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 09 Seite 30 cal shaft. The line load is set according to the line load obtained by radial force measurements at test temperature of the corresponding RSS. The measurement principle is also identical to the RSS setup with the exclusion that the primary chamber is fully filled with lubricant, while the secondary chamber is only filled up to mid shaft, since the RFT systems is flipped in order to improve accessibility to the load unit. Measurements based on the hydrostatic pressure in chamber-1 have two small disadvantages in this setup. The reader must be cautioned, that the secondary seal might possibly influence the reverse pumping results in this configuration. Tightness of the auxiliary seal therefore was checked regularly. Additionally, preliminary investigations with both chambers fully flooded have shown, that the amount of fluid level decrease in chamber 1 matches the amount of fluid level increase in chamber 2 with good agreement, indicating that the secondary seal does not significantly influence the measurement. The second one regards the movement of the shaft into chamber-1, which leads to more a wavy oil level and little noise in the signal. The main objective of this setup was the quantification and definite proof of the existence of a reverse pumping effect in the tribologically equivalent system for RSS. For the comparative investigation, elastomer ring samples made from the same material as the corresponding seals were used (FKM). Reverse pumping rates were determined and compared to RSS reverse pumping values at 7.5 m/ s and 10 m/ s. The shaft surface was ground. An oil sump temperature of 60 °C was chosen for this comparative measurement. Again, the system reverse pumping value PV sys I ,(+) during counter-clockwise rotation was evaluated. 4.2 Results in the RFT-System Measurements (figure 12) with this setup show a similar development of the hydrostatic pressure in the primary chamber as for the RSS. The tests shown here started at approximately 5,000 Pa. This is the upper level of the riser pipe. Small deviations in the starting values result from very small differences in the filling level. The decline in pressure is a proof for the existence of the reverse pumping effect in the RFT. A comparison of a RSS and the RFT using identical material pairings under identical tribological test conditions (i.e. load, speed, oil sump temperature, lubricant, surface finish) reveals, that the reverse pumping values can even be quantified with small deviations compared to RSS. Two fresh samples for every system have been evaluated. Every measurement at every speed was repeated twice. The comparison to the RSS measurements shows, that under identical operating conditions the RFT has -25 %, respectively -18 % lower reverse pumping values, than the RSS-system. It is very noticeable, that within the research project FVA 578 II [39] a difference in planar wear between the two systems of the same amount about +28 % (RFT to Aus Wissenschaft und Forschung 31 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 Housing Bore Chamber- 1 Auxiliary Seal Drive Shaft Chamber - 2 Cover Plate PT-100 Pressure Sensor Riser Pipes Ring Sample Conical Shaft Load Unit Figure 11: Exploded cross sectional view of the two-chamber-principle test setup for the reverse pumping value measurement of elastomer ring samples used in the tribologically equivalent system for radial shaft seals (the ring cone tribometer). The ring sample can be statically loaded with a central load unit based on springs. TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 09 Seite 31 The RFT results provide novel information for the further understanding of the tribological processes in the tribologically equivalent system. A comparison between the tribologically equivalent system and the RSS-system has shown, that the reverse pumping of both systems is in a similar order of magnitude. This offers a lot of further potential to use the tribologically equivalent system for further investigation on elastomer material. 5 Acknowledgement Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) - Project-ID: 172116086 (SFB 926) - SFB 926, MA 6719/ 1-1 and KO 1220/ 26-1. Nomenclature Abbreviation Meaning Unit d RP shaft radius m g gravity constant m/ s 2 K absolute of the incline of the pressure slope Pa/ s PV reverse pumping value µl/ m v sliding speed m/ s ρ lubricant density kg/ m 3 Aus Wissenschaft und Forschung 32 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 RSS) in the same material combination (PAO / FKM / AISI 5115) after a 5,000 km endurance test was reported. These wear results were obtained prior to an optimization of the contact angle, which has not been applied yet in the measurements presented here. Wear and reverse pumping values seem to have a certain relation. With the adaption of the geometry of the ring sample and the coned shaft to the desired RSS geometry, statements about the reverse pumping values could possibly be made in an early stage of material or design development using the RFT. The possibilities of this test rig could also be used to set up systems with a desired reverse pumping values, like neutral. 4 Summary A device for the online-determination of reverse pumping values with very fine pressure sensors based on the two chamber principle by B RITZ [12] was introduced for a RSS-system and a tribologically equivalent test setup. With the presented device reverse pumping values of radial shaft seal systems for different lubricants and shaft surfaces were successfully determined. Investigations have confirmed, that the reverse pumping values are almost independent of the relative speed if normalized with respect to the sliding distance, but are clearly impacted by base oil type and viscosity. With a small modification on the test rig towards the tribologically equivalent system for RSS, the existence of the pumping mechanism in the tribologically equivalent system were also confirmed and for the first time quantified. Figure 12: Left: Comparison of the development of the hydrostatic pressure in the fully filled test chamber for shaft seal and ring sample system under identical conditions. Combination: FKM / ground shaft / PAO / 60 °C. Right: Quantitative comparison of the reverse pumping values in the two formerly mention systems. TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 09 Seite 32 AISI American Iron and Steel Institute CHD case hardness depth CNC computerized numerical control DFG Deutsche Forschungsgemeinschaft DIN Deutsches Institut für Normung FEM finite element method FKM fluorelastomer FVA Forschungsvereinigung Antriebstechnik MIN mineral oil PAO polyalphaolefin oil PG polyalkylene glycol oil RFT ring cone tribometer RSS radial shaft seal Literatur [1] Prem, E. u. Vogt, R.: Der Simmerring. Zuverlässigkeit von Beginn an. Grundlagen zur Schadensprävention. https: / / www.gupta-verlag.de/ nachrichten/ firmenschriften / 6093/ grundlagen-zur-schadenspraevention, abgerufen am: 02.01.2017 [2] Horve, L. A.: Shaft Seals for Dynamic Applications. Mechanical Engineering (Marcel Dekker, Inc.), Bd. 107. New York [u.a.]: Marcel Dekker 1996 [3] Baart, P., Lugt, P. M. u. Prakash, B.: Review of the lubrication, sealing, and pumping mechanisms in oiland grease-lubricated radial lip seals. Proceedings of the Institution of Mechanical Engineers, Part J: Journal of Engineering Tribology 223 (2009) 3, S. 347-358 [4] Kammüller, M. u. Müller, H. K.: Physikalische Ursachen der Dichtwirkung von Radial-Wellendichtringen. ATZ Automobiltechnische Zeitschrift 88 (1986), S. 39-45 [5] Kammüller, M.: Zur Abdichtwirkung von Radial-Wellendichtringen. Zugl.: Stuttgart, Univ., Diss., 1986. Berichte aus dem Institut für Maschinenelemente und Gestaltungslehre, Nr. 20. Stuttgart: Inst. für Maschinenelemente u. Gestaltungslehre 1986 [6] Horve, L.: The correlation of rotary shaft radial lip seal service reliability and pumping ability to wear track roughness and microasperity formation. SAE Transactions (1991 (Cars)) 910530, S. 620-627 [7] Jagger, E. T.: Rotary Shaft Seals: The Sealing Mechanism of Synthetic Rubber Seals Running at Atmospheric Pressure. Proceedings of the Institution of Mechanical Engineers 171 (1957) 18, S. 597-616 [8] Symons, J. D.: Elastohydrodynamic Sealing Systems. SAE Technical Paper Series Paper 730049 (1973) [9] SAE 1002; 1973. Seals-Evaluation of Elastohydrodynamic [10] Hermann, W. u. Seffler, H. W.: Neue Erkenntnisse für den Abdichtmechanismus von Radial-Wellendichtringen. New Knowledges Regarding Sealing Mechanisms of Oil Seals. ATZ-Automobiltechnische Zeitschrift 87 (1985) 9, S. 475-484 [11] Kawahara, Y. u. Hirabayashi, H.: A Study of Sealing Phenomena on Oil Seals. A S L E Transactions 22 (1979) 1, S. 46-55 [12] Britz, S.: Ein Beitrag zur Erfassung der Funktionsprinzipien dynamischer Wellendichtungen unter besonderer Berücksichtigung des Radialwellendichtrings, TU Kaiserslautern Dissertation. Kaiserslautern 1988 [13] Kawahara, Y., Abe, M., Hirabayashi, H. u. Matsushima, A.: Effect of Surface Condition of Lip on Sealing Phenomena of Oil Seals. SAE Technical Paper (1978) 780405 [14] R. Vogt and D. E. Johnston: The Sealing Performance of Elastomer Rotary Lip Seals on Turned Shafts. JOURNAL OF PASSENGER CARS 107 (1998), S. 1067-1076 [15] Schuler, P.: Einfluss von Grenzflächeneffekten auf den Dichtmechanismus der Radial-Wellendichtung, Universität Stuttgart Phd. Thesis. Stuttgart 2014 [16] Fritzsche, R.: Ein Beitrag zur Untersuchung der Verhaltens von Radialwellendichtringen aus Elastomer- und Polytetrafluoräthylen-Material bei speziellen Betriebsbedingungen., Technische Universität Kaiserslautern Dissertation. Kaiserslautern 1994 [17] Ruhl, C.: Ein Beitrag zur Wirkungsweise von Radialwellendichtungen unter Berücksichtigung von rauheitsinduzierter Hydrodynamik und radialen Verlagerungen der Gegenlauffläche. Zugl.: Kaiserslautern, Univ., Diss., 2001. Fortschritt-Berichte VDI Reihe 1, Konstruktionstechnik, Maschinenelemente, Bd. 339. Düsseldorf: VDI- Verl. 2001 [18] Kunstfeld, T.: Einfluss der Wellenoberfläche auf das Dichtverhalten von Radial-Wellendichtungen, Universität Stuttgart Phd. Thesis. Stuttgart 2005 [19] Einfluss der Wellenlaufflächen-Topographie auf das Dichtsystem der Radialwellendichtung, Abschlussbericht, Oberflächentopographie: Vorhaben Nr. 269. Forschen im Maschinenbau Nr. 297, Schmuker, S., Frankfurt am Main 2007 [20] Klaiber, M.: Betriebs- und Benetzungseigenschaften im Dichtsystem Radial-Wellendichtung am Beispiel von additivierten synthetischen Schmierölen, Universität Stuttgart Pdh. Thesis. Stuttgart 2013 [21] Klaiber, M., Bauer, F. u. Haas, W.: Compatibility of Oil Additives and Elastomeric Radial Lip Seals. Tribologie + Schmierungstechnik 60 (2013) 3, S. 60-68 [22] Remppis, M.: Untersuchungen zum Förderverhalten von Dichtsystemen mit Radial-Wellendichtringen aus Elastomer, Universität Stuttgart Dissertation. Stuttgart 2016 [23] Berechenbare Dichtgüte II. Rechnerische Abschätzung der Dichtgüte von Radial-Wellendichtringen durch Kenntnis der Systemparametereinflüsse II, Forschungsvorhaben Nr. 617 II, FVA 617 II, 2017 [24] Totz, J., Baumann, M. u. Bauer, F.: Einfluss der Konditionierung auf die Dichtfunktion. Antriebstechnik (2019) 4, S. 147-153 [25] Salant, R. F.: Elastohydrodynamic Model of the Rotary Lip Seal. Journal of Tribology 118 (1996) 2, S. 292-296 Aus Wissenschaft und Forschung 33 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 09 Seite 33 [37] Deutsche Norm DIN 3761 Part 2; 1983. Rotary shaft lip type seals for automobils; applications, abgerufen am: 25.08.2017 [38]Verschleißu. Reibungsprüfung. Vergleichende Reibungs- und Verschleißuntersuchungen durch Experiment und Simulation an Elastomeren für Dichtungsanwendungen in der Antriebstechnik, Forschungsvorhaben Nr. 578 I. FVA Heft Nr. Nr. 979, FVA 578 I, 2011 [39] Praxistaugliches RWDR-Triboersatzsystem. Praxistaugliche Prüfmethodik für Reibungs- und Verschleißuntersuchungen am tribologischen Ersatzsystem von RWDR, Forschungsvorhaben Nr. 578 II. FVA Heft Nr. N.A., FVA 578 II, 2019 [40] Burkhart, C., Emrich, S., Magyar, B., Kopnarski, M. u. Sauer, B.: Nachbildung und Analyse von Schadensmechanismen bei Dichtringen im tribologischen Ersatzsystem. Tribologie + Schmierungstechnik (2019) 4-5, S. 9- 18 [41] Jia, X., Jung, S., Haas, W. u. Salant, R. F.: Numerical simulation and experimental study of shaft pumping by plunge ground shafts with rotary lip seals. Tribology International 48 (2012), S. 155-161 [42] Raab, H. u. Haas, W.: Tribologische Partner: Radialwellendichtring und Gegenlauffläche. Antiebstechnik 38 (1999) 4, S. 133-135 [43] Thielen, S., Breuninger, P., Hotz, H., Burkhart, C., Schollmayer, T., Sauer, B., Antonyuk, S., Kirsch, B. u. Aurich, J. C.: Improving the tribological properties of radial shaft seal countersurfaces using experimental micro peening and classical shot peening processes. Tribology International (2020), S. 106764 [44] Estorff, O. von u. Wollesen, V.: Ein Tribometer zur Nachbildung der tribologischen Verhältnisse im Dichtzonenbereich von Radialwellendichtingen. Beitrag 05. In: Radialwellendichtringe: VI. Hamburger Dichtungstechnisches Kolloquium. Hamburg: Techn. Univ 2006 [45] Frick, A., Muralidharan, V. u. Borm, M.: A novel rotational cone tribometer to approach the tribological contact between polymers and a rotational shaft. 57. Tribologie Fachtagung. 2016, S. 1-10 [46] Burkhart, C., Weyrich, D., Magyar, B., Sauer, B.: Experimental Determination and Comparison of the Contact Temperature of Radial Shaft Seals and its Derived Tribological System. International Sealing Conference. 2018, S. 197-218 Aus Wissenschaft und Forschung 34 Tribologie + Schmierungstechnik · 67. Jahrgang · 5-6/ 2020 DOI 10.30419/ TuS-2020-0025 [26] Matus, M., Pelzer, V., Wennehorst, B., Poll, G.: Continuous Measurement of the Reverse Pumping Effect of Radial Shaft Seals. International Sealing Conference. 2018 [27] Poltavchenko, V.: Horizontalförderprüfstand, Hannover. https: / / www.imkt.uni-hannover.de/ fileadmin/ imkt/ pdf/ Horizontalfoerderpruefstand.pdf, abgerufen am: 28.10.2019 [28] Wälzlagerdichtungen. Schutzdichtungen für Wälzlager, Forschungvorhaben FVA Nr. 432 I, FVA 432 I, 2008 [29] Ottink, K.: Betriebsverhalten von Wälzlagerschutzdichtungen - Experimentelle Untersuchungen und Berechnungsansätze, Gottfried Wilhelm Leibniz Universität Hannover Phd. Thesis. Hannover 2014 [30] Burkhart, C., Emrich, S., Kopnarski, M. u. Sauer, B.: Excessive shaft wear due to radial shaft seals in lubricated environment. Part I: Analysis and mechanisms. Wear 460- 461 (2020), S. 203419 [31] DIN 3760; 1996. Rotary shaft lip seals (withdrawn) [32] Burkhart, C., Thielen, S., Peter, K. u. Sauer, B.: Online determination of reverse pumping values of radial shaft seals and their tribologically equivalent systemlip seals. In: Fatemi, A. (Hrsg.): 22th International Colloquium on Tribology 2020. 28. to 30. January 2020, Stuttgart/ Ostfildern. Ostfildern: Techn. Akad. Esslingen 2020, S. 89-90 [33] Macherauch, E. u. Zoch, H.-W.: Praktikum in Werkstoffkunde. 91 ausführliche Versuche aus wichtigen Gebieten der Werkstofftechnik. Studium. Wiesbaden: Vieweg+ Teubner Verlag / Springer Fachmedien Wiesbaden GmbH Wiesbaden 2011 [34] Deutsche Norm DIN 3761 Part 2; 1983. Rotary shaft lip type seals for automobils; applications, abgerufen am: 25.08.2017 [35] Thielen, S., Magyar, B., Sauer, B., Schneider, F., Mayer, P., Kirsch, B., Müller, R., Harbou, E. v. u. Aurich, J. C.: Functional investigation of zero lead radial shaft seal counter-surfaces turned with a special method. Tribology International (2017) [36] Frölich, D., Magyar, B., Sauer, B., Mayer, P., Kirsch, B., Aurich, J. C., Skorupski, R., Smaga, M., Beck, T. u. Eifler, D.: Investigation of wear resistance of dry and cryogenic turned metastable austenitic steel shafts and dry turned and ground carburized steel shafts in the radial shaft seal ring system. Wear 328-329 (2015), S. 123-131 TuS_5_6_2020.qxp_TuS_Muster_2020 09.12.20 16: 09 Seite 34