velocity limits for various seals under standard conditions as a function of the shaft diameter and the seal material, e.g. [1]. However, the temperature at the seal depends very much on the heat dissipation possibilities via the shaft and, if possible, via the oil bath. If there are additional heat sources such as bearings near to the sealing area, these also has to be considered. Müller [2] provides an initial estimate of the temperature increase Δϑ at the seal contact as a function of the circumferential velocity v by (1) Leichner [3] used this approach to estimate temperature increases in the seal contact of approx. 15 K at 1500 rpm and 30 K at 3000 rpm for a shaft diameter of 78 mm. With these contact temperatures he simulates the temperature distributions in the seal using a thermal FE model. The results were temperature distributions with circular ring-like isotherms around the contact in the cross-section of the seal. In a stationary state, the heat generated in contact is divided into a heat flow through the shaft and a heat flow through the seal. However, due to the filigree cross-section of the seal and the poor thermal conductivity of elastomers, only a small proportion of the heat flow passes through the seal. Engelke [4] has assumed the distribution in relation to the thermal conductivity ratio. The thermal conductivity coefficient of steel is approx. 40 W/ m/ K and that of elastomers approx. 0.2 W/ m/ K. y = 2.5 Ks m . Science and Research 26 Tribologie + Schmierungstechnik · volume 71 · issue 3/ 2024 DOI 10.24053/ TuS-2024-0014 1 Introduction Sealing applications are critical if the seals are used at high temperatures. High velocities in the seal contact cause high friction losses and lead to a localised temperature rise on the seal contact and in the lubricating film. Therefore, dynamic shaft seals in particular quickly reach the limits of their operating conditions. If there are also high ambient temperatures, the operating conditions for the sealing application are even more difficult. This can be caused by increased oil bath temperatures or increased ambient temperatures as well as very warm components in the immediate vicinity of the seal, e.g. combustion engines, exhaust systems or process engineering reactors. Use cases under grease lubrication or minimum oil lubrication are also critical due to the lack of heat dissipation through an oil bath. This results in a strong selfheating of the seal contact at high velocities. Consequences may be ageing of the seal or the lubricant or the formation of oil carbon. The aim of this publication is to present test options for shaft seals under high temperatures and to discuss their results. 2 State of the art The temperature at the seal is often a limiting factor for sealing applications. Increased temperatures lead to faster ageing of the seal material. The very frequently used sealing material acrylonitrile butadiene rubber (NBR) is particularly sensitive to high temperatures. Therefore, seals made from hydrogenated acrylonitrile butadiene rubber (HNBR) for raised temperatures and seals made of fluororubber (FKM) for high temperatures can be used. The application limits of the seals depend on the temperature and the duration of exposure resp. the temperature profile. An initial indication is provided by diagrams from the seal manufacturers, which specify Rotary shaft seals at high temperatures Matthias Kröger, Jim Gerschler, Ringo Nepp, Christian Berndt* Presented at the GfT Conference 2024 In many applications, dynamic seals are used at high temperatures which can lead to many problems like damage or ageing of elastomer and lubricant or in formation of oil carbon. A recently built seal test rig enables very high temperatures up to above 250 °C with grease lubrication. Experiments show the influence of high temperatures on function and friction of seals. Keywords Dynamic shaft seals, high temperature, friction losses, oil and grease lubrication, test rig, ageing Abstract * Prof. Dr.-Ing. Matthias Kröger M. Sc. Jim Gerschler Dr.-Ing. Ringo Nepp Dr.-Ing. Christian Berndt Institute of Machine Elements, Design and Manufacturing Technische Universität Bergakademie Freiberg Agricolastr. 1, 09599 Freiberg According to Engelke, only approx. 0.5 % of the heat flow is transferred via the seal and over 99 % via the shaft. Engelke [4] and later Ottink [5] also calculate the temperature rise at the seal contact. It amounts to between 10 K and 17 K per 1 W friction power and per 1 mm 2 contact area. Investigations by Berndt [6] with minimum oil lubrication without an oil bath showed a heating of 2.6 K/ W depending on the friction power of a radial shaft seal without spring with a diameter of 45 mm for his test setup. Estimating a contact area of approx. 13 mm 2 , Berndt’s seal contacts heat up two to three times more than Engelke’s resp. Ottink’s seal contact. In particular, this is due to the thermodynamic boundary conditions of the different test set-ups. The friction of seals primarily consists of fluid friction and hysteresis friction. At very low velocities or start-up processes, adhesive friction also has to be taken into account, but this isn’t considered here, cp. [7, 8]. By increasing the temperature in the contact, the most important parameters for fluid friction and hysteresis friction are strongly shifted. As the temperature increases, the dynamic viscosity η of the lubricant decreases by powers of ten, see Figure 1 left. This makes it more difficult to build up the lubricating film and results in smaller lubricating film thickness h resp. the lubricating film thickness in relation to the standard deviation of roughness, see Figure 1 centre. This results in higher shear rates in the lubricating film. Assuming a Newtonian fluid, the shear stress τ in the fluid film for a circumferential velocity v is calculated as (2) = . It can be seen that a decrease in viscosity η directly reduces the shear stress in the fluid and thus the fluid friction torque. However, the decrease in lubricating film thickness increases the shear stress. The extremely strong decrease in viscosity dominates at the rotary shaft seal, so that the frictional torque decreases with increasing temperature T, see Figure 1 on the right. The modelling assumptions on which the results are based can be found in dissertation of Berndt [6]. It should be noted that various constant temperatures were specified for this simulation. In real operation, the frictional torque of the seal leads to increasing frictional power and higher temperatures as the velocity increases. Hysteresis friction is also influenced by the temperature, as the viscoelastic properties of the seal change. Hysteresis friction is caused by local cyclic deformation of the seal at the roughness of shaft and the damping effect of the viscoelastic elastomer material. The relevant frequency range at the increased temperature can be converted into a frequency range at reference temperature by means of the temperature-frequency equivalence using the WLF transformation of Williams, Landel and Ferry [9], see Figure 2 left. The figure shows the lower cut-off frequency (1 and 2) resp. the upper cut-off frequency (3 and 4) for 20 °C in blue (2 and 4) and 100 °C in red (1 and 3). For the analysed ground counterface, the hysteresis friction decreases with increasing temperature, see Figure 2 on the right. When interpreting the hysteresis friction results, it should be noted that the simulation of hysteresis friction was calculated without lubrication. This means that there is no separating lubricating film considered which reduces hysteresis friction at high velocities. It can be seen that both fluid friction and hysteresis friction decrease with temperature. This does not apply to Science and Research 27 Tribologie + Schmierungstechnik · volume 71 · issue 3/ 2024 DOI 10.24053/ TuS-2024-0014 Friction coefficient [-] Velocity [m/ s] Oil temperature [°C] Oil viscosity [Pa s] Velocity [m/ s] Rel. film thickness [-] Figure 1: Influence of temperature on viscosity of 4 lubricating oils (ISO VG 15, ISO VG 32, ISO VG 100, ISO VG 460) (left), simulated influence of contact temperature on relative lubricating film thickness (centre) and fluid friction coefficient (right) using the example of a radial shaft seal without spring for the lubricating oil ISO VG 100 In the RWDR seal test rig, the frictional torque of the seal is measured using a torque sensor which is connected to the seal via seal support and a thermal insulation to reduce the temperature at the sensor. As the seal support represents the housing and is motionless, there are no inertial forces in the measurement signal, even with rapid speed changes, which would be the case with a shaftattached torque sensor. The test shaft as the counterface of seal is fixed via a clamping element and is mounted on a horizontal spindle. The connection to the motor is given by a clutch and optionally a gearbox. The design allows very good accessibility to the seal and test shaft. The friction torque of the seal is measured as well as the controlled rotation speed up to 6000 rpm without gearbox and the temperature at a point on the seal close to the contact using an infrared thermal sensor. To vary the ambient temperature, a heated thermal box can be fitted over the seal and seal support, whereby the Science and Research 28 Tribologie + Schmierungstechnik · volume 71 · issue 3/ 2024 DOI 10.24053/ TuS-2024-0014 very high temperatures resp. very low sliding velocities, at which no suitable lubricating film builds up and higher viscosities resp. lower temperatures lead to lower frictional torques, cp. Teichert [8]. 3 Test rigs for high temperatures There are many test rigs for shaft seals that seal against an oil bath and can therefore be used for applications such as gearbox seals. Here, the temperature at the seal can be influenced very well by influencing the oil temperature, as the heat is primarily dissipated to the oil via the shaft. In the case of minimum oil lubrication or grease lubrication, the heat must be dissipated to the environment via the shaft and the seal by convection. Then, the ambient temperature must be varied using a thermal box, see Figure 3, to analyse the influence of temperature. Frequency [Hz] Complex modulus [MPa] Friction coefficient [-] Velocity [m/ s] Measurement Simulation Figure 2: Analysis of the influence of frequency resp. temperature on the viscoelastic material properties, here the complex modulus, (left) and hysteresis friction coefficient in dependence of velocity (right) Test head Bearing block Drive Test head Bearing Drive Base Counter part Seal k Seal support Thermal isolation Torque sensor Clutch Motor Ther Thermal box M Bearing block Figure 3: Photographs and sectional view of the structure of the RWDR seal test rig with optional thermal box (dashed line) according to Berndt [6] torque sensor remains outside the thermal box and is therefore not heated. The thermal box is indicated by a dashed line in Figure 3 and enables ambient temperatures of up to 120 °C. By measuring the temperature at the torque sensor, a temperature correction of the measured values is implemented. For higher temperatures up to above 250 °C, a new hightemperature seal test rig has been developed that realises the temperature via a thermal chamber with adapted control, see Figure 4. Here, the drive is connected via a toothed belt and a long vertical shaft that extends into the thermal box from below. The application-specific test specimen has an axial running shaft seal and is mounted above the shaft. The torque support of the seal is provided by a long casing around the shaft and is connected to the torque sensor outside the thermal box via a thermal insulation. Holes in the casing and suitable insulation ensure that the shaft and casing expand to approximately the same extent when heated. As a result, the overlap and the axial preload of the axial running seal remain largely unchanged. The design can also be used for radial shaft seals with minor modifications of seal support and conterface. Due to the long shaft with bearings only outside the thermal chamber, operation is possible at low and medium rotation speeds below the critical rotation speed and precise balancing of the shaft is necessary. 4 Experimental investigations Using the thermal box with the first test rig, Figure 3, tests were carried out on radial shaft seals (RWDR) without springs at ambient temperatures of 25 °C, 40 °C and 60 °C. Figure 5 top shows for the 3 ambient temperatures the temperatures at the seal which are mainly shifted in parallel. The friction-specific heating can be added to the ambient temperature for a rough estimate. However, a closer look shows slight differences in the distances between the three coloured curves. Three lubricating oils with very different viscosity classes (ISO VG 32, ISO VG 100, ISO VG 460) were analysed. The corresponding measured friction coefficients, see Figure 5 bottom, show a strong change in the friction characteristics with temperature, especially for the oil (ISO VG 460) with high-viscosity. A first example of the investigation on the high-temperature seal test rig is shown in Figure 6 using an axially running seal. In this test the seal run without lubricant. After running in for 120 minutes at room temperature without heating, the thermal box was heated up from 140 °C to 210 °C in 10 °C steps. For this, a rotation speed of 60 rpm with periodically changing direction of rotation was set. The alternating direction of rotation was used to compensate the temperature-dependent zero-point drift of the torque sensor, assuming equal friction torques in both directions. For this purpose, the mean value of the measured friction coefficient in both directions was subtracted from the friction torque signal. Alternatively, the sensor can be calibrated as a function of the temperature in the thermal chamber. It was found that the friction torque decreased significantly when heated to 140 °C for the axially running shaft seal used. Further heating led to an additional decrease in friction torque. At the thermal chamber temperatures of 200 °C and 210 °C, the frictional torque also Science and Research 29 Tribologie + Schmierungstechnik · volume 71 · issue 3/ 2024 DOI 10.24053/ TuS-2024-0014 Seal Casing Shaft Thermal chamber Torque sensor Drive system Thermal isolation Figure 4: Photograph and sectional view of the structure of the high-temperature seal test rig 5 Conclusion Elevated temperatures are particularly critical in the case of minimum oil lubrication or grease lubrication. The influence of both moderately increased ambient temperatures on the sealing contact temperature and the seal Science and Research 30 Tribologie + Schmierungstechnik · volume 71 · issue 3/ 2024 DOI 10.24053/ TuS-2024-0014 initially decreases during the 60-minute measurement at a constant temperature, but then increased again. This is an indication of changes in the tribological system and may be caused by changes in the seal material at the high temperatures. These can be investigated in more detail by analyses after the test run. Velocity [m/ s] Friction coefficient [-] Temperature [°C] Velocity [m/ s] Velocity [m/ s] Velocity [m/ s] Velocity [m/ s] Velocity [m/ s] Friction coefficient [-] Temperature [°C] Friction coefficient [-] Temperature [°C] Figure 5: Temperature (top) at the seal contact and coefficient of friction (bottom) of a radial shaft seal without spring at three ambient temperatures (blue: 25 °C, green: 40 °C, red: 60 °C) for three different oil viscosities (left: ISO VG 32, centre: ISO VG 100, right: ISO VG 460) according to Berndt [6] Figure 6: Frictional torque (only maxima and minima of torque are shown) and controlled temperature curve of a temperature rise test with cyclic changing rotation directions on the high-temperature seal test rig of an axially running shaft seal, here tested without lubricant friction as well as greatly increased temperatures on the frictional torque could be investigated using two specially test rigs. Several influences on the tribological effects of the temperature were shown. The friction torque curves showed initial changes of the seal properties at high temperatures above 200 °C. Literature [1] Freudenberg Sealing Technologies GmbH: Technisches Handbuch, Kapitel 1: Simmerringe und Rotationsdichtungen, 2015. [2] Heinz K. Müller: Abdichtung bewegter Maschinenteile: Funktion - Gestaltung - Berechnung - Anwendung. Waiblingen: Medienverlag Müller, 1990. [3] Tim Leichner: Prognose der Dichtlippenfolgefähigkeit von RWDR bei dynamisch verlagerter Welle. Maschinenelemente- und Getriebetechnik-Berichte 9. Technische Universität Kaiserslautern, 2012. [4] Tobias Engelke: Einfluss der Elastomer-Schmierstoff- Kombination auf das Betriebsverhalten von Radialwellendichtringen. Dissertation. Leibniz Universität Hannover, 2011. [5] Kathrin Ottink: Betriebsverhalten von Wälzlagerschutzdichtungen: Experimentelle Untersuchungen und Berechnungsansätze. Dissertation. Leibniz Universität Hannover, 2014. [6] Christian Berndt: Tribologie von Radial-Wellendichtungen. Dissertation. Technische Universität Bergakademie Freiberg, 2019. [7] Ringo Nepp: Experimentelle und theoretische Adhäsionsanalysen an Reifen-Profilklötzen und O-Ring-Dichtungen. Dissertation. Technische Universität Bergakademie Freiberg, 2017. [8] Robert Teichert: Tribologie von fettgeschmierten Radialwellendichtungen. Dissertation. Technische Universität Bergakademie Freiberg, 2022. [9] Malcolm L. Williams, Robert F. Landel, John D. Ferry. The temperature dependence of relaxation mechanisms in amorphous polymers and other glass-forming liquids. Journal of the American Chemical Society 77.14 (1955), pp. 3701-3707. Science and Research 31 Tribologie + Schmierungstechnik · volume 71 · issue 3/ 2024 DOI 10.24053/ TuS-2024-0014