Assessing workability of greased bearings after long-term storage

Here, we developed a technique to assess the workability of sealed-for-life greased rolling bearings after a long-term storage. In this framework, we devised a model of equivalent transition between the conditions of natural ageing under daily and seasonally fluctuating temperature, and the conditions of accelerated thermal ageing at a constant high temperature. The tested bearings were thermally aged, and then their steady state friction and outer ring temperature were examined in a custom high-speed spindle. These results were compared to the performance of a reference new bearing tested under the same loading conditions. Our findings suggest that long-term storage can significantly degrade the performance of sealed-for-life greased rolling bearings. However, a proper running-in can substantially deter the ageing-driven degradation of the bearings.


Introduction
The presence of a lubricant layer separating rolling elements, cages and raceways is a vital requirement for long-life and high-efficiency operation of rolling bearings. The most common lubricant applied today to rolling bearings is grease [1,2], in part because replacing oil with grease can reduce sealing problems and enable simpler designs to be pursued.
Lubricating grease is a complex system consisting of oil, several additives and thickener. The actual lubrication is mainly provided by the base oil, which can be either mineral or synthetic. The additives are compounds used to improve both bulk and surface properties of the oil, such as resistance to oxidation [3] and interaction with a solid component surface defining the lubrication efficiency in a boundary lubrication regime [4]. The thickener consists of metallic soap molecules creating a fibre-based network structure [1,3], which is able to hold the oil inside by a com-bination of physical and chemical forces [5,6] and to slowly release it into the contact over a long period. The mechanism of the oil release is, however, still unclear, and there are works suggesting that the lubricating film is formed as a result of oil bleeding [7−9] or destruction of the soap matrix [10−12].
There are many applications, such as low-noise bearings, high-precision bearings or high-speed spindle bearings, that may require lifetime lubrication [13]. In such cases, bearing service life is defined by the grease degradation over time. This degradation (or ageing) manifests itself in a mechanical destruction of grease molecules within the contact zone, and in temperatureaccelerated chemical changes that lead to the appearance of various deposits and oxidation products, to the alteration of viscosity and lubricity, and to oil bleeding, affecting both friction and wear [14−17]. Interestingly, though there were successful attempts to predict maximum period of reliable service of grease [18], due to a high complexity of the processes involved, the effects of ageing are determined today solely based on experimental techniques.
The methods used can be generally subdivided into two categories. The first includes the tests of mechanically and thermally loaded greased bearings that are run until failure [19−21], which is usually determined by a certain change in the bearing temperature or the drive motor torque. Grease samples are taken from the different locations in the bearing both before and after failure, and changes in viscosity and chemical composition are assessed [22,23]. The second category involves laboratory simulations of both mechanical and thermal ageing of grease [12,16,24]. Mechanical ageing can be modelled by forcing the lubricant through a sub-millimetre diameter capillary repeatedly [14]. Thermal ageing is most influenced by oxidation [15], and, in doing accelerated tests, the grease is heated in contact with actual solid, which may serve as a catalyst [1], while, as is generally accepted, a rise of 10 °C effectively doubles the rate of oxidation [15,25,26]. Laboratory simulations, however, are rarely analysed in relation to lubrication performance in real bearings [12,15]. Another topic that did not receive much research attention is the effect of grease ageing on the operation of mechanisms designed for a very long storage [27].
In light of the above, the purpose of this work was to develop a technique to assess the workability of sealed-for-life greased rolling bearings after a long-term storage.

Model of equivalent transition
Finding the correct conditions of accelerated ageing that is equivalent to a real storage requires taking into account the daily and seasonal temperature fluctuations in the storage area. Here, we show how a proper equivalent transition between the conditions of accelerated and normal ageing can be devised based on a generally acknowledged rule that a rise of 10 °C effectively doubles the rate of oxidation [15,25,26].
If we view ageing as an oxidative degradation reaction, and further assume that an incremental amount of reaction products, dP, forms during the increment of time dt at a reaction speed V(T), when T is some given temperature, then we can write: The total amount of reaction products, t P , formed during the storage time, s t , is then given by Based on the above rule of chemical reaction acceleration, the reaction speed , where k is acceleration factor, and r T is some reference temperature. This allows us to write a general expression for the reaction speed: Now, let us consider the two cases of accelerated and normal ageing, in which an equal total amount of reaction products, t P , should form. In accelerated ageing, t P is expected to form at the constant equivalent (relatively) high temperature  e T T held during the short (accelerated) period of time  In normal ageing, t P forms at the time-dependent (relatively) low temperature  ( ) T T t fluctuating during the long (normal) period of time  s n t t . In this case, Eq. (4) becomes r n ( ) Finally, combining Eqs.
Thus, knowing the daily and seasonal temperature fluctuations, ( ) T t , and the duration of normal storage, |www.Springer.com/journal/40544 | Friction http://friction.tsinghuajournals.com n t , we can choose an equivalent temperature e T and determine the time a t needed for equivalent accelerated ageing.

Evaluation of ageing conditions
To determine the conditions for accelerated ageing, we used the data on the characteristic daily temperature fluctuations in a desert, where hot and cold seasons last for 75% and 25% of a year, respectively (Fig. 1). The storage time was set to be 17 years, and the acceleration factor k was chosen to be equal to 2 according to what is accepted in testing grease ageing [15].
Given that the sealed-for-life greased rolling bearings used in this work (see Section 3.1 for details) are limited to the temperature of 149 °C , we performed accelerated ageing at the equivalent temperature, T e , of 143 °C to work at sufficiently high but safe temperature. Plugged into Eq. (7), these data yielded a total accelerated ageing time, t a , of 105 hours (Fig. 2).

Experimental details 3.1 Bearings and test conditions
The bearings used in this work were three deep groove spindle and turbine bearings C9201FFT6 GJ-284 2850-3050 (ceramic balls, d=12 mm, D=32 mm, B= 15.9 mm, Kluber Asonic HQ 72-102 grease functional at -40 °C −180 °C , The Barden Corporation Ltd., Plymouth, UK) able to operate at speeds of up to 71 krpm and at temperatures of up to 149 °C (cage limitations). Two bearings were artificially aged in order to simulate a 17-year-long storage before testing their frictional performance, with one of them being also run-in before ageing (about half an hour at 30 N and 20 krpm, in the frame of the system adjustments). The accelerated ageing was done by keeping the bearings in a furnace at a constant temperature of 143 °C for 105 hours (see Section 2 for details). In order to provide enough oxygen for the oxidation-driven ageing, the atmosphere inside the furnace was continuously renewed. The third bearing served as a reference and was tested without any preliminary treatment. All tests were performed under temperature and relative humidity conditions of 20−22 °C and 45−55%, respectively.

Equipment
The tests were performed with a rig (Fig. 3) based on a high-speed motor MFM-10120/11 HJND-60 (Fischer AG, Herzogenbuchsee, Switzerland) and a custom spindle we constructed to examine bearings under high loads and speeds. Rotation is transmitted from the motor to the spindle by a mechanical fuse/clutch made of a plastic platelet capable of breaking down if rotational load exceeds the accepted motor torque of 0.3 N·m. The test unit consists of a tested bearing and measurement/loading fixtures designed for easy assembly/disassembly on/from the spindle. The tested bearing is prevented from a friction-driven rotation by a tangential connection to a force gauge, which allows measurement of the friction moment, and is loaded axially using a dead weight applied through a pulley. To provide on-line monitoring of the temperature and the friction moment in the tested bearing, a multifunctional data acquisition board NI-USB-6351 (National Instruments Co., Austin, Texas) is operated by a custom code written using LabVIEW software 492 Friction 7(5): 489-496 (2019) | https://mc03.manuscriptcentral.com/friction package (National Instruments Co., Austin, Texas). The signals going into the data acquisition board are measured using a K-type thermocouple connected to an amplifier ITMA2003 (Red Lion Controls, York, Pennsylvania), and a load cell LRM200 (FUTEK Advanced Sensor Technology, Irvine, California) connected to a strain gage amplifier AE101 (HBM, Darmstadt, Germany).

Test procedure
Each test sequence began with assembling the examined bearing on the spindle, fixing the thermocouple to the outer bearing ring and applying the axial load of 400 N. Then, the tests were run at speeds of 12, 24, 36, 48, 60, 72, 84 and 90 krpm. At each speed the tests continued until steady state conditions of temperature and friction moment were reached. Then the data was saved and the system was accelerated to the next speed level. The tested bearing was considered to have failed when the friction moment and the outer ring temperature started rising sharply. In this case the experiment was stopped. Figure 4 shows a characteristic chart of the friction moment and the outer ring temperature recorded as a function of rotational speed with a brand new bearing. It is evident that each increase in rotational speed results in the temperature rise. The reason for this is that, regardless of the main mechanism of energy dissipation, be it either sliding of the cage against the other bearing components or a viscous deformation of the grease [28], more heat is generated when the rolling distance increases. Obviously, the bearing that rotates faster moves a greater distance during the same time and, hence, generates more heat.

Results and discussion
The changes in friction moment are not monotonic, with the friction signal fluctuating around the same value during the normal operation of the tested bearing. These fluctuations can be associated with local running-in due to changes in working conditions and with different processes dominating the bearing operation at different speeds, as explained below. The bearing failure seems to be initiated by the temperature increase, which leads to a steep rise in friction when  the temperature reaches the threshold value of about 150 °C that is specified as the highest allowed temperature by the bearing manufacturer. At the time of failure, both the temperature and the friction moment grow rapidly until either the test rig is stopped or the mechanical fuse/clutch between the motor and the spindle is broken.
The average friction moment and outer ring temperature measured with aged and new bearings in the steady state at different rotational speeds are presented in Fig. 5. Observing the behaviour of the reference new bearing, we see that friction first decreases and then starts increasing with increasing speed. This may be explained in the following way. At low speeds, such reasons as the temperature-driven reduction of the grease viscosity or a growing centrifugal force and the related cage deformation leading to a decrease of the contact area between the balls and the cage may dominate the frictional behaviour. At high speeds, the effect of constantly growing temperature and corresponding thermal expansion leading to the growth of the contact area and more intensive frictional interactions between the bearing components may become more significant, with a temperature of about 50 °C being a threshold for the change in behaviour. Along this line of thought, it should be mentioned that the presence of an external heat source might also have a tremendous effect on the bearing behaviour. We have learned during preliminary tests that, at the transition from 48 to 60 krpm, the front spindle bearing, whose housing was designed to have a sliding fit, slightly moves forward. This leads to a significant relaxation of the stresses generated due to a thermal expansion of the shaft and a subsequent decrease in the spindle bearings' temperature. Since the test bearing is assembled on the same shaft at the distance of only 25 mm from the front spindle bearing, both friction moment and temperature of the tested bearing are also affected significantly by this relaxation as can be seen in Fig. 5.
The behaviour of the aged new bearing differed significantly from that of the reference new bearing. The temperature increased much more steeply with the rotational speed, so the aged new bearing came to the critical temperature of about 160 °C already at 48 krpm and failed at 60 krpm, not being able to reach its maximum designed operational speed of 71 krpm. The friction moment of the aged new bearing was lower than that of the reference bearing at the low rotational speed, but, being governed by the sharply rising temperature, surpassed the reference values already at 36 krpm, well before approaching its speed limit of 60 krpm. Given that most of the mechanical energy expended in bearings is dissipated through the  | https://mc03.manuscriptcentral.com/friction cage sliding and the lubricant film deformation [28], this behaviour is clearly associated with the grease ageing, as a result of which oil molecules are damaged via carbon-carbon chain scission due to thermal decomposition [15,22,23]. These changes lead to lower temperature and friction at low speeds, but interfere with the grease's ability to protect surfaces from direct contact at higher speeds. Presumably weaker bonds between the oil molecules and the working surface let the protective grease layer be removed more easily, which results in earlier bearing failure under extreme loading.
A very interesting result was obtained with the bearing that was run-in under low loads and speeds before ageing. Surprisingly, despite being aged, this bearing demonstrated similar temperatures and friction moments to those of the reference new bearing, and was even able to operate at higher speeds, eventually failing at 90 krpm as opposed to 84 krpm, at which the reference bearing failed. This result can probably be explained by noting that the churning of excess grease and the smoothening of raceways [29] are further supported by the mechano-chemical modification of the topmost surface layers that can also take place during running-in. The mechano-chemically modified layers can be less sensitive to the oil composition [30] and can allow for a proper operation even with a significantly degraded grease. The latter assumption, however, is yet to be verified, though the main finding remains the same: new greased-for-life bearings have to be run-in [31] before they go to a long-term storage.

Conclusion
The paper shows that it is possible to find a consistent equivalent transition between the conditions of natural ageing under daily and seasonally fluctuating temperature, and the conditions of accelerated thermal ageing at a constant high temperature. The test results obtained with artificially aged and reference new bearings suggest that long-term storage can significantly degrade the performance of sealed-for-life greased rolling bearings. It is also evident that a proper running-in performed prior to the long-term storage can substantially deter the ageing-driven degradation of the greased bearings.