A New Test Rig to Study Rolling Element Bearing Thermomechanical Behavior

Release Time：12 Jul,2019

Introduction

    The rolling element bearing (REB) is an essential component in mechanical transmission to reduce friction between rotating parts. Now, with the development of the electrical motor in mechanical industry, REBs may work at very high-rotation speed. It leads to an increase of REB power losses and temperatures. In literature several approaches are used to estimate the REB thermomechanical behavior. Hence, for some applications, several divergent viewpoints can be found on thermal behavior modeling according to the dissipation sources taken into account. In order to overcome these discrepancies, a specific test rig was designed to obtain information on REB thermomechanical behavior. This test rig allows for obtaining REB power losses thanks to torque measurement. To study high-speed applications, the (N.dm) product expected volume is higher than one million.

    Rolling element bearings are widely used in mechanical transmission to reduce friction between two rotating parts. With the further development of the electrical motor in mechanical industry, REBs operate more and more at high rotational speed. For these applications, REBs power losses can be predominant in mechanical transmissions. Several global models can estimate the REB resistive torque (Refs.1–2). Hence, for some applications, several divergent viewpoints can be found between these models. In order to overcome these discrepancies, some measurements are required. In the literature, several tests rigs are presented to measure REB torque loss. A first set of test rigs measure torque loss on REB outer ring. In this case, the REB outer ring has to be mounted in the inner ring of a hydrostatic bearing. The torque is measured via a load sensor located on a beam between the REB outer ring and hydrostatic bearing housing (Ref.3). Brecher et al. (Ref.4) used a hydrostatic bearing to measure the REB torque loss for high-speed application. In this test rig a telemetry system is used to measure the REB inner ring temperature. Neurouth et al. (Ref.5) have also used this design to measure the REB frictional torque for grease-lubricated thrust ball bearings. However the hydrostatic bearing can be complex to use and modifies the outer ring thermal behavior. REB torque loss can also be measured with strain gauges located on the housing. Hannon (Ref.6) developed a test rig for four types of REBs within a similar size range. A slip ring allows measuring the REB inner ring temperature. In this test rig, four identical REBs are mounted on the main shaft. The global torque loss is divided by four in order to obtain the REB torque loss. The REBs’ torque loss is measured thanks to a strain gauged torque table. This test rig has been developed for low-rotational speed and strong radial load conditions (up to 260kN). Pinel et al. (Refs.7–8) developed a test rig for a 35mm bore diameter angular-contact ball bearing under thrust load and for very high-speed application; the maximum rotational speed is equal to 72,000rpm, which corresponds to a (N.dm) product equals to 3.4 million. The REB torque is measured with strain gauges located near the end of an arm that prevents the housing from rotation. However, this measurement can be complex to realize. Finally, REB torque loss can be measured on the inner ring by using a torquemeter. In this case, the torquemeter measures the global torque of the shaft. However, some components can affect this measurement (seals, REBs mounting, etc.). Takabi et al. (Ref.9) designed an REB test rig to study the deep-groove thermal behavior in oil bath lubrication. The test rig is composed of two REB mountings and one test bearing. A torque sensor measures the torque loss of the system.         REB tests rigs with vertical shaft are also presented in the literature. These test rigs allow testing thrust ball bearings (Ref.10) or cylindrical (Ref.11) and tapered roller bearings (Refs.12–13) under axial load. The torque measurement is realized with a torquemeter located on the vertical shaft. To finish, some test rigs have been realized only for one application. Ke et al (Ref.14) developed a specific REB test rig to study the thermal characteristics of double-row tapered roller bearings of a high-speed locomotive. Blake and Truman (Ref.15) designed a test rig to measure the running torque of tapered roller bearings. The abovementioned test rigs are dedicated to one operating condition or one size of REB. The new test rig developed in this study is dedicated to a wide range of REB dimensions and for different operating conditions. In the first section of this paper a new REB test rig design is presented. The second part of this paper is dedicated to the first experimental results. The New Test Rig Design Specifications of this test rig. The new test rig has been designed to be able to study the thermomechanical behavior of several kinds of REBs and for different lubrication conditions (grease, oil bath and oil jet). This modularity has been a key point during the test rig design. The tested REB outer diameter is between 72mm and 150mm. A lot of REBs can be tested in this test rig: deep groove ball bearings, angular contact ball bearings, roller bearings, etc. Radial and axial load can be applied on the tested REB. Test rig operation. Thus far, several tests rigs have been presented. For each test rig, a specific method is used to measure the REB torque loss. In this new test rig, REB torque loss measurement is divided into two steps:

? Calibration phase: Four identical REBs are mounted and work in the same operating condition (rotational speed, radial load, lubrication). These REBs are jetlubricated with the same lubricant at a given oil injection temperature. The torque loss is divided by four in order to isolate the torque loss of one REB.

? Measure phase: Two REBs (calibration block) are removed and replaced by the tested REB (Fig.2). The torque loss of mounting blocks (determined in calibration phase) is subtracted to the global torque measurement in order to isolate the contribution of the tested REB. This architecture requires a specific test rig design. The main challenge is to connect the blocks while maintaining a correct concentricity of the main shaft. Labyrinth seals are used to guarantee oil tightness and to limit power losses. Deep grove ball bearings are used in in the mounting blocks; their characteristics are presented in Table 1; they have been chosen in order to work at high-rotational speed. Test rig components: REB lubrication. The REB test rig is composed of two oil tanks. The first one is dedicated to the lubrication of calibration and mounting blocks. These REBs are always lubricated with the same oil at the same temperature (around 70°C).

    The lubricant properties are presented in Table 2. The second reservoir is dedicated to the tested REB; it allows testing different lubricants. Gear pumps impose an oil flow between 0 and 1.5 L/min on the REB. Each oil tank is surrounded by hot plates to warm the lubricant. The oil pipes are thermally isolated to reduce heat exchange with the ambient air, allowing for an oil injection temperature even higher than 100°C. The electro spindle maximum rotational speed is equal to 18,000rpm. A hydraulic jack allows applying a radial load on REB up to 20kN.

Conclusion

    This research work aims to present a new REB test rig dedicated to the study of the thermomechanical behavior of this mechanical component. In the first part of this paper, this new test rig is presented. It has been designed to study different kinds of REBs and for different operating conditions. The second part is dedicated to the first results that have been obtained. An 85mm pitch diameter deep groove ball bearing was tested under different operating conditions. Experimental results have been compared with global models of power losses. This part underlines that these models can estimate the REB torque loss for lowspeed application. However, when the (N.dm) product tends toward a million, the REB torque loss increases suddenly; this increase is not correctly taken into account in the global models. Moreover, the influence of the REB thermal behavior and the oil flow rate on the REB torque loss were highlighted.