BENCHMARKING OF TRIBOLOGY IN SLIDING BEARINGS BY ACOUSTIC EMISSION

Transcription

1 BENCHMARKING OF TRIBOLOGY IN SLIDING BEARINGS BY ACOUSTIC EMISSION B. Ziegler 1, R. Eichelbeck 1, H.-J. Schwalbe 1, K. Wierzcholski 2 and A. Miszczak 2 1 University of Applied Sciences, Wiesenstraße 14, Gießen, Germany; 2 Gdynia Maritime University, ul. Morska 83, PL Gdynia, Poland Keywords: Tribology, friction, slide bearing, sliding processes Abstract Friction behavior in slide bearings arises from friction mechanisms, such as boundary, mixed and (elasto)hydrodynamic friction. The friction process cannot be described online by established measurements. Different sliding processes cause characteristic acoustic emission (AE). The RMS-value of the AE depends on friction measurement quantity. In this investigation, the AE parameters are correlated with temperature, shear rate and dynamic viscosity of the lubricant. 1. Preliminaries The friction process in slide bearings cannot be described online by established measurement methods. Different sliding processes (boundary, mixed, hydrodynamic friction according to the Stribeck curve) cause characteristic acoustic emission (AE). This paper describes the benchmarking of tribology in slide bearings by using the characteristic AE to identify the sliding process. The examination was carried out at the slide-bearing test stand of the University of Applied Sciences Giessen-Friedberg. The construction of the test stand, the experiments and the results are described. The examinations were made possible due to the cooperation between the Maritime University of Gdynia and the University of Applied Sciences, Gießen under the EU-project TOK-FP6, MTKD-CT , Biobearing (transfer of knowledge). 2. Description of the Test Stand Onto the base frame of a tribometer, which was used as chassis, the slide-journal-bearing test stand was constructed with two bearings under test on one common shaft. To indicate the friction conditions, the AE analysis, a specialized mechanical system as well as temperature measurements were used. One main aim of the construction was to reduce most of the influences on the AE signals that are generated in the bearing under test and caused by mechanical oscillations of the drive of the bearing shaft, Fig. 1. The speed-controlled electric motor drives the intermediate shaft via a toothed belt, which itself is connected via a curved-tooth gear coupling with the main shaft. The curved-tooth gear coupling is used to compensate the misalignments of the shafts and - most important - to disconnect the spurious mechanical couplings between the main and intermediate shafts. The main shaft is supported by hydrostatic slide bearings, to avoid mechanical disturbances caused by the rolling-element bearings. The lubricant supply of the bearings is ensured by an external pump module. The shaft under test, which is made of an austenitic steel, will be connected to the main shaft by a hydraulic clamping device. Because the austenitic steel itself has a huge damping attribute in case of AE waves, the remaining high frequency disturbances are reduced to their minimum. The test rings are fixed onto the test shaft by thermal shrinking, Fig. 2. The AE sensor (PAC R6D) is placed at the bearing carrier (Fig. 3). The sensor transforms the vibration, that are generated in the bearing, depending on the sliding process, via a couplant of high viscosity grease, into electrical signals [5, 6], which are fed to a preamplifer (PAC 2/4/6C; 60-dB gain) with a band-pass filter of khz. The filtered and amplified AE signals are converted into an energy equivalent dc voltage by the root-mean-square module [5] EWGAE, Cracow UT

2 Fig. 1. Drive of the bearing shaft. The test stand is powered by an electric motor. Fig dimensional plot of the bearing shaft. The shaft under test is fixed concentric at the main shaft through a hydraulic clamping device. The test rings of the bearings are shrunk onto the shaft under test. Fig. 3. Suspension of the 1 st slide bearing under test with measuring devices (AE-sensor, thermoelement) and the lubricant supply. 185

3 The bearing carrier is connected to the suspension via an interlayer of vibration-damping material. This suspension that realizes the induction of the bearing load as well as the measurement of the friction force is constructed on a base plate, which is connected to the load cells by two steel ropes, Fig. 4. The load cells (Burster, Type ) have a range of 1000 N each and are connected to a carrier amplifier (HBM, MGT231; ±10V output). The center lines of the load cells form 60º angle with an adapter. The needed bearing load is conducted into the adapter via a steel rope. The whole construction of the suspension forms an equilateral triangle. At the centroid of the triangle there is nearly the central point of the bearing bore. This construction of the suspension enables us to measure the friction forces in the bearing under test while testing. The conducted bearing load is allocated symmetrically to both load cells if the shafts do not rotate. While rotating, the friction force (depending on the sliding process) produces a torque in the suspension. This torque relieves the load cell 1 and stresses the load cell 2 at the exactly equal amount. The friction force F R can be determined by the given geometrical connections [2]. The output signals of force, RMS-value and thermoelements are recorded with a multichannel measuring device (HBM, model UPM 60 Fa), connected to the control computer. For this computer, a special DELPHI-based software was installed. All test measurands are simultaneously recorded and continuously saved on the hard disc. The evaluation and the graphic presentation of the outcome are made via the EXCEL program. 3. Experiments Fig. 4, Drawing of the suspension of one test module. To eliminate the temperature influence from the hydrostatic bearings onto the bearings under test (conducted through the shaft under test), first the hydrostatic slide bearings were heated for ~45 minutes to their operating temperature by pumping the lubricant through them. After that, the experiments were started. To get different bearing capacities, the steps listed in Table 1 have been carried out. Depending on their current bearing capacity, the bearings under test are heated at each point of operation with a different temperature gradient. The temperature of the oil strongly affects the 186

4 viscosity of the lubricant. To reduce this influence there was a need to measure all values of each operation point synchronously within a few seconds after the start. Table 1 Steps of the experiments. step no. bearing load F L [N] shaft speed min. [min 1 ] shaft speed max. [min 1 ] shaft speed steps [min 1 ] To characterize the pairing friction parts their roughness is measured by a surface-roughness tester (Perthen, Pertometer C50), and the hardness with a hardness tester (Emco, M4C 075 G3R). The material of the test rings was analyzed by an emission spectrometer (Belec, Vario lab). Sleeve: (according to Fa Ferderal Mogul, Glyco), multilayer copper bearing, G18 Fa Glyco PbSn14Cu8, surface hardness 25 HV, (own measurements), width = 38 mm, inner diameter, d 1 = 82,255mm Shaft, test rings: width = 50 mm, external diameter d 2 = mm, material: S235, quality: , mean surface hardness: 179 HV 10, mean roughness of the surface: R z = R max = 13.5 µm. According to the existing dimensions of the bearings and using the equation (1), the relative bearing clearance ψ is : ψ = d 1 d mm 82.07mm = = (1) d mm where d 1 : inner diameter, d 2 : shaft diameter. The used oil under test was base oil, SN 600, Fuchs, Germany. 4. Results In Fig. 5, one can see the behavior of the 1st slide bearing under test identified by the measured friction force and the RMS-value versus the bearing capacity. There are three different bearing capacity ranges. These ranges (So-range I to So-range III) are calculated by the Sommerfeldnumber, which is defined as follows [4]: So = p m ψ 2 (2) η ω where: p m - mean surface pressure [Pa], ψ - relative bearing play [-], η - dynamic viscosity [Pa s], and ω - angular velocity [s -1 ]. For each operation point, the bearing capacity was calculated by equation (2). The dynamic oil viscosity depends strongly on the oil temperature, which is nearly the same as the measured bearing back temperature. The dynamic viscosity of the oil at different temperatures η ϑ was calculated according to equation (3), [1, 4]. 160 ϑ +95 C 0,1819 ln η 40 k η ϑ = k e (3) where: ϑ - temperature of the oil [ C], η 40 - nominal dynamic viscosity at 40 C [Pa s], and k - constant = Pa s. In the So-range I (from 0.19 down to 0.077), both the friction force value and the RMS-value show a continuous behavior with a low spread. The bearing load is 100 N, the bearing temperature starts at 51 C and increases during the test up to 53 C. The values increase by increasing the shaft speed. The friction force value increases by virtue of the increase in viscous friction in the 187

5 clearance [3]. The RMS-value increases by virtue of the increase in the shear rate of the lubricant [6, 7]. In the So-range II, the bearing capacity is in the low duty area (from So = down to So = 0.028). The bearing load is 80 N, the bearing temperature starts with 30 C and increases during the test up to 35 C. The characteristics of the friction force and of the RMS-value show the same dependence on the increase of the shaft speed as in the So-range I, but with a higher amount. The explanation for that is the higher viscosity of the lubricant due to the lower temperature than in So-range I. In the So-range III, the bearing capacity is in the lowest duty area (from So = down to So = 0.02). The bearing load is 60 N, the bearing temperature starts also at 30 C and increases during the test up to 35 C and therefore with the same oil viscosity as in the So-range II. The characteristics of the friction force and of the RMS-value show the same dependence on the increase of the shaft speed like in the So-range II, but with a lower amount. The explanation for that is the lower bearing load. The RMS-value at the lowest speed shows a discontinuity, because this value is too high. This is caused by a solid friction contact of the asperities of the pairing friction parts. The bearing is unstable at this very low bearing capacity. Besides that the solid friction generates a stronger AE than a hydrodynamic friction. The experimental results of the 2nd bearing under test are similar to the results of the 1st bearing under test and therfore are not mentioned further. Fig. 5. RMS-values (broken line) and friction force values (solid line) of the three measurement steps vs. the Sommerfeld-number. The diagram shows the results of the 1st bearing under test. 5. Conclusion The investigation shows a clear dependence of the friction force on the friction conditions in the hydrodynamic slide bearing. Higher oil viscosity leads to a higher friction within the clearance and therefore to a higher friction force. The friction force also increases due to an increase of the bearing load. The AE reacts with a lower inertia in regard to the changes of the friction conditions. Higher bearing loads, higher shaft speeds and higher oil viscosities each leads to higher RMS-values, but with different intensities. In opposite to the friction force, AE reacts spontaneously as well as with a high amount on the solid contact. Further investigations are planned with constant bearing temperature at the beginning of each operation point and higher bearing capacity. 188

6 6. References [1] Czichos, H., Habig, K.-H., Tribologie - Handbuch Reibung und Verschleiß, ed. 2003, Vieweg & Sohn Verlag, Wiesbaden. [2] Eichelbeck, R., Diplomarbeit SoSe 08, University of Applied Sciences Gießen-Friedberg, [3] Haberhauer, H., Bodenstein, F., Maschinenelemente, 14 ed., Springer Verlag, Berlin, [4] Roloff/Matek, Maschinenelemente Formelsammlung, 14 ed., Vieweg Verlag, Braunschweig, [5] Ziegler, B., Contribution of Acoustic Emission into optimal bearing lubrication., Journal of Kones, 14 (4), 2007, [6] Ziegler, B., Miszczak, A., Acoustic Emission as a Friction force Indicator after Test Stands Experiments. Journal of Kones, 14 (4), 2007, [7] Ziegler, B., Schwalbe, H.-J., Miszczak, A., Nachweis des Reibzustandes in Gleitlagern mit Hilfe der Schallemission. DGZfP Berichtsband BB 105 D, Vortrag 17, ISBN , Deutsche Gesellschaft für zerstörungsfreies Prüfen e.v., Berlin. 189