Dependency of bearing noise properties on surfaces lubrication/Pavirsitepimo itaka guoliu keliamam triuksmui.

Augustaitis, V.K. ; Bucinskas, V.

1. Introduction

Noise of technical equipment usually assumed to be negative effect. Noise affects personnel and brings its negative effect of operating machinery and equipment. Technical equipment with rotational movement has different types of bearings, which are known as technical noise generators.

Modern materials in bearings create some specific problems with noise intensity and this paper touches some aspects of noise generation in steel and ceramic bearings. To make research of such task it is necessary to built a model of noise generation and to find parameters of noise generation mechanism.

Noise of bearings is created by vibrating solid bodies, which is transmitting further to air. It is known [1] that sound filters solid body vibrations due to its own physical properties of wave media.

Vibration of bearing components has different nature, but for operable ones the main part of noise in 2 5 Hz area belongs to friction noise [2-8].

2. Aim and formulation of research

This research is intended to find dependency between bearing material hardness, lubrication of bearing and vibration characteristics as well as acoustic noise.

The research presented consists from several parts. The first step is to accept model of friction noise generation in a bearing, based on tribologic properties of bearing surfaces and other sources of vibration are neglected. This statement is used only in case of experimental setup; in real machinery it can't be used directly. Noise generation hypothesis is presented in Fig. 1. There are two cases of vibration generation - when two surfaces are sliding (as shown in Fig. 1, a), and direction of surfaces movement is opposite in tangential direction or when the surfaces are rolling and direction of surfaces movement is coincident in tangential direction.

In this paper only rolling mode is assumed.

[FIGURE 1 OMITTED]

Surface of contact area are random peaked and only statistically evaluated surface pitch [R.sub.t] and average roughness [R.sub.z] are available to evaluate in this model. So, piece of some surface projection, showed on Fig. 1, b can be accessed directly by another surface to support force, or support surface is modified by lubricant film, which is represented by two positions of oil surface as "thin oil" and "thick oil".

These surfaces are interacting with peaks, "sticking out" from lubricant film with their own [R.sub.z] and [R.sub.t].

This creates different conditions in excitation of vibrations when the surfaces are moving in coinciding tangentiall directions (rolling) or in opposite tangentiall directions (sliding).

Then for a single surface cross section is possible to write the following dependencies

[f.sub.sl] = [[omega].sub.1][R.sub.1] + [[omega].sub.2][R.sub.2] / [R.sub.t] (1)

where [f.sub.sl] is average statistical frequency in sliding; [[omega].sub.1], [[omega].sub.2] are angular speed of surfaces; [R.sub.1], [R.sub.2] are radii of rotation for these surfaces; [R.sub.t] is equivalent pitch for a roughness.

In case of rolling average theoretical frequency can be expressed as

[f.sub.sl] = [[omega].sub.1][R.sub.1] / [R.sub.t] = [[omega].sub.2][R.sub.2] / [R.sub.t] (2)

Dependencies (1) and (2) are very rough estimation of kinematical excitation in contact of friction noise, which separates modes of movement of contact surfaces cross-sections. Real frequencies are generated from real surface relief, so the number of frequencies in huge, but center values follows this model [8 - 9].

This model has also more modifications, but the main ideas can be accepted also in slightly simplified form [10 - 11].

3. Modelling of contact

In order to look on contact area behavior, FEM model of one side bearing contact, which allows to evaluate distribution of stresses in contact and shows active area in noise generation was created.

As the basis for this model a bearing No. 208 was taken, but the model can be used in other applications. Contact model was created using SolidWorks software and FEM analysis was performed on CosmosWorks 2010. FEM model (Fig. 2) was created using one fragment from bearing No. 6208. The fragment of outer ring was taken with angular size, which corresponds to another neighboring ball contact places. For contact modeling here was chosen contact type "surface -surface", while in the model little gap was created and bonding place was unknown. Because of ball shape, restrictions in both horizontal axis directions were fixed by creation axis, only vertical ball movement was allowed. Outer ring fragment was fixed in all 3 axes direction on the outer surface. In order to run such model, special stabilizing spring was used.

All elements were taken as 1st order tetrahedron, contact was assumed to be Coloumb's. Load of the ball was applied in the area of estimated contact area of other ring, size - 80% of bearing radial load, according the shape of free body diagram.

[FIGURE 2 OMITTED]

Solution of this task is presented in Fig. 4, where inner surface of the bearing and ball contact surface are shown. Configuration of contact on ring running surface and ball are slightly shifted from the axis of symmetry due to global friction in the contact.

[FIGURE 3 OMITTED]

The analysis of stresses and displacements shows slight change in stresses in the area contact proves model of friction noise, because great gradient in stresses would create another scenario in vibration excitation.

4. Methodology of research

This experimental research was performed in Braunschweig technical university (Germany) for ceramic bearings and in Hannover Leibnitz technical university (Germany) was made extensive steel bearing research.

Initially contact area of the bearing was measured for roughness. Profile of contact, which is shown in Fig. 4, was evaluated for the main parameters and statistically proved values or [R.sub.t] and [R.sub.z] were defined. These values were basic in the definition of desired noise frequency range. It is necessary to take into account, that statistically proved data on surface of contact differs within certain tolerance and phase of vibration from touching roughness profile is also unknown. The frequency of resulting vibration from outer ring of the bearing was recorded and further analyzed. Spectrum of vibration accelerations is a result of such research and the main information supplier.

In case of nonlubricated bearing, ceramic bearing with very similar surface parameters was tested in another test rig. This test rig was intended to record acoustic pressure. Duration of the test in this test rig was very short in order to avoid heating. Partial axial load was applied, because installation of ceramic bearing was not tightening enough the outer ring and there was internal gap. Loads on ceramic bearing and steel bearing were made proportional to their maximum load (80% of it) and axial load was assumed not influential.

[FIGURE 4 OMITTED]

Bearings (steel and ceramic) were built into special setups and rotated with corresponding load and rotational speed. Because of different design of setup and bearing size, rotational frequency was taken so, that linear rotational velocity of the rolling bodies should be the same - 600 rpm for steel bearing and 1000 rpm for ceramic bearings.

5. Results

Results of performed research are shown below graphically. Fig. 5 shows steel bearing vibration acceleration signal in time. This sample was lubricated by industrial lubricant 9.

[FIGURE 5 OMITTED]

Vibration signal output from the steel bearing was recorded; signal discretion is 0.00002 s, which corresponds 50 kHz of sampling rate. From such signal frequency spectrum was calculated using FFT. As it is seen from vibrational spectrum (Fig. 6), the main frequency peak is about 3000 Hz and 7000 Hz as the second harmonics and low V subharmonic on 1500 Hz. Higher frequency range in spectrum has low accuracy because of sampling rate. These frequencies do not fit to ball pass or inner bearing ring revolution rate.

[FIGURE 6 OMITTED]

In case of lubricant t68 (Fig. 7), the spectrum was shifted to higher side and the values of vibration amplitudes are significantly higher, what means lower film thickness, as stated in initial model. The main frequency range shifted to 2600 Hz, correspondingly the second harmonics to 7200 Hz. During these tests lubricant film thickness was not measured, but the frequency change for more than 10% shows that the number of sticking roughness peaks correspondingly increased about the same number.

[FIGURE 7 OMITTED]

In case of ceramic bearing analysis (the measuring of outer ring vibration was made using laser sensor) another software was used. As it is possible to see in processed spectrum, ceramic bearings create much higher values of vibrations (Fig. 8, a) and correspondingly higher acoustic pressure. Higher level of vibration amplitudes is caused due to higher hardness of bearing material. Absence of lubricant in contact enables to touch much more peaks in contact area and energy of elementary impact between roughness peaks makes much higher.

[FIGURE 8 OMITTED]

Result of such investigation evidently show that more accurate model for vibration frequency evaluation is necessary, because the presence of prescribed frequency in the range of 3000 Hz and below has coinciding sub harmonics and higher harmonics.

6. Conclusions

After analysis of different types of bearings, it is possible to state, that material of a bearing and lubrication influence on bearing vibrations and bearing noise is significant. Harder materials for bearings, such as ceramics, without the use of lubricant increase noise values and vibration energy and only soft media between ceramic parts can make effect of vibration energy decrease. Correspondingly, in case of soft media, like thick oil film, increase resistance moment in the bearing and cause higher heating. In case of oil film breakage, dry friction due to dry surface increase energy of vibration (expressed in acceleration amplitude shift to higher frequency range) and higher values of local stresses (contact in surface peaks) develop heat and surface damage. As particular conclusion it is possible to state:

1. Difference in dynamic characteristics of bearing vibration in the same surface interaction velocity with bearings from different material is defined by surface hardness and particularly rolling/sliding values.

2. Lowering of bearing vibration and noise with different lubricant properties are directly influenced by lubricant film efficient thickness and surface "sticking roughness", load in the bearing can be expressed in the terms of efficient film thickness;

3. Rolling and momentum sliding mode of rolling elements inside bearing is characterizing by vibration spectrum frequencies significantly.

References

[1.] Othman, M.O.; Elkholy, A.H.; Seireg. A.A. 1990. Experimental investigation of frictional noise and surface-roughness characteristics, Experimental Mechanics 30(4): 328-331.

[2.] Bot, A. Le; Chakra, E. Bou. 2010, Measurement of friction noise versus contact area of rough surfaces weakly loaded, Tribology Letters 37(2): 273-281.

[3.] Othman, M. O.; Elkholy, A.H. 1990. Surface-roughness measurement using dry friction noise, Experimental Mechanics 30(3): 309-312.

[4.] Yokoi, M.; Nakai, M. 1982. A fundamental study on frictional noise, Bulletin JSME 22(173): 1665-1671.

[5.] Sergienko, V.P.; Bukharov, S.N.; Kupreev, A.V. 2008. Noise and vibration in brake systems of vehicles. Part 1: Experimental procedures, Journal of Friction and Wear 29(3): 234-241.

[6.] Sergienko, V.P.; Bukharov, S.N. 2009. Vibration and noise in brake systems of vehicles. Part 2: Theoretical investigation techniques, Journal of Friction and Wear 30(2): 216-226.

[7.] Akay, A. 2002. Acoustics of friction, Journal of Acoustics 111(4): 1525-1548.

[8.] Nakano, K. 2006. Two dimensionless parameters controlling the occurrence of stick-slip motion in a 1-DOF system with Coulomb friction, Tribology Letters 24(2): 91-98.

[9.] Perret-Liaudet, J.; Rigaud, E. 2006. Response of an impacting Hertzian contact to an order-2 sub harmonic excitation: theory and experiments, Journal of Sound and Vibration 96(1-2): 319-333.

[10.] Deleau, F.; Mazuyer, D.; Koenen, A. 2009. Sliding friction at elastomer/glass contact: influence of the wetting conditions and instability analysis, Tribology International 42(1): 149-159.

[11.] Salau, T.A.; Adedokun, A.O.; Oke, A.S. 2006. Surface area determination of ragged metallic bodies, Mechanika 5(61): 66-70.

[12.] Vekteris, V.; Cereska, A.; Jurevicius, M. 2008. Comparable analysis of vibrodiagnostics results of rotory components with different type bearings, Vibroen gineering 10(2): 251-255.

[13.] Kanapeckiene, L.; Kaklauskas, A.; Zavadskas, E.K.; Seniut, M. 2010. Integrated knowledge management model and system for construction projects, Engineering Applications of Artificial Intelligence 23(7): 1200-1215.

[14.] Ragulskis, K.; Dabkevicius, A.; Kibirkstis, E.; Bivainis, V.; Miliunas, V.; Ragulskis, L 2009. Investigation of vibrations of a multilayered polymeric film, Mechanika 6(80): 30-36.

Received January 21, 2011

Accepted May 30, 2011

V. K. Augustaitis, Vilnius Gediminas Technical University, Basanaviciaus 28, 03224 Vilnius, Lithuania, E-mail: vytautas.augustaitis@vgtu. lt

V. Bucinskas, Vilnius Gediminas Technical University, Basanaviciaus 28, 03224 Vilnius, Lithuania, E-mail: vytautas.bucinskas@vgtu.lt