Jtam.dvi JOURNAL OF THEORETICAL AND APPLIED MECHANICS 46, 1, pp. 171-184, Warsaw 2008 ANALYSIS OF LONGITUDINAL TANGENTIAL CONTACT VIBRATION EFFECT ON FRICTION FORCE USING COULOMB AND DAHL MODELS Mariusz Leus Paweł Gutowski Technical University of Szczecin, Faculty of Mechanical Engineering, Szczecin, Poland e-mail: mariusz.leus@ps.pl; pawel.gutowski@ps.pl The paper presents results of analysis of changes of the friction force in sliding motion affecting a solid body subjected to excited longitudinal tangential contact vibration.The studywas conductedusing two friction models: firstly, the classicalCoulombmodel related to rigid bodymotion on a non-deformable base and secondly, theDahlmodelwhich takes into account tangential contact deformability including the phenomenon of ”pre-sliding displacement”. It was demonstrated that in the case of vi- brationmotionwith a low amplitude (i.e.motion in amicro scale, which is exemplified by longitudinal contact vibration), the Coulomb friction model is not adequate to describe the friction force. It was also shown that the friction force can be reduced in one vibration cycle without instantaneous change of the vector sign of this force, which in litera- ture is often quoted as the main reason for friction force reduction at longitudinal tangential contact vibration. Key words: friction models, friction force, tangential contact vibration 1. Introduction Many studies on the influence of contact micro vibration on the friction force carried out over the past several decades suggest that this force can be redu- ced by this vibration excitation both in the normal (Godfrey, 1967; Tolstoi et al., 1973; Budanov et al., 1980; Hess and Soom, 1991; Grudziński andKostek, 2005) and tangential direction to the plane of contact (Pohlman andLehfeldt, 1966; Mitskevich, 1968; Skare and Stahl, 1992; Katoh, 1993; Sase et al., 1995, 1997; Siegert and Ulmer, 1998, 2001; Kutomi 1999; Littmann et al., 2001a,b; 172 M. Leus, P. Gutowski Storck et al., 2002; Kumar andHutchings, 2004). In the case of tangential vi- bration,most studieswere carried outwithin ultrasonic vibration range,which is connected with the fact that this vibration plays an important role among methods aiming at improvingmanufacturing techniques through lowering fric- tion forces between a tool and aworkpiece, particularly in plastic forming and machining. Most of the studies on the influence of tangential micro vibra- tion on the friction forcemainly concern practical applications of vibration in various technological processes omitting the mathematical description of the phenomenon. The first theoretical model describing the influence of longitudinal tangen- tial contact vibration on the friction forcewas presented byMitskevich (1968). According to thismodel, it is possible to reduce the average friction force thro- ugh a cyclic and instantaneous change of the vector sign of this force registered in every vibration period provided that the amplitude va of the vibration ve- locity is higher than the constant component vc of the slidingmotion velocity. This phenomenon described as the friction vector effect is commonly regarded as one of the most important mechanisms which can lower the friction force in sliding motion for excited longitudinal tangential contact micro vibration (Eaves et al., 1975; Siegert andUlmer, 1998, 2001; Skare and Stahl, 1992). On the basis of the same assumption, Littmann et al. (2001a,b) as well as Kumar andHutchings (2004) using classical frictionCoulomb’s law elaborated succes- sive models allowing them to determine the friction force duringmacroscopic sliding of one body against the other while longitudinal tangential contact vibration takes place. The calculation results of the friction force for these models show si- gnificant discrepancies in comparison with the results obtained in empiri- cal research, which were carried out by Littmann et al. (2001a,b), Storck et al. (2002) or Kumar and Hutchings (2004). A much better consisten- cy of the calculation results as compared with the experimental results gi- ven in the above quoted articles by Littmann et al. (2001a,b) and Storck et al. (2002) was obtained by Tsai and Tseng (2006) for the model develo- ped by Dahl (1968, 1976) which takes into account contact deformability in the tangential direction, or with the use of the elasto-plastic friction model proposed by Dupont et al. (2000, 2002), which is a development of Dahl’s model. The present study attempts to explain, through numerical analysis, physi- cal reasons accounting for amuchbetter agreement of friction force calculation results obtainedusing theDahlmodelwith the experimental data as compared with other models based on classical Coulomb’s friction law. Analysis of longitudinal tangential contact... 173 2. The investigated object In the numerical calculations of the friction force in sliding motion taking place in the presence of excited longitudinal tangential contact vibration, the authors used a model (Fig.1) in which body Awith themass mmoves with the relative velocity vr with regard to base B. Fig. 1. The model used for calculations The velocity vr is the superposition of two component velocities vc and vv vr = vc+vv (2.1) where vc – a constant component, vc = const, connected with motion in the macroscopic scale (without vibration) vv – a variable component, vv = vv(t), connected with excited con- tact vibration (motion in the microscopic scale). When harmonic motion is assumed, the variable component can be given in the form vv(t)=xaωcos(ωt) (2.2) where xa – amplitude of excited vibration ω – circular frequency of excited vibration t – time. The product xaω is the amplitude va of the velocity of excited vibration va =xaω (2.3) Hence, equation (2.1) can be written as vr(t)= vc+vacos(ωt) (2.4) 174 M. Leus, P. Gutowski In the Coulomb friction model, it is assumed that the interacting surfaces of contact are ideally rigid. The friction force FT can be given by FT =FC sgn(vr) (2.5) where FC =µFN (2.6) FC is the Coulomb friction force, µ – coefficient of kinetic friction and FN – normal contact reaction of the contact area. From relation (2.5), it can be seen that in this model the friction force is constant in its absolute value and only its plus/minus sign can change, depending on the sign of the relative velocity vr of sliding. In fact, machined surfaces which compose the contact area are not ideally smooth. The process of machining causes deflections from the ideal state. These include roughness,waviness and shape errors.As a result, these surfaces do not adhere one to an other ideally throughout all the nominal area of contact, but instead they only adhere in somemicro-areas (Fig.2a) distributed in a random fashion on this surface. Fig. 2. Real contact of two bodies and its model: (a) a scheme of contact of two machined surfaces, (b) the Dahl frictionmodel showing the distribution of the total displacement x of a rigid body on the elastic z and plastic w components In the 1970s, the so called dynamic friction models were developed, inclu- ding the model elaborated by Dahl (1968, 1976). In this model the asperities are modelled by means of micro-springs (Fig.2b), which when a tangential load is applied deflect in the direction of frictional resistance force. If the de- flection is big enough (i.e. when the friction force reaches itsmaximumvalue), the contact is broken-away and then sliding takes place. Analysis of longitudinal tangential contact... 175 According to theDahlmodel, the displacement x of a rigid body is elasto- plastic and can be broken down into two components: an elastic z and a plastic w x= z+w (2.7) The elastic component is connected with elastic deflections of asperities in the tangential direction. Its mutual relation with the friction force FT can be given by FT = ktz (2.8) where kt is the tangential contact stiffness coefficient. This deflection can be described with a differential equation (Dahl, 1976) dz dt = vr ( 1− kt FC sgn(vr)z )α (2.9) The parameter α in this equation, defines the shape of a curve describing the dependence of the tangential deflections on tangential force. For brittle materials, the value of this parameter is in the range of 0 < α < 1, and for ductile materials it is α­ 1 (Bliman, 1992). 3. Numerical calculations Fornumerical calculations of the friction force the followingparameters of exci- ted vibration were taken: frequency f =60kHz and amplitude xa =0.7µm. They are identical to those which were assumed by Littmann et al. (2001a) in their experimental investigation of the longitudinal contact vibration in- fluence on the friction force. The following data were also used: mass of the die m = 0.02kg (Littmann, 2006), kinetic friction coefficient µ = 0.1, va- lue of the parameter α = 1 and the tangential contact stiffness coefficient kt = 0.056N/µm. The last coefficient was determined on the basis of data provided in the Tsai and Tseng (2006) paper. For the assumed frequency f and amplitude xa, taking into account that ω=2πf, the amplitude of vibration velocity va =0.264m/s was determined. Itwasdoneusingequation (2.3).The friction force calculationswereperformed for a sequence of the velocity vr value with a constant componentvc equal to: vc = 0.324, 0.264, 0.230, 0.199, 0.146, 0.095m/s, for which Littmann et al. (2001b) carried out experimental measurements of the friction force. For the Coulombmodel, the value of the FT force was determined from relation (2.5), and for the Dahl model it was determined from relations (2.8) and (2.9). The 176 M. Leus, P. Gutowski results of numerical calculations for the assumed values of motion parameters are shown in graphical forms in Figures 3 and 4. Fig. 3. The friction force for the Coulomb andDahl models when the amplitude of vibration velocity va is less than or equal to the constant component vc: (a) va =0.264m/svc). As it can be seen in graphs presented in Fig.4, in such a case, the results may differ considerably depending on which friction model is chosen. In the Coulomb friction model, for the whole range of velocities which satisfy the condition that va > vc, a change of sign of the relative sliding velocity results in an immediate change of sign of the friction force (Fig.4). As a result, the calculated values of the average friction force for one period of vibration are much smaller than tho- se determined experimentally, which is seen in the papers by Littmann et al. (2001a,b) and Storck et al. (2002). The calculation results using the Dahl model (Fig.4) demonstrate that a change of an instantaneous friction force in the presence of tangential contact vibration isnotabrupt.For small differencesbetween va and vc (Fig.4a,b), the value of the instantaneous friction force is loweredwithout the sign change (its Analysis of longitudinal tangential contact... 177 Fig. 4. The friction force for the Coulomb andDahl models when the amplitude of vibration velocity va (va =0.264m/s) is greater than the constant component vc: (a) vc =0.230m/s, (b) vc =0.199m/s, (c) vc =0.146m/s, (d) vc =0.095m/s sense is opposite to vc). For largedifferencesbetween va and vc (Fig.4c,d), the instantaneous friction force is gradually reduced to zero and then it changes its sign and starts growing in the opposite direction (in accordance with vc). It is connected with the direction of elastic deflection of the contact area asperities relative to themotion trajectory of the rigidbody,which is illustrated inFig.5. In both analyzed models, for each considered value of the velocity vc, an 178 M. Leus, P. Gutowski Fig. 5. Diagrams illustrating the relation between the friction force FT and the elastic deflection of asperities in the contact area; (a) va =0.264m/svc =0.199m/s, (c) va =0.264m/s>vc =0.095m/s average friction force FT in one vibration period T =2π/ω was determined. This force was evaluated from the following relation FT = 1 n n ∑ i=1 FT(t+ i∆t) (3.1) Analysis of longitudinal tangential contact... 179 where n is the number of time steps in one vibration period n= 2π ω∆t (3.2) The percentage reduction S of the average friction force, calculated from Dahl’s andCoulomb’smodels,which tookplace asa result of excited tangential contact vibrations, was also evaluated S= ( 1− FT FC ) ·100% (3.3) The results of FT and S obtained in the calculations are presented in a ta- bular form inTable 1where they are comparedwith the results of experiments carried out by Littmann et al. (2001a). Table 1. A comparison between numerical calculations and experimental re- sults (va =0.264m/s) Velocity vc [m/s] Average friction force, FT [N] Coulomb Dahl Experiment model model (Littmann et al., 2001a) 0.324 0.0196 0.0196 0.0197 0.264 0.0196 0.0196 0.0187 0.230 0.0131 0.0181 0.0179 0.199 0.0107 0.0156 0.0168 0.146 0.00724 0.0109 0.0112 0.095 0.00454 0.0066 0.0077 Velocity vc [m/s] Reduction of average friction force, S [%] Coulomb Dahl Experiment model model (Littmann et al., 2001a) 0.324 0 0 −0.56 0.264 0 0 4.32 0.230 33.2 7.65 8.45 0.199 45.4 20.4 14.1 0.146 63.1 44.4 42.9 0.095 76.8 66.2 60.6 This comparison clearly shows that the Dahl model provides a far better consistency as compared with the experimental results than other models, in- 180 M. Leus, P. Gutowski cluding the classical Coulomb frictionmodel which does not take into account tangential contact stiffness. The friction force value calculated using the Dahl model – equations (2.8) and (2.9) – is a function of contact stiffness in the tangential direction. This is why the quantitative discrepancy between the results obtained using this model and those calculated using theCoulombmodel significantly depends on the contact stiffness in the tangential direction. Fig. 6. The friction force for the Coulomb andDahl models depending on the stiffness of contact in the tangential direction, when va >vc; (a) va =0.264m/s and vc =0.230m/s, (b) va =0.264m/s and vc =0.146m/s This problem is graphically illustrated in Fig.6, which compares the fric- tion force FT calculated using the Dahl model for contacts with various tan- gential stiffness with the friction force obtained using the Coulomb model. Figure 6a presents the friction force FT for velocity vc = 0.230m/s and am- plitude of vibration velocity va = 0.264m/s, whereas Fig.6b presents the friction force FT for vc = 0.146m/s and va = 0.264m/s. Tables 2 and 3 present a comparison of the average friction force FT within one period for different contact stiffness parameters calculated using the Dahl model. They are compared with respective FT values calculated using the Coulombmodel for the above given values of vc and va. The graphspresented inFig.6 and the comparison shown inTables 2 and3 clearly demonstrate that discrepancies in the friction force FT estimated in one period of longitudinal tangential contact vibration using the Coulomb and Dahl models decrease as the contact stiffness increases in the tangential direction. Analysis of longitudinal tangential contact... 181 Table2.Average friction force FT according to theCoulombandDahlmodels depending on values of kt, for vc =0.230m/s and va =0.264m/s Coefficient of Average friction force FT [N] contact stiffness Dahl Coulomb kt [N/µm] model model 0.030 0.0186 0.0131 0.056 0.0181 0.120 0.0172 0.240 0.0160 0.480 0.0148 0.960 0.0139 Table3.Average friction force FT according to theCoulombandDahlmodels depending on values of kt, for vc =0.146m/s and va =0.264m/s Coefficient of Average friction force FT [N] contact stiffness Dahl Coulomb kt [N/µm] model model 0.030 0.0127 0.00724 0.056 0.0109 0.120 0.00877 0.240 0.00783 0.480 0.00754 0.960 0.00742 4. Summary The presented above analysis demonstrates that depending on the mathe- matical model used in numerical calculations the results of the influence of longitudinal tangential contact vibration on the friction force vary. A marke- dly better consistency as compared with the experimental data provided in the literature (Littmann et al., 2001a,b; Storck et al., 2002) is obtained using the Dahl model. The analysis of numerical calculations conducted using the Dahl model showed that the friction force in sliding motion in the presence of excited longitudinal tangential contact vibration can be reducedwithout changing the sign of the friction force vector.Theview that follows fromCoulomb’s law, and 182 M. Leus, P. Gutowski which is still currently adopted bymany authors that themainmechanism of reducing friction force in sliding motion in the presence of tangential contact vibration results in a cyclic, instantaneous sign change of this force, in the light of the presently obtained results has not been confirmed. References 1. BudanovB.V., KudinovV.A., Tolstoi D.M., 1980, Interaction of friction and vibration, Soviet Journal of Friction and Wear, 1, 79-89 2. BlimanP.A., 1992,Mathematical studyof theDahl’s frictionmodel,European Journal of Mechanics, A/Solids, 11, 6, 835-848 3. Eaves A.E., Smith A.W., Waterhouse W.J., Sansome D.H., 1975, Re- viewof theapplicationofultrasonicvibrations todeformingmetals,Ultrasonics, 13, 4, 162-170 4. Dahl P.R., 1968, A solid friction model, Technical Report TOR-0158H3107– 18I-1, The Aerospace Corporation, El Segundo, CA 5. Dahl P.R., 1976, Solid friction damping of mechanical vibrations, AIAA Jo- urnal, 14, 12, 1675-1682 6. Dupont P., ArmstrongB., Hayward V., 2000, Elasto-plastic frictionmo- del: contact compliance and stiction,Proceedings of the AmericanControl Con- ference, Chicago: AACC, 1072-1077 7. DupontP.,HaywardV.,ArmstrongB.,AltpeterF., 2002, Single state elasto-plastic frictionmodels, IEEE Transactions of Automatic Control, 47, 5, 787-792 8. Godfrey D., 1967, Vibration reduces metal to metal contact and causes an apparent reduction in friction,ALSE Transactions, 10, 183-192 9. Grudziński K., Kostek R., 2005, Influence of normal micro-vibrations in contact on sliding motion of solid body, Journal of Theoretical and Applied Mechanics, 43, 1, 37-49 10. Hess D.P., Soom A., 1991, Normal vibrations and friction under harmonic loads: Part I – hertzian contacts, Journal of Tribology, 113, 80-86 11. KatohK., 1993,Active control of friction using ultrasonic vibration, Japanese Journal of Tribology, 38, 1019-1025 12. Kumar V.C., Hutchings I.M., 2004, Reduction of sliding friction of metals by the application of longitudinal or transverse ultrasonic vibration,Tribology International, 37, 833-840 Analysis of longitudinal tangential contact... 183 13. Kutomi H., Sase N., Fujii M., 1999, Development of friction controller, Proceedings of the International Conference AMPT’99, I, 605-612 14. Littmann W., 2006, Private correspondence 15. Littmann W., Storck H., Wallaschek J., 2001a, Reduction of friction using piezoelectrically excited ultrasonic vibrations,Proceedings of SPIE’s 8th Annual International Symposium on Smart Structures and Materials, Bellin- gham,Washington, 302-311 16. Littmann W., Storck H., Wallaschek J., 2001b, Sliding friction in the presence of ultrasonic oscillations: superposition of longitudinal oscillations, Archive of Applied Mechanics, 71, 549-554 17. MitskevichA.M., 1968,Motion of a body over tangentially vibrating surface, taking into account of friction, Soviet Physics-Acoustics, 13, 348-351 18. Pohlman R., Lehfeldt E., 1966, Influence of ultrasonic vibrations on me- tallic friction,Ultrasonics, 4, 178-185 19. Sase N., Chandy M., Suzuki K., Kumagai T., Fujii H., 1995, Reduc- tion of frictionwithout lubrication,Proceedings of the International Conference AMPT’95, III, 1298-1304 20. Sase N., Kurahashi T., Fujii M., Kutomi H., Fujii H., 1997, Control of friction coefficient between metal surfaces, Proceedings of the International Conference AMPT’97, II, 609-615 21. SiegertK.,UlmerJ., 1998,Reduction of sliding frictionbyultrasonicwaves, Production Engineering, 5, 9-12 22. Siegert K., Ulmer J., 2001, Superimposing ultrasonic waves on the dies in tube andwire drawing, Journal of EngineeringMaterials and Technology, 123, 517-523 23. Skare T., Stahl J.E., 1992, Static and dynamic friction processes under the influence of external vibrations,Wear, 154, 177-192 24. Storck H., Littmann W., Wallaschek J., Maik M., 2002, The effect of friction reduction in presence of ultrasonic vibrations and its relevance to traveling wave ultrasonic motors,Ultrasonics, 40, 379-383 25. Tolstoi D.M., BorisovaG.A., Grigorova S.R., 1973, Friction regulation by perpendicular oscillation, Soviet Phisics-Doklady, 17, 9, 907-909 26. Tsai C.C., Tseng C.H., 2006, The effect of friction reduction in presence of in-plane vibrations,Archive of Applied Mechanics, 75, 164-176 184 M. Leus, P. Gutowski Analiza wpływu drgań kontaktowych stycznych wzdłużnych na siłę tarcia przy wykorzystaniu modelu Coulomba i modelu Dahla Streszczenie Wpracy przedstawionowyniki analizy zmian siły tarcia w ruchu ślizgowym ciała sztywnego przy występowaniu wymuszonych drgań kontaktowych stycznych wzdłuż- nych tego ciała. Badania przeprowadzono przy wykorzystaniu dwóch modeli tarcia: klasycznegomodeluCoulombaodniesionegodo ruchu ciała sztywnegopo nieodkształ- calnympodłożu orazmodelu Dahla uwzględniającego podatność kontaktową styczną styku. Wykazano, że w przypadku występowania ruchu drgającego o małej wartości amplitudy (ruch w skali mikro, którego przykładem są drgania kontaktowe styczne) model tarcia Coulomba jest nieadekwatny do opisu siły tarcia. Wykazano również, że siła tarcia może ulec obniżeniu bez występowania zjawiska chwilowej zmiany zna- ku wektora tej siły, które w literaturze często podawane jest za główną przyczynę obniżenia siły tarcia przy występowaniu kontaktowych drgań wzdłużnych. Manuscript received March 9, 2007; accepted for print October 8, 2007