JOURNAL OF THEORETICAL AND APPLIED MECHANICS 43, 1, pp. 37-49, Warsaw 2005 INFLUENCE OF NORMAL MICRO-VIBRATIONS IN CONTACT ON SLIDING MOTION OF SOLID BODY Karol Grudziński Robert Kostek Department of Mechanics and Machine Elements,Technical University of Szczecin e-mail: karol.grudzinski@ps.pl The paper shows results of some simulation studies on: (i) normal con- tact micro-vibrations in a simple model of a friction system and (ii) the influence of such vibrations upon reduction of friction forces in the sys- tem.Amodel of friction contactwith two degrees of freedom is adopted, considering non-linear normal contact flexibility of rough surfaces acting on each other. Themechanismswhich cause the reduction of the friction force are investigated and described. It is proved that the main cause of the decrease of the friction force, if normal contact micro-vibrations are present, is due to certain dynamic processes taking place within the contact region. This decrease is not due to lower values of the friction coefficient, nor lowermeanvalues of the normal reactionand real contact area, as it is often quoted in many publications. Key words: mechanical systems, vibration, friction reduction 1. Introduction According to many to date investigations, there exists powerful mutual interaction between friction and vibrations in mechanical systems (see Tol- stoi, 1967; Godfrey, 1967; Budanov et al.,1980; Broniec and Lenkiewicz, 1982; Sakamoto, 1987; Martins et al., 1990; Hess and Soom, 1991; Hess and Wagh, 1995; Sk̊are and St̊al, 1992; Kohei and Kyosuke, 2001; Grudziński and We- dman, 1999; Grudziński et al., 2000; Grudziński and Kostek, 2003). Friction may excite vibrations of different kinds, and vibrations in turn may have si- gnificant influence on the current value of the friction force and its variation in time. In specific conditions, in dynamic systems, friction may cause nonu- niform sliding in the form of stick-slip motion or tangential quasiharmonic 38 K.Grudziński, R.Kostek oscillations. On the other hand, through excitation and maintaining of pro- perly directed micro-vibrations, it is possible to decrease the friction force as well as the stick-slip effect, thus improving the uniformity of sliding motion. These two aspects of the friction-vibration interaction have long been (Tolstoi, 1967; Godfrey, 1967) and are also currently (Pfeiffer, 1999; Kohei andKyosu- ke, 2001; VDI-Berichte, 2002) the subject of interest of investigators who deal with the friction and dynamics of machinery. Reductions of the friction force observed in investigations, wherevibrations between contacting bodies are present, are usually explained in the literature on the subject through decreased values of the friction coefficient (Godfrey, 1967; Broniec and Lenkiewicz, 1982; Sk̊are and St̊ahl, 1992) or the decreased meanvalue of the normal reactionwithin the contact area (Tolstoi et al., 1973; Budanov et al., 1980; Hess andSoom, 1991). This is a significant simplification of the considered problem that can lead to false conclusions. Despite the fact that there is a great deal of knowledge on that subject and a wealth of expe- rimental data, there is a lack of appropriate models and theories which would enable reliable understanding of the interaction between friction and vibration in dynamic systems and their qualitative and quantitative investigations with the aid of computer simulation methods. The goal of this paper is to prove that themain cause of the decrease in the friction force in mechanical systems, where normal contact micro-vibrations are present, are complex non-linear dynamic processes occurring where so- lid bodies with rough surfaces come into contact, and not a decrease in the friction coefficient or mean normal reaction within the contact area. In the considerations contained herein, the friction coefficient has been assumed to be a constant quantity. 2. Theoretical fundamentals The investigated object is a system of two bodies in planar friction contact as shown in Fig.1a. This consists of a slider and a slideway. The surfaces in friction contact are rough (Fig.1b) and create an elastic interface which is modelled by a system of non-linear springs (Fig.1c). The slider, with the mass m, is assumed to be a rigid block. The load-deflection characteristic of the interface is presented in Fig.1d. On the grounds of many investigations (see Levina, 1965, Back et al., 1973; Martins et al., 1990), it may be well described by an involution formula as follows Influence of normal micro-vibrations in contact... 39 δ= enp m1 (2.1) where: δ denotes the normal deflection of the interface, p – nominal contact pressure, and en and m1 are interface constants depending on the actual type of material and surface treatment of the friction pair, to be determined experimentally. The model (Fig.1c,d) represents, in simplification, a working unit of a machine tool, which is moved along a slideway. The research carried out by Levina (1965), Ostrovski (1965), Dolby and Bell (1971) has shown that for the range of interface pressures and surface finishes used in the joint faces of machine tools, the above formof equation fits the experimental results. Tables with experimental values of the constants en and m1 for several combinations of materials and surfaces finishes can be found in Back et al. (1973). In theadoptedmodel of the interface (Fig.1) only elastic normaldeflections are taken into account.Thedampingandtangential deflections at the interface are neglected in this work for the sake of simplificity. These factors will been taken into account in farther investigations. As a consequence of the contact flexibility, the slider may micro-vibrate in the normal direction – both freely and as a result of the excitation, see Tolstoi (1967); Budanov et al. (1980). Fig. 1. Model of a simple mechanical systemwith non-linear contact flexibility and friction A constant force F is applied to the slider (Fig.1c) in the direction of the x axis.The force tends to introduce slidingmotion.Thismotion is impairedby 40 K.Grudziński, R.Kostek the friction force Ft acting in the contact region.The friction force is expressed by the Coulomb law of dry friction as follows if ẋ=0 then |Ft| ¬µ|R| else Ft =−µ|R|sgnẋ (2.2) where µ is a constant friction coefficient and R=R[y(t)] is the resultant nor- mal reaction in the contact region,which dependson the co-ordinate y(t).This co-ordinate describes the approach and separation of the bodies in contact in the normal direction to the sliding surfaces. The deflection of the interface δ= y if y> 0 and δ=0 if y¬ 0; (y< 0 corresponds to a loss of contact). The normal reaction in the contact region is computed from the following formula, see Martins et al. (1990) R(y)= { −Scnym2 for y= δ > 0 0 for y¬ 0 (2.3) where Smeans the nominal contact area between the slider and slideway, and cn,m2 are constant parameterswhich characterise the elastic properties of the contact interface m2 = 1 m1 cn = e −1/m1 n The equation ofmotion to describe the contact freemicro-vibrations of the slider in the normal direction y, is as follows mÿ−R(y)=mg (2.4) where g denotes the acceleration of gravity. The equation of slidingmotion for the slider (Fig.1c) may be expressed in the following general form mẍ−F −Ft =0 (2.5) As equations (2.4) and (2.5) are non-linear, Runge-Kutta’s numerical me- thod of integration of the fourth order was used to calculate the solution. 3. Numerical investigations of normal contact free micro-vibrations Normal contact free vibrations of the slider (Fig.1c) are determined from equation (2.5), considering formula (2.4). For numerical computations, the Influence of normal micro-vibrations in contact... 41 following data were adopted: m=1kg, cn = 4.52 · 1016N/m4, m2 = 2, S = 42.735 · 10−4m2, g = 9.81m/s2. The computations were carried out assuming for t=0 three different initial conditions: (1) y=0 ẏ=0 (2) y= 1 3 y0 ẏ=0 (3) y= 2 3 y0 ẏ=0 where y0 = δ0 =2.25 ·10−7m, and δ0 is the deflection of the interface corre- sponding to the static equilibrium position of the slider. The results of com- putations are presented on graphs in Fig.2 and Fig.3. Fig. 2. A load-deflection characteristic of the interface (a) and results of computer simulations illustrating free (non-damped) normal contact micro-vibrations versus time (b), (c) and on the phase plane (d) for three different amplitudes of vibration Figure 2a shows the load-deflection characteristic of the considered contact interface, where the working point C is marked up, which corresponds to the static equilibrium position of the slider. The frequency of small normal contact vibrations is close to the frequency of linear contact vibrations at that point and may be calculated from equation f0 = (1/2π) √ Kn/m, where Kn = 2 √ mgScn is the contact stiffness at the point C of the characteristic, and m is the mass of the slider. Substituting the above mentioned data to that formula, the frequency of natural normal contact micro-vibrations was determined to be f0 =1485Hz. 42 K.Grudziński, R.Kostek Fig. 3. Normal contact vibrations (a) and contact forces (b) as functions of time The free vibrations are asymmetric with respect to the static equilibrium position. Their natural frequency is dependent not only on the position of the working point C, on the curve (Fig.2a), but on their amplitude as well. As the amplitude rises, the asymmetry of contact vibrations increase and the fre- quency decreases (Fig.2b,c). Figure 2d shows phase portraits of the discussed contact vibrations. To be more precise, the normal micro-vibrations and for- ces acting within the contact region are presented in Fig.3. Figure 3a shows a graph of free undamaged normal contact vibration andFig.3b – time histories of the normal reaction R and friction force Ft that is determined by means of the equation Ft =µ|R| assuming µ=0.1. The contact vibrations make the slider ascend and descend alternately (Fig.3a). The normal reaction R in the contact region, and thus the friction force Ft, areperiodicvariables (Fig.3b).The inertial force originating fromthe vibratingmass of the slider has large influence upon these forces. It causes the normal reaction in the contact region to fluctuate in a wide range of values. In the case concerned (Fig.3), the said reaction varies from nearly 0 up to approximately 3 times of its static value. The specified range within which the dynamic load varies is a result of the maximum free vibrations at which the contact between the sliding surfaces has not been lost yet. The range of variation of the normal reaction for smaller vibrations is shown in Fig.2. 4. Numerical investigations of friction forces and sliding motion For numerical investigations of friction forces andmotion of the slider, the model shown in Fig.1c loaded with a tangential constant drive force F was adopted. If in the said systemnormal contact vibrations are not present, then sliding motion of the slider will take place when F ­Ft0 =µmg (4.1) Influence of normal micro-vibrations in contact... 43 where Ft0 = µmg denotes the nominal static friction force in the absence of contact vibrations. If there are contact vibrations, then slidingmotion of the slidermay occur at lower values of the tangential force F where motion will be intermittent, i.e. of the stick-slip type. The condition for motion to occur may be written down as follows Ft0 >F >−µRmax =µm(g− ÿ)min =Ftmin (4.2) where Ftmin denotes the minimum of the static friction force and ÿ is the acceleration of vibrating motion in the direction normal to the friction plane. The latter dependson the currentamplitudeof contact vibration,whose course is asymmetric relative to the static equilibrium position (Fig.2 and Fig.3). The mean values of the normal reaction and static friction force do not undergo any change here, whereas their instantaneous values do vary within a wide range. In such circumstances, small periodical slip may appear for a driving force F far less than Ft0 determined by formula (4.1). Equation (2.6) was used in order to determine motion of the slider due to a constant tangential force F , under consideration of the periodically variable friction force Ft determined from formulae (2.2) and (2.3). The data for com- putations relevant to the system in questionwas assumed as in Section 3. The computations were made for the tangential force F =0.5N, which is ca. 50% less than the nominal static friction force Ft0 = µmg = 0.981N for the case when there are nonormal contact vibrations in the system.The results of com- puter simulations are shown inFig.4. If there are no contact vibrations (Fig.4, columns 1 and 1’), then the normal reaction R and friction force Ft assume constant values (R=R0 =mg, Ft =Ft0 =µmg), and because F F) and final- ly, when the velocity ẋ drops to zero, the slider stops. Then, typical cyclical stick-slip motion occurs consisting of the sliding and sticking phase (Fig.5). In the sliding phase, besides the traction F and friction force Ft, there is also a force of inertia of themass m. The friction force is described by the formula Ft = F −mẍ. During the sticking phase (when ẋ = 0 and ẍ = 0) the real friction force Ftr = F . This follows from the condition of static equilibrium. Influence of normal micro-vibrations in contact... 45 The potential value of the static friction force during this time (ts) is much higher than F , (Fig.5a,b, thin lines). Fig. 5. Detailed curves of the friction force Ft and velocity ẋ of the slider in stick-slip motion induced by normal contact micro-vibrations with two greatest amplitudes (from Fig.4) Figure 5c shows variations of the friction force Ft in a function of the in- stantaneous velocity ẋ of the slider, and Fig.5f presents a relation between the displacement x and velocity ẋ of sliding motion in the phase plane. Dia- grams in Fig.5c,f concern the sliding phase of intermittent motion. For some determined contact vibrations, stick-slip motion exhibits a stable limit cycle. Times of both individual phases of stick-slip motion are very short and prac- tically unnoticeable. The effect of such micro-slips may not be apparent until some time, e.g. when the nuts on bolts work loose. At greater vibrations, that motion is seen to be continuouswith a constant velocity. The presence of such motion is usually explained as a result of reduction in the friction coefficient or an increased distance between mating surfaces, lower mean values of the normal deflection at the interface, and the normal reaction R as well as real contact area (see Tolstoi et al., 1973; Sakamoto, 1987). In the case concerned, none of these factors is present. The decrease of the friction force and the occurrence of motion are both due to rapidly changing normal displacements and dynamic contact forces excited by vibrations. Figures 4 and 5 show results of simulations for contact vibrations at which the friction surfaces are not yet separated from each other. If the vibrations are of greater amplitudes, there will be a momentary loss of contact, and during this time, the contact forces R and Ft will equal zero. An example of such vibrations and their influence upon the course of sliding motion, where the tangential force F =0.5N is applied, are illustrated by graphs presented 46 K.Grudziński, R.Kostek in Fig.6. Such situations are usually accompanied by specific sound effects (chatter). In theory, any small tangential force F may, in that case, cause motionof the slider in thedirectionof the xaxis. Inpractice, thismaymanifest itself, e.g. in the form of vibrating items drifting on a resting tabletop even if the slope of the latter is very small (nearly zero). Fig. 6. Time histories of normal contact freemicro-vibrationswith amomentary loss of contact (a), (b), (c) and their influence upon the normal reaction R, friction force Ft (d), (e) and kinematic values x,ẋ, ẍ (f), (g), (h) In general, one may conclude that the occurrence of sliding motion of the slider and its behaviour in time depend on the value of the tangential force F (which is lower than the static friction force Ft0 =µmg) and amplitude A of normal contact vibrations. Figure 7 presents the influence of these factors on themean velocity v= ẋ of slidingmotion for µ=0.1. If ymin ­ 0, the bodies remain in contact all the time (Fig.4 and Fig.5), and if ymin < 0 it leads to a periodical loss of contact between the sliding surfaces (Fig.6). The possibility of amomentary loss of contact between the sliding surfaces has been taken into account in the computations. In Figure 7, the area of variations of controllable data (F and A) is noticeable, inwhich slidingmotion does not occur (v = 0). This takes place for appropriately small amplitudes of vibrations and values of the tangential force. If the vibration amplitude is A > 0.225µm, which was the case, then some momentary losses of contact Influence of normal micro-vibrations in contact... 47 will occur, and the sliderwillmove even for very small values of the tangential force F (Fig.7). Fig. 7. Influence of the tangential force F and amplitude A of normal micro-vibrations in the contact area on the mean velocity of sliding Motion of the slider with a constant mean velocity v for F 0. Should F > Ft0, then the slider will move with an acceleration dependent, to a certain extent, on the current amplitude of contact vibrations. These cases are of little interest and were not analysed in detail. 5. Conclusions The adopted model as well as simulation tests carried out in this study both provide theoretical justification and show the mechanism of reduction of the friction force for planar contact of solids if normal micro-vibrations take place. It has been shown that the influence that normal contact micro- vibrations have on the reduction of the friction force, which is often observed and confirmed in many experiments, does not necessarily have to have any connection with the decrease either in the friction coefficient or the mean value of the normal reaction in contact. This influence primarily results from periodic dynamic processes that take place in the frictional contact. In order to properly understand and describe these processes, it is necessary to adopt the frictional contact model with at least two degrees of freedom that would take into account the non-linear normal flexibility of the rough surfaces. 48 K.Grudziński, R.Kostek Acknowledgement The authors gratefully acknowledge the financial support of this work by the Volkswagen-Foundationwithin the framework of themain research project Investiga- tion of nonlinear dynamical effects in production systems. References 1. Back N., Burdekin M., Cowley A., 1973, Review of research on fixed and sliding joints,Proc. 13th International Machine Tool Design and Research Conference (Edited by S.A. Tobias and Koenigsberger), Macmillan, London, 87-97 2. Broniec Z., Lenkiewicz W., 1982, Static friction processes under dynamic loads and vibration,Wear, 80, 261-271 3. BudanovB.V.,KudinovV.A., Tolstoi D.M., 1980, Interaction of friction and vibration, Sov. J. Friction and Wear, 1, 1, 79-89 4. Dolby M., Bell R., 1971, The contact stiffness of joints at low apparent interface pressure,Annals of CIRP, 19, 67-79 5. Godfrey D., 1967, Vibration reduces metal to metal contact and causes an apparent reduction in friction,ASLE Trans., 10, 183-192 6. GrudzińskiK.,WedmanS., 1999,Asimulation investigation into self-excited stick-slip motion in a mass-spring system under variable normal load, Tenth World Congress on the Theory of Machine andMechanisms, Oulu, Finland, 4, 1662-1667 7. Grudziński K., Konowalski K., Kostek R., 2000, Untersuchung norma- ler Schwingungen der gefügten Körper bei Führungssystemen von Maschinen, International Congress MATAR, Praha, Section 1, 236-242 8. GrudzińskiK.,KostekR., 2003,Modelowanie i analizanieliniowychkontak- towych mikrodrgań normalnych wymuszonych siłą harmoniczną, Zeszyty Na- ukowe Katedry Mechaniki Stosowanej Politechniki Śląskiej, 20, 141-146 9. Hess D.P., Soom J.T., 1991, Normal vibrations and friction under harmonic load:Part I –Hertzian contacts,Part II -Roughplanar contacts,Trans.ASME, Journal of Tribology, 113, 80-92 10. Hess D.P., Wagh N.J., 1995, Evaluating surface roughness from contact vi- brations,Trans. ASME, Journal of Tribology, 117, 60-64 11. Kohei I., KyosukeO., 2001,Analysis of bouncing vibration of 2-DOFmodel of tripad contact slider overa randomwavydisk surface,Trans.ASME, Journal of Tribology, 123, 159-167 Influence of normal micro-vibrations in contact... 49 12. LevinaZ.M., 1965,Calculationof contactdeformationon slideways,Machines and Tooling, 36, 1, 1-8 13. Martins J.A.C., Oden J.T., Simões F.M.F., 1990, A study of static and kinetic friction, Int. J. Engng. Sci., 28, 1, 29-92 14. Ostrovski V.I., 1965, The influence of machining tools on slideway contact stiffness,Machines and Tooling, 36, 1, 17-19 15. Pfeiffer F., 1999, Unilateral multibody dynamics, 2000 Kluwer Academic Publishers, Netherland,Meccanica, 34, 437-451 16. SakamotoT., 1987,Normaldisplacementanddynamic frictioncharacteristics in a stick-slip process,Trib. Intl., 20, 25-31 17. Skåre D.M., St̊ahl J.E., 1992, Static and dynamic friction processes under the influence of external vibrations,Wear, 154, 177-192 18. Tolstoi D.M., 1967, Significance of normal degree of freedom and natural normal vibration in contact friction,Wear, 10, 199-213 19. Tolstoi D.M., BorisovaG.A.,Grigorova S.R., 1973, Friction regulation by perpendicular oscillation, Soviet Phisics-Doklady, 17, 9, 907-909 20. VDI-Bericht 1736, 2002,Reibungund Schwingungen inFahrzeugen,Maschinen und Anlagen, VDI-Geselschaft EntwicklungKonstruktion Vertrieb, Düsseldorf Wpływ mikrodrgań normalnych w kontakcie na ruch ślizgowy ciała stałego Streszczenie W artykule przedstawiono wyniki symulacyjnych badań mikrodrgań w kierunku normalnymwkontakcie ciał stałych o powierzchniach chropowatych oraz ichwpływu na redukcję siły tarcia. Przyjęto model o dwóch stopniach swobody, uwzględniający nieliniową podatność kontaktową normalną oddziaływujących na siebie powierzchni. Zbadano i opisano mechanizm powodujący zmniejszenie oporu tarcia w kontakcie, w przypadkuwystępowaniaw nimdrgań normalnych.Wykazano, że głównąprzyczy- ną redukcji oporu tarcia w takim przypadku są pewne procesy dynamiczne występu- jącewkontakcie, a nie spadkiwartościwspółczynnika tarcia, ani też średniej wartości reakcji normalnej, jak to się podaje w wielu publikacjach. Manuscript received March 2, 2004; accepted for print July 14, 2004