Abstract
Epistemic uncertainty, the uncertainty in the physical model used to represent a phenomenon, has a significant effect on the predictions of simulations of mechanical systems, particularly in systems with impact events. Impact dynamics can have a significant effect on a system’s functionality, stability, wear, and failure. Because high-fidelity models of systems with impacts often are too computationally intensive to be useful as design tools, rigid body dynamics and reduced order model simulations are used often, with the impact events modeled by ad hoc methods such as a constant coefficient of restitution or penalty stiffness. The choice of impact model, though, can have significant ramifications on design predictions. The effects of both epistemic and aleatoric (parametric) uncertainty in the choice of contact model are investigated in this paper for a representative multiple-degree of freedom mechanical system. Six contact models are considered in the analysis: two different constant coefficient of restitution models, a piecewise-linear stiffness and damping (i.e., Kelvin–Voight) model, two similar elastic-plastic constitutive models, and one dissimilar elastic-plastic constitutive model. Results show that the optimal mechanism design for each contact model appears extremely different. Further, the effects due to epistemic uncertainty are differentiated clearly in the response from the effects due to aleatoric uncertainty. Lastly, when the mechanisms are optimized to be robust against aleatoric uncertainty, the resulting designs show some robustness against epistemic uncertainty.
Similar content being viewed by others
References
Brake, M.R.: An analytical elastic-perfectly plastic contact model. Int. J. Solids Struct. 49, 3129–3141 (2012)
Brake, M.R.: The effect of the contact model on the impact-vibration response of continuous and discrete systems. J. Sound Vib. 332, 3849–3878 (2013)
Blankenship, G.W., Kahraman, A.: Steady state forced response of a mechanical oscillator with combined parametric excitation and clearance type non-linearity. J. Sound Vib. 185, 743–765 (1995)
Jerrelind, J., Stensson, A.: Nonlinear dynamics of parts in engineering systems. Chaos Solitons Fractals 11, 2413–2428 (2000)
Nayeri, R.D., Masri, S.F., Caffrey, J.P.: Studies of the performance of multi-unit impact damers under stochastic excitation. ASME J. Vib. Acoust. 129, 239–251 (2007)
Païdoussis, M.P., Li, G.X.: Cross-flow-induced chaotic vibrations of heat-exchanger tubes impacting on loose supports. J. Sound Vib. 152, 305–326 (1992)
Zuo, L., Curnier, A.: Non-linear real and complex modes of conewise linear systems. J. Sound Vib. 174, 289–313 (1994)
Blazejczyk-Okolewska, B., Czolczynski, K., Kapitaniak, T.: Classification principles of types of mechanical systems with impacts—fundamental assumptions and rules. Eur. J. Mech. A 23, 517–537 (2004)
Ervin, E.K., Wickert, J.A.: Repetitive impact response of a beam structure subjected to harmonic base excitation. J. Sound Vib. 307, 2–19 (2007)
Shaw, S.W.: The dynamics of a harmonically excited system having rigid amplitude constraints part 2: chaotic motions and global bifurcations. ASME J. Appl. Mech. 52, 459–464 (1985)
Krylov, S., Barnea, D.I.: Bouncing mode electrostatically actuated scanning micromirror for video applications. Smart Mater. Struct. 14, 1281–1296 (2005)
Deshpande, S., Mehta, S., Jazar, G.N.: Optimization of secondary suspension of piecewise linear vibration isolation systems. Int. J. Mech. Sci. 48, 341–377 (2006)
Ramírez, R., Pöschel, T., Brilliantov, N.V., Schwager, T.: Coefficient of restitution of colliding viscoelastic spheres. Phys. Rev. E 60, 4465–4472 (1999)
Ismail, K.A., Stronge, W.J.: Impact of viscoplastic bodies: dissipation and restitution. ASME J. Appl. Mech. 75, 061011-1–061011-5 (2008)
Stronge, W.J.: Impacts in mechanical systems: analysis and modelling. Contact problems for elasto-plastic impact in multi-body systems, vol. 551, pp. 189–234. Springer, Heidelberg (2000)
Zhang, X., Vu-Quoc, L.: Modeling the dependence of the coefficient of restitution on the impact velocity in elasto-plastic collisions. Int.J. Impact Eng. 27, 317–341 (2002)
Jackson, R.L., Green, I., Marghitu, D.B.: Predicting the coefficient of restitution of impacting elastic-perfectly plastic spheres. Nonlinear Dyn. 60, 217–229 (2010)
Kogut, L., Etsion, I.: Elastic-plastic contact analysis of a sphere and a rigid flat. ASME J. Appl. Mech. 69, 657–662 (2002)
Etsion, I., Kligerman, Y., Kadin, Y.: Unloading of an elastic-plastic loaded spherical contact. Int. J. Solids Struct. 42, 3716–3729 (2005)
Thornton, C.: Coefficient of restitution for collinear collisions of elastic-perfectly plastic spheres. ASME J. Appl. Mech. 64, 383–386 (1997)
Wagg, D.J., Bishop, S.R.: A note on modelling multi-degree-of-freedom vibro-impact systems using coefficient of restitution models. J. Sound Vib. 236, 176–184 (2000)
Tatara, Y., Moriwaki, N.: Study on impact of equivalent two bodies (coefficients of restitution of spheres of brass, lead, glass, porcelain and agate, and the material properties). Bull. JSME 25, 631–637 (1982)
Kharaz, A.H., Gorham, D.A.: A study of the restitution coefficient in elastic-plastic impact. Philos. Mag. A 80, 549–559 (2000)
Minamoto, H., Kawamura, S.: Effects of material strain rate sensitivity in low speed impact between two identical spheres. Int. J. Impact Eng. 36, 680–686 (2009)
Minamoto, H., Kawamura, S.: Moderately high speed impact of two identical spheres. Int. J. Impact Eng. 38, 123–129 (2011)
Rigaud, E., Perret-Liaudet, J.: Experiments and numerical results on non-linear vibrations of an impacting Hertzian contact. part 1: Harmonic excitation. J. Sound Vib. 265, 289–307 (2003)
Brake, M. R., Reu, P. L., VanGoethem, D. J., Bejarano, M. V., Sumali, A.: Experimental validation of an elastic-plastic contact model, in: ASME 2011 International Mechanical Engineering Congress and Exposition, Denver, CO (2011)
Du, Y., Wang, S.: Energy dissipation in normal elastoplastic impact between two spheres. ASME J. Appl. Mech. 76, 061010-1–061010-8 (2009)
Jackson, R.L., Green, I.: A finite element study of elasto-plastic hemispherical contact against a rigid flat. ASME J. Tribol. 127, 343–354 (2005)
Jackson, R.L., Chusoipin, I., Green, I.: A finite element study of the residual stress and deformation in hemispherical contacts. ASME J. Tribol. 127, 484–493 (2005)
Brake, M.R.: IMEX-a: an adaptive, fifth order implicit–explicit integration scheme, SAND2013-4299. Sandia National Laboratories, Albuquerque (2013)
Ascher, U.M., Ruuth, S.J., Spiteri, R.J.: Implicit-explicit Runge-Kutta methods for time-dependent partial differential equations. Appl. Numer. Math. 25, 151–167 (1997)
Frick, T. M., Sobek, E., Reavis, J. R.: Overview on the development and implementation of methodologies to compute vibration and wear of steam generator tubes, in: ASME symposium on flow-Induced vibrations in heat exchanges, New Orleans, LA, (1984)
Knudsen, J., Massih, A.R.: Vibro-impact dynamics of a periodically forced beam. ASME J. Press. Vessel Technol. 122, 210–221 (2000)
Knudsen, J., Massih, A.R.: Dynamic stability of weakly damped oscillators with elastic impacts and wear. J. Sound Vib. 263, 175–204 (2003)
Knudsen, J., Massih, A.R.: Impact oscillations and wear of loosely supported rod subject to harmonic load. J. Sound Vib. 278, 1025–1050 (2004)
Helton, J.C., Davis, F.J.: Latin hypercube sampling and the propagation of uncertainty in analyses of complex systems, SAND2001-0417. Sandia National Laboratories, Albuquerque (2001)
Mühlenbein, H., Schomisch, M., Born, J.: The parallel genetic algorithm as function optimizer. Parallel Comput. 17, 619–632 (1991)
Chipperfield, A. J., Fleming, P. J.: The MATLAB genetic algorithm toolbox, IEE Colloquium on Applied Control Techniques Using MATLAB London, UK, January 1995 (1995) 10/1-10/4
Houck, C.R., Joines, J.A., Kay, M.G.: Comparison of genetic algorithms, random restart, and two-opt switching for solving large location-allocation problems. Comput. Oper. Res. 23, 587–596 (1996)
Johnson, K.L.: Contact Mech. Cambridge University Press, Cambridge (1985)
Hertz, H.: Über die berührung fester elastischer körper (On the contact of elastic solids). Journal fur die Reine und Andgewandte Mathematik 92, 156–171 (1882)
Tabor, D.: A simple theory of static and dynamic hardness. Proc. R. Soc. Lond. A 192, 247–274 (1948)
Becker, E.B., Carey, G.F., Oden, J.T.: Finite elements: an introduction. Prentice-Hall, Englewood Cliffs (1981)
Yan, S.L., Li, L.Y.: Finite element analysis of cyclic indentation of an elastic-perfectly plastic half-space by a rigid sphere. Proc. Inst. Mech. Eng. Part C 217, 505–514 (2003)
Acknowledgments
The author gratefully acknowledges the support and collaboration of his colleagues, especially Dannelle Aragon, Doug VanGoethem, Anton Sumali, and Jill Blecke.
Author information
Authors and Affiliations
Corresponding author
Additional information
Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporations, for the U.S. Department of Energy’s National Nuclear Security Administration under Contract DE-AC04-94AL85000.
Appendix: Contact model descriptions
Appendix: Contact model descriptions
For each of the contact models, the geometry shown in Fig. 17 is used: the models are derived based on the contact between two spheres (or a sphere and a flat, which is represented as a sphere with an infinite radius), and the primary degree of freedom is the contact displacement \(\delta \), which is the relative displacement of the center of mass of the two spheres. In Fig. 17, the illustrative drawing uses a reference frame in which the center of mass of the second sphere is not changed, and thus \(\delta \) is defined as the displacement of the center of mass of the first sphere. The elastic-plastic contact models assumes that the two elastic-perfectly plastic spheres (labeled by \(i=1\), 2) with radii \(r_i\) shown in Fig. 17 have elastic modulus \(E_i\), Poisson’s ratio \(\nu _i\), yield strength \(\sigma _{Yi}\), density \(\rho _i\), and Brinell hardness \(H_i\), are smooth such that their asperities are small compared to the displacements (specifically, that the asperity heights are much smaller than the yield displacement \(\delta _y\), defined in what follows), and that friction is negligible. The normal displacement between the two spheres is \(\delta \), and the contact radius is \(a\). The bulk parameters used to represent the spheres are given as the effective elastic modulus \(E=((1-\nu _1^2)/E_1+(1-\nu _2^2)/E_2)^{-1}\) and radius \(r=(r_1^{-1}+r_2^{-1})^{-1}\) [41], and the yield strength \(\sigma _y=\min \left[ \sigma _{y1},\sigma _{y2}\right] \).
1.1 Coefficient of restitution contact model
The coefficient of restitution \(e\) is a useful measure of the energy dissipation from an impact event at a given velocity. For two bodies with velocities just before an impact of \(v_{1i}\) and \(v_{2i}\), the velocities just after the impact \(v_{2o}\) and \(v_{1o}\) are governed by
This equation assumes that the two bodies have mass \(m_1\) and \(m_2\), that the impact is instantaneous, and that the energy dissipation of impact is independent of impact velocity (thus a constant \(e\)).
1.2 Piecewise-linear contact model
The penalty method of contact is a Kelvin–Voight model described by a piecewise-linear stiffness and damping
where \(\dot{\delta }\) is the derivative of \(\delta \) with respect to time. The stiffness \(K_P=10\hbox { kN/mm}\) and damping \(C_P=280\hbox { Ns/m}\) are chosen to represent a sufficiently stiff impact that is tuned to match the energy dissipation predictions of Brake’s contact model. The value of 10 kN/mm is a commonly used approximation in many finite element simulations and is several orders of magnitude larger than the other stiffnesses in the system investigated. The damping coefficient is calibrated to match the energy dissipation predictions of Brake’s contact model at velocities above 4 m/s.
1.3 Brake’s contact model
The formulation for Brake’s contact model is given in [1] and summarized here. Contact is divided into four regimes: elastic, mixed elastic-plastic, fully plastic, and unloading.
1.3.1 The elastic regime
The elastic regime for each of the elastic-plastic models discussed all use Hertz’s solution [42] for the constitutive relationship between the spheres
with the contact radius determined by \(a=\sqrt{r\delta }\). The elastic regime spans from the initiation of contact (\(\delta =0\)) until yield (\(\delta =\delta _y\)). In Brake’s contact model, yield is determined based on the stress field that develops in elastic contact between two bodies [41] and the von Mises’ criterion. Defining the maximum amplitude of the stress field
over the distance \(z\) into the surface, where \(\nu \) is Poisson’s ratio for the material that will yield first,
1.3.2 The plastic regime
In the plastic regime, it is assumed that the pressure \(p_0\) over the contact area is uniform and there is a linear relationship between the contact force and \(\delta \)
Since both the contact radius \(a_p\) and displacement \(\delta _p\) at the fully plastic regime’s inception are fixed in Eqs. 16 and 17, \(\xi \) is chosen for compatibility. The maximum contact pressure is based on the definition of hardness given in [43] \(p_0=Hg10^6\), where \(g\) is acceleration due to gravity and is necessary for the unit conversion from the Brinell hardness \(H\) (given in \(\hbox {kgf/mm}^2\)) to Pa, and is defined by the hardness of both materials involved in the contact
The contact radius \(a_p\) and displacement \(\delta _p\) at the onset of fully plastic behavior are then defined via
The incipient plastic contact force is thus defined as \(F_p=p_0\pi a_p^2\).
1.3.3 The mixed elastic-plastic regime
To model the mixed elastic-plastic regime, the set of cubic Hermite polynomials [44] is used to enforce continuity between the end of the elastic regime and the beginning of the plastic regime
These polynomials are chosen for convenience; other sets of splines that are \(\mathcal {C}^1\) continuous would also be suitable. The net effect of these equations is to enforce both \(\mathcal {C}^0\) and \(\mathcal {C}^1\) continuity for the relationship between \(\delta \) and both \(F_C\) and \(a\). The \(\cdot ^\prime \) indicates a derivative with respect to \(\delta \).
1.3.4 The restitution phase
Unloading is assumed to be an elastic process [15, 17, 45] with no reverse yielding. In each regime, unloading follows
and \(\bar{\delta }\) is related to \(\bar{r}\) by enforcing continuity. Defining the maximum displacement and maximum force during the loading phase as \(\delta _m\) and \(F_m\), respectively,
The physical interpretation of \(\bar{r}\) and \(\bar{\delta }\) is that the contacting surfaces deform both in terms of the radius of curvature over the contact area (\(\bar{r}\)), and in terms of a residual deformation (\(\bar{\delta }\)).
Unloading from the elastic regime: Impacts that stay purely within the elastic regime result in no plastic deformation. As a result \(\bar{r}=r\) and \(\bar{\delta }=0\).
Unloading from the plastic regime: The contact radius is assumed to be approximately the same as the radius of a spherical indentation with depth \(\delta _m\) and width \(2a_m\). From the trigonometric properties of a chord
and \(\bar{\delta }\) is found via Eq. 26.
Unloading from the mixed elastic-plastic regime: The deformed radius is approximated using the cubic Hermite polynomials again. In this case, the deformed radius at \(\delta _m=\delta _y\) is \(r\) by definition and has the derivative with respect to \(\delta \) of 0. Likewise, the deformed radius at \(\delta _m=\delta _p\) is defined as \(\bar{r}_p\) and has derivative with respect to \(\delta \) of \(\bar{r}_p^\prime \), which is found numerically using a forward difference calculation. This yields
1.4 Etsion’s contact model
Etsion’s contact model [18, 19] uses the same Hertzian loading process as Brake’s contact model in the elastic zone. Following the onset of yield, correlations derived from high-fidelity finite element simulations are used to model the mixed elastic-plastic and the plastic regimes. Thus, the coefficients in Eqs. 30 through 36 are determined from best-fit curves applied to the results of high-fidelity finite element simulations in [18, 19]. The displacement at the inception of yield
is defined in terms of the hardness \(H\) of the more compliant material and the hardness coefficient \(K=0.454+0.41\nu \) (where \(\nu \) is the Poisson’s ratio of the more compliant material).
1.4.1 Loading after yield
The analysis in [18] observes that there are three distinct regions following the inception of yield: the development of the plastic region below the contact surface in which the entire contact surface is still elastic (for \(1\le \delta /\delta _y\le 6\)), the expansion of the plastic region to include an annular plastic region on the contact surface (\(6\le \delta /\delta _y\le 68\)), and a fully plastic contact area (\(68\le \delta /\delta _y\)). In the regime where the extent of the plastic deformation is entirely below the surface
where \(a_y\) and \(F_y\) are the contact radius and contact force at yield (\(\delta =\delta _y\)), and for the regimes where the plastic deformation extends to the contact surface
1.4.2 Unloading
Similar to the loading model is developed, the unloading model is based on best-fit correlations to numerical simulations [19]. Defining the residual interference \(\delta _r\) as the permanent deformation due to plasticity, the correlations
are developed with respect to the maximum contact force \(F_m\), contact radius \(a_m\), and deformation \(\delta _m\) during loading.
1.5 Thornton’s contact model
Thornton’s contact model [20] is divided into two loading regimes: elastic and plastic. The contact model uses the same Hertzian solution for the elastic regime as both Brake’s and Etsion’s contact models. The onset of yield is determined by equating the kinetic energy of a sphere before impact with the work done during elastic loading to the point of yield, which is defined by the contact pressure equalling the yield stress
1.5.1 Loading in the plastic regime
Thornton’s model does not consider a transition from elastic to plastic behavior. The elastic regime is defined to abruptly transition to the plastic regime for the yield force
The plastic contact force is then found by considering a Hertzian pressure distribution that is truncated for all pressures above the yield stress
1.5.2 Unloading
Unloading for Thornton’s contact model is an elastic process. In the elastic regime, there is no plastic deformation, and thus the unloading curve is identical to the loading curve. In the plastic regime, the deformed radius of curvature \(\bar{r}\) and residual deformation \(\bar{\delta }\) are related to the maximum contact force during loading \(F_m\) and maximum displacement \(\delta _m\) via
Rights and permissions
About this article
Cite this article
Brake, M.R. The role of epistemic uncertainty of contact models in the design and optimization of mechanical systems with aleatoric uncertainty. Nonlinear Dyn 77, 899–922 (2014). https://doi.org/10.1007/s11071-014-1350-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11071-014-1350-0