CP620, Shock Compression of Condensed Matter - 2001 edited by M. D. Furnish, N. N. Thadhani, and Y. Horie © 2002 American Institute of Physics 0-7354-0068-7/02/$ 19.00 NON-NEWTONIAN VISCOSITY EFFECTS AT SHOCKED FLUID INTERFACES * STEVEN M. VALONE Materials Science and Technology Division, Los Alamos National Laboratory, Los Alamos, NM 87545 Abstract. The regime under consideration here is one in which a polymeric material is shocked strongly enough to behave as a fluid, but not as a simple fluid. The present work takes a case-study approach in which several models of shear-rate-dependent viscosities are applied to a RichtmyerMeshkov shock instability at high-viscosity polymeric interface. The early instability growth regime is modeled with the viscosity dependence of Mikaelian and the late growth regime is modeled with an aerodynamic viscous-drag model. Shear-rate dependencies of the viscosities can change the shape of the early growth, while altering the decay in growth rate at later times. Both of these effects should be observable in experiments. The results suggest that the shear-rate relaxation times will need to be known to better than an order of magnitude for prediction purposes. INTRODUCTION Inviscid, compressible shock passage The subject of hydrodynamic instabilities in shocked fluid interfaces began very soon after shock physics became a field of inquiry. The fluids studied in early investigations were primarily gases (1,2), treated as incompressible, inviscid materials. Several later efforts (3,4) introduced viscous effects into the evolution of the instability fronts. It was only in the early to middle 1990's that compressibility effects (5) and drag effects (6,7) were introduced. However, a general picture of the flow characteristics associated with the Richtmyer-Meshkov instability (8,9) emerged. The evolution is broken down into three stages (Fig. 1): An initial, inviscid compressible stage associated with passage of the shock wave through the material interface; a front growth stage which is linear in the perturbation amplitude, if treated inviscidly, and exponentially in time, if treated with Newtonian viscosity (10); and an incompressible flow stage characterized by hydrodynamic drag. The second stage will be growth rate Early, viscous incompressible \ Late, viscous incompressible asymptote Figure 1. Growth stages of a material interface instability as a shock wave passes from the light fluid to the heavy one. referred to as the early stage and the third stage will be referred as the late stage. Presently, instability growth in viscous, compressible materials have not been treated at the level of (e.g) ZhangandSohn(ll). The evolution equations for the instability growth rates are valid for non-Newtonian fluids. It will always be assumed that the evolution begins at the early stage. We will study the results of intro- * Work supported by the US Department of Energy under contract W-7405-ENG-36 to the University of California. 295 proposed and lead to different asymptotic behavior in the instability growth rate of the peak amplitude. In these models, it is necessary to separate the evolution of the spikes (heavy fluid) and the bubbles (light fluid). All of the models take the form ducing non-Newtonian viscosity into existing interface evolution models. The purpose of this inquiry is to gauge the sensitivity of the evolution to the viscosity model and, specifically to shear rate, and to discern whether or not non-Newtonian effects should be experimentally observable. h"=-Ch'lhVA, MODELS where A is some characteristic length-scale for the perturbation and C is some function of r| and A. The models differ in their assignments for A and C. For A, Alon et al. (7) (AHOS model) set it to the ratio of volume-to-surf ace area which is ~A,. The asymptotic decay for the instability growth rate goes as 1/t. Dimonte and Schneider (12) (DS model) and Cheng et al. (13) (COS model) set A = h. The asymptotic decay then ranges between 1 and 1/t. For C, CAHOS = cd fAHOs(A)> where cd is a dimensionless drag coefficient and fAHOs(A) = * for bubbles and (1-A)/(1+A) for spikes. The DG model takes CDS = cd with no dependence on A. Finally, the CGS model (14) takes C = cd.fCGS(A), where fCGS(A) = (l-A)/2 for bubbles and (l+A)/2 for spikes. The presence of A couples the spike and bubble flows. The model used for cd is that for simple laminar flow around a sphere, cd = 24/Re, where Re = (A/2)2e'p/r| and s' = (h'spike SDike + h\bubble If TI is constant in time, Eqs. (3-4) can be integrated analytically. For arbitrary t|, Eqs. (3-4) are simple to integrate numerically. To date, only constant r| have been considered (7-13) Viscosity Models: The inviscid, T| = 0, and constant (zero-shear-rate) viscosity, r| = TJQ, models have been mentioned in regards to earlier work. We consider three shear-rate-dependent models which have been used in various applications of polymer viscosity modeling. No attempt is made here to find optimal models for particular applications or to be comprehensive in covering all important types of models. The two models are associated with Bird (14) andGraessley (15) Early-time Growth Models: The original RM model takes the growth rate as constant after the passage of the shock wave. Letting h be the peak amplitude of a single-mode instability and h' its time derivative, the growth-rate constant is h' = Ah 0 kAu, (1) where A is the Atwood number, h0 is the initial amplitude of the perturbation in the interface, k is the wavenumber of the perturbation, and Au is the velocity jump after passage of the shock. The Atwood number is the ratio = (p 2 -Pi)/(p 2 + P i ) > (2) where p j is the density of the material which is shocked first and p2 is the density of the material shocked second as the shock front traverses the interface. The wavenumber k = 2n/K, where A, is the wavelength of the perturbation. Mikaelian (3, 10) derives a viscosity dependence to the early-time growth by analyzing the lowest-order moment expansion of the momentum equation for incompressible flow: h' = Ah0 k Au exp(-k2T|t), (4) (3) where r| is the viscosity at time t. Mikaelian considers Newtonian viscosity only. The growth rate is further modified here with the decay modeled by Zhang and Sohn (11). This modification entails dividing the RHS of Eq. (3) by a polynomial function of A, k, and t. The early-time models are considered to be valid during the time required for the perturbation to grow to ~ 0.1 A. Eqs. (1) and 3 are considered valid after the compression time, estimated as the shock-passage time, ho/D, where D is the shock speed. Late-time Drag Models: Several different versions of the drag evolution equations have been a 296 (5) and (6) respectively, where TJ^ is the infinite shear-rate viscosity, £' is the shear rate, T is a characteristic relaxation time, and a is a constant. The function f in Eq. (6) is a given in (15). The transcendental equation must be solved for each value of T|G. Fig. 2 illustrates the shear-rate dependence of the three models. The Eyring (16) and Graessley models are fairly similar. The Bird model can be calibrated as needed for work hardening or shear thinning. 0.00 0.00 0.05 0.10 0.15 0.20 0.25 t(ns) Figure 3. Shear-rate dependence of the early-time growth. T is a relaxation time. 0.1 E '•5 0.01 - 3 - 2 - 1 0 1 2 3 Iog10i 5s/oi(dimensionless) Figure 2. Shear-rate dependence of various viscosity models mentioned in the text. I is a relaxation time. o.oo 0.330 0.335 0.340 0.345 0.350 0.355 0.360 t(ns) SIMULATIONS AND RESULTS Figure 4. Shear-rate dependence of the DS growth rate. I is a relaxation time. Applying the Newtonian, Bird, Graessley viscosity models to both the early and late stages illustrates the differences and capabilities of the models. Only the DS drag model will be used. The early-time first-order ODE and the latetime second-order ODEs are integrated with a 5thorder Runga-Kutta algorithm. The same initial amplitude, aQ = 0.1 Jim, change in shock speed across the interface, Au = 1 km/s, and T|0 = 103 were used in all cases. See the results in Figs. 3 and 4. At early times, the viscosity model dramatically influences the shape of the growth rate curve. High shear-rate effects cause even high-viscosity fluids (in terms of TIO) to behave inviscidly, making it possible for them to exhibit interface instability. The duration of the growth is strongly influenced by both T and r|0. During latestage growth, i and T|0 control the decay rate in the growth profile. These results suggest that T and r|0 determined for one growth/viscosity model combination are unlikely to be usable for another model combination. Newtonian viscosity models, as has always been suspected, seem entirely inappropriate over any significant range of shock strengths. Order-of-magnitude changes in either T or r|0 297 8. should be readily observable in experiments on either growth stage. The ID-ODE approaches for modeling growth may be valuable as a viscositymodel screening tool . The justification requires a posteriori comparison with experiment. 9. REFERENCES 1. 2. 3. 4. 5. 6. 7. Lord Rayleigh, "Investigation of the Character of the Equilibrium of an Incompressible Heavy Fluid of Variable Density," Scientific Papers Vol. //, (Cambridge University Press, Cambridge, England, 1900), p. 200. G. I. Taylor, "The Instability of Liquid Surfaces When Accelerated in a Direction Perpendicular to Their Plane," Proc. Roy. Soc. A 201, 192 (1950). S. Chandrasekhar, "The Character of the Equilibrium of an Incompressible Heavy Viscous Fluid of Variable Density," Proc. Camb. Phil. Soc. 51,162(1955). D. E. Grady, "Strain-rate Dependence of the Effective Viscosity Under Steady-Wave Shock Compression," Appl. Phys. Lett. 38, 825 (1981). Y. Yang, Q. Zhang, and D. H. Sharp, "Small Amplitude Theory of Richtmyer-Meshkov Instability," Phys. Fluids 6, 1856 (1994). J. Hecht, U. Alon, and D. Shvarts, "Potential Flow Models of Rayleigh-Taylor and RichtmyerMeshkov Bubble Fronts," Phys. Fluids 6, 4019 (1994). U. Alon, J. Hecht, D. Ofer, and D. Shvarts, "Power Laws and Similarity of Rayleigh-Taylor and Richtmyer-Meshkov Mixing Fronts at All Density Ratios," Phys. Rev. Lett. 74, 534 (1995). 10. 11. 12. 13. 14. 15. 16. 298 R. D. Ritchmyer, "Taylor Instability in Shock Acceleration of Compressible Fluids," University of California, Los Alamos Scientific Laboratory Report LA-1914, July (1954) and Commun. Pure Appl. Math. 13, 297 (1960). E. E. Meshkov, "Instability of the Interface of Two Gases Accelerated by a Shock," Fluid Dyn. 43, 101 (1969); "Instability of a Shock Wave Accelerated Interface Between Two Gases," NASA Tech. Trans., NASA TTF-13, 074 (1970). K. Mikaelian, "Effect of Viscosity on RayleighTaylor and Richtmyer-Meshkov Instabilities," Phys. Rev. £47,375(1993). Q. Zhang and S.-I. Sohn, "Nonlinear Theory of Unstable Fluid Mixing Driven by Shock Wave," Phys. Fluids 9, 1106(1997). G. Dimonte and M.Schneider, "Density Ratio Dependence of Rayleigh-Taylor Mixing for Sustained and Impulsive Acceleration Histories," Phys. Fluids 12, 1070 (2000). B. Cheng, J. Glimm, and D. H. Sharp, "Density Dependence of Rayleigh-Taylor and RichtmyerMeshkov Mixing Fronts," Phys. Lett. A 268, 366 (2000). S. G. Bardenhagen, M. G. Stout, and G. T. Gray, "Three-Dimensional, Finite Deformation, Viscoplastic Constitutive Models for Polymeric Materials," Mech. Mater. 25, 235 (1997). W. M. Graessley, "Viscosity of Entangling Polydisperse Polymers," J. Chem. Phys. 47, 1942 (1967). H. Eyring, "Viscosity, Plasticity, and Diffusion as Examples of Absolute Reaction Rates," J. Chem. Phys. 4, 283 (1936).
© Copyright 2026 Paperzz