Inviscid stability analysis of parallel bubbly flows Suhas S Jain - - PowerPoint PPT Presentation

inviscid stability analysis of parallel bubbly flows
SMART_READER_LITE
LIVE PREVIEW

Inviscid stability analysis of parallel bubbly flows Suhas S Jain - - PowerPoint PPT Presentation

Introduction and formulation Stability analysis Inviscid stability analysis of parallel bubbly flows Suhas S Jain ME451B Final Project Presentation, Stanford University Wednesday 12 th December, 2018 Suhas S Jain Stability of bubbly flows 1


slide-1
SLIDE 1

Introduction and formulation Stability analysis

Inviscid stability analysis of parallel bubbly flows

Suhas S Jain

ME451B Final Project Presentation, Stanford University

Wednesday 12th December, 2018

Suhas S Jain Stability of bubbly flows 1 / 23

slide-2
SLIDE 2

Introduction and formulation Stability analysis

Introduction and Motivation

motivation bubbly flows are ubiquitous in nature. even at low void fractions, their presence can significantly change

sound speed attenuation characteristics inertia of the medium

crucial to understand the dynamical properties of the medium.

  • bjectives

derive governing equations and disturbance relations for bubble-liquid mixture. perform stability analysis of spacewise problem of inviscid bubbly shear flow, following d’Agostino et al., JFM, 1997. study the effect of presence of bubbles.

Suhas S Jain Stability of bubbly flows 2 / 23

slide-3
SLIDE 3

Introduction and formulation Stability analysis

Basic equations

Individual phase continuity equation (IPCE): ∂ρiαi ∂t + ∇ · (ρi uiαi) = 0, ρi and αi are the density and volume fraction of phase i. Individual phase momentum equation (IPME): ∂ρi ui ∂t + ∇ · (ρi ui ⊗ ui + pi1) = 0,

Suhas S Jain Stability of bubbly flows 3 / 23

slide-4
SLIDE 4

Introduction and formulation Stability analysis

Derivation of mixture continuity equation

Starting with IPCE for liquid phase: ∂ρlαl ∂t + ∇ · (ρl ulαl) = 0, Rewrite, 1 ρl Dρl Dt + 1 αl Dαl Dt + ∇ · ul = 0, If p = f(ρ, s), for an isentropic process, Dp Dt = c2 Dρ Dt where, c is the speed of sound. Using this above, 1 ρlc2

l

Dpl Dt + 1 αl Dαl Dt + ∇ · ul = 0,

Suhas S Jain Stability of bubbly flows 4 / 23

slide-5
SLIDE 5

Introduction and formulation Stability analysis

Introducing terminologies

Let, β - number of bubbles per unit liquid volume n - number of bubbles per unit total volume τ - individual bubble volume αb - volume fraction of bubbles = nτ Now, 1 + βτ = 1 +

  • #
  • liq. vol.
  • τ = 1 + gas vol.
  • liq. vol. = tot. vol.
  • liq. vol.

n = #

  • tot. vol. =

#

  • liq. vol. ∗ liq. vol
  • tot. vol =

β 1 + βτ αb = βτ 1 + βτ Substituting in the equation, Mixture continuity equation:

  • 1

1 + βτ Dβτ Dt − 1 ρlc2

l

Dpl Dt = ∇ · ul, where, D/Dt = ∂/∂t + ul · ∇ and τ = 4/3πR3 (assuming spherical bubbles)

Suhas S Jain Stability of bubbly flows 5 / 23

slide-6
SLIDE 6

Introduction and formulation Stability analysis

Mixture momentum equation

Start with IPME for liquid phase: ∂ρl ul ∂t + ∇ · (ρl ul ⊗ ul + pl1) = 0, Rewriting,

  • ul

∂ρlαl ∂t + ∇ · (ρlαl ul)

  • + ρlαl

∂ ul ∂t + ul · ∇ ul

  • = −

∇pl, Mixture momentum equation: ρl(1 − αb) D ul Dt = − ∇pl

Suhas S Jain Stability of bubbly flows 6 / 23

slide-7
SLIDE 7

Introduction and formulation Stability analysis

Closure

Assuming volumetric mode of oscillation of the bubbles, modified Rayleigh-Plesset equation (also called as Keller-Miksis equation)

  • 1 − 1

cl ˙ R

  • R ¨

R + 3 2 ˙ R2 1 − 1 3cl ˙ R

  • =
  • 1 + 1

cl ˙ R pR(t) + pl(t + R/cl) ρl

  • + R

ρlcl ˙ pR(t) where, dots are D/Dt, pR is the liquid pressure at bubble surface and pl is the driving pressure. Boundary condition pb(t) = pR(t) + 2 σ R + 4µ ˙ R R where, pb is the uniform bubble internal pressure, σ is the surface tension and µ is the liquid viscosity. [Keller & Miksis, J. Acoust. Soc. Am., 1980.]

Suhas S Jain Stability of bubbly flows 7 / 23

slide-8
SLIDE 8

Introduction and formulation Stability analysis

Final system

Mixture continuity equation:

  • 1

1 + βτ Dβτ Dt − 1 ρlc2

l

Dpl Dt = ∇ · ul, Mixture momentum equation: ρl(1 − αb) D ul Dt = − ∇pl modified Rayleigh-Plesset equation (also called as Keller-Miksis equation):

  • 1 − 1

cl ˙ R

  • R ¨

R + 3 2 ˙ R2 1 − 1 3cl ˙ R

  • =
  • 1 + 1

cl ˙ R pR(t) + pl(t + R/cl) ρl

  • + R

ρlcl ˙ pR(t) Boundary condition: pb(t) = pR(t) + 2 σ R + 4µ ˙ R R

Suhas S Jain Stability of bubbly flows 8 / 23

slide-9
SLIDE 9

Introduction and formulation Stability analysis

Stability analysis of 2D parallel flows

Let, ul = U(y)ˆ ex + ˜ u(x, y, t) and vl = ˜ v(x, y, t) pl = p0 + ˜ p(x, y, t) Rl = R0 + ˜ R(x, y, t) Let, αb → α, ρl → ρ, cl → c, Substituting these in mass and momentum equations, linearizing and subtracting base flow and (assuming β to be uniform), disturbance mass equation 3α R0 ˆ D ˜ R ˆ Dt − 1 ρc2 ˆ D˜ p ˆ Dt = ∇ · ˜ u, where, ˆ D/ ˆ Dt = ∂/∂t + U∂/∂x disturbance momentum equation ρ(1 − α) ∂˜ u ∂t + U ∂˜ u ∂x + U′˜ v

  • = − ∂ ˜

p ∂x ρ(1 − α) ∂˜ v ∂t + U ∂˜ v ∂x

  • = − ∂ ˜

p ∂y where, prime denotes ∂/∂y

Suhas S Jain Stability of bubbly flows 9 / 23

slide-10
SLIDE 10

Introduction and formulation Stability analysis

Continued

If gas is assumed to behave polytropically, then, pb = pb0 R0 R 3γ Linearizing, pb = pb0

  • 1 − 3γ R0

R

  • Substituting these in Keller-Miksis equation and the boundary condition,

linearizing and subtracting base flow, disturbance equation for bubble dynamic response ρ ¨ ˜ R + pb0 3γ ˜ R R2 − 2σ ˜ R R3 + 4µ ˙ ˜ R R2 + pb0 3γ ˙ ˜ R cR0 − 2σ ˙ ˜ R cR2 + 4µ ¨ ˜ R R0c = − ˜ p R0 where, dots represent ˆ D/ ˆ Dt

Suhas S Jain Stability of bubbly flows 10 / 23

slide-11
SLIDE 11

Introduction and formulation Stability analysis

Normal mode assumption

Now, making an ansatz for the disturbance, ˜ u = ˆ u(y)ei(kx−ωt) ˜ v = ˆ v(y)ei(kx−ωt) ˜ p = ˆ p(y)ei(kx−ωt) ˜ R = ˆ R(y)ei(kx−ωt) Substituting these in the disturbance equations, disturbance mass equation ikˆ u + ˆ v′ = −i 3γ R0 ωL ˆ R + i ωL ρc2 ˆ p disturbance momentum equation ρ(1 − α)(−iωLˆ u + U′ˆ v) = −ikˆ p ρ(1 − α)(iωLˆ v) = ˆ p′ where, ωL = ω − Uk is the Lagrangian frequency.

Suhas S Jain Stability of bubbly flows 11 / 23

slide-12
SLIDE 12

Introduction and formulation Stability analysis

Continued

disturbance equation for bubble dynamic response

ω2

L

  • inertial

− iωLλ

damping

+ ω2

b

  • compressiblity
  • ˆ

R = −

  • 1 + i ωLR0

c

  • ˆ

p ρR0 where, λ = ω2

LR0

c

acoustical

− 4µ ρR2

viscous

+

  • thermal = 0
  • is the damping coefficient and

ω2

b = pb03γ

ρR2 − 2σ R3 is the natural frequency of the bubble

Suhas S Jain Stability of bubbly flows 12 / 23

slide-13
SLIDE 13

Introduction and formulation Stability analysis

Dispersion relation for homogeneous medium

Let both x and y be homogeneous. Repeating the whole process again, Then eliminating ˆ u, ˆ v and ˆ R from the 4 disturbance equations ⇒ wave equation for ˆ p, 3α R2 (1 − α)(1 + iω R0

c )

(−ω2 − iωλ + ω2

b) + (1 − α)

c2

  • ω2 −
  • k2

x + k2 y

  • =0 ⇒ dispersion relation

ˆ p = 0 speed of propagation of harmonic disturbance ω in the bubbly mixture medium 1 c2

m(ω) = 3α

R2 (1 − α)(1 + iω R0

c )

(−ω2 − iωλ + ω2

b) + (1 − α)

c2

Suhas S Jain Stability of bubbly flows 13 / 23

slide-14
SLIDE 14

Introduction and formulation Stability analysis

Going back to the parallel flow setup

Eliminating ˆ R and ˆ p from the 4 disturbance equations, equivalent Rayleigh system for bubbly flows ˆ u′ = ikˆ v − i U′′ ωL ˆ v − i U′ kc2

m(ωL)

  • iωLˆ

u − U′ˆ v

  • ˆ

v′ = −ikˆ u + ωL kc2

m(ωL)

  • iωLˆ

u − U′ˆ v

  • In the limit of cm → ∞,

ˆ u′ = ikˆ v − i U′′ ωL ˆ v ˆ v′ = −ikˆ u Eliminating ˆ u, it reduces to the classical Rayleigh equation, (U − c)(D2 − k2)ˆ v − U′′ˆ v = 0 where, D is ∂/∂y

Suhas S Jain Stability of bubbly flows 14 / 23

slide-15
SLIDE 15

Introduction and formulation Stability analysis

Base flow and boundary conditions

Base flow: Inviscid shear layer U(y) = U1 + U2 2 + U2 − U1 2 tanh( y δ ) Boundary conditions: When y → ±∞ ⇒ U = const, then the system reduces to, ˆ u′ = ikˆ v ˆ v′ = −ikˆ u + i ω2

L

kc2

m(ωL) ˆ

u This admits a close form solution as, Asymptotic solutions ˆ v = Ae±y(k2−ω2

L/c2 m)1/2

ˆ u = ±A ik (k2 − ω2

L/c2 m)1/2 e±y(k2−ω2

L/c2 m)1/2

where A is an arbitrary complex constant.

Suhas S Jain Stability of bubbly flows 15 / 23

slide-16
SLIDE 16

Introduction and formulation Stability analysis

Solution procedure

Method: Shooting method. Spacewise problem: complex k and real ω is assumed. Procedure Guess a complex eigenvalue k. Choose A such that initial conditions at y = −nδ simplifies, where n ≫ 1 (n = 5 in this project) ˆ v = 1 ˆ u = ik (k2 − ω2

L/c2 m)1/2

Integrate upto y = nδ (using RK4 in this project). Check if the solution is continuous with the asymptotic solution at y = nδ ˆ u = − ik (k2 − ω2

L/c2 m)1/2 ˆ

v Iteratively correct eigenvalue k until convergence (using 2D Newton-Raphson method in this project)

Suhas S Jain Stability of bubbly flows 16 / 23

slide-17
SLIDE 17

Introduction and formulation Stability analysis

Verification of the solver

d’Agostino et al. (1997) verified their solver against Michalke, JFM, (1965) results for cm → ∞. same reference used here.

0.1 0.2 0.3 0.4 0.5 0.05 0.1 0.15 0.2 0.25

Figure 1: Verification against values from Table 1. of Michalke, JFM, (1965). In the present project, ω∗ values from 0 to 0.5 has been used with a step size of 0.005. Solution is computed from lower ω∗ to higher. −k∗

i values from previous ω∗ is used as

an initial guess for next ω∗.

Suhas S Jain Stability of bubbly flows 17 / 23

slide-18
SLIDE 18

Introduction and formulation Stability analysis

Effect of presence of bubbles

0.1 0.2 0.3 0.4 0.5

  • 0.05

0.05 0.1 0.15 0.2 0.25

Figure 2: Presence of bubbles have a stabilizing effect on the flow. Left: current work, Right: d’Agostino et al., JFM, (1997). In all cases α = 0.01 and R0 = 0.01.

as ω∗

b0 decreases (approaches towards excitation frequency ω∗), flow stabilizes.

ω∗

b0 ≫ ω∗ ⇒ fluid behaves barotropically → asymptotes to single-phase

behavior.

Suhas S Jain Stability of bubbly flows 18 / 23

slide-19
SLIDE 19

Introduction and formulation Stability analysis

Maximum amplification rate

5 10 15 20 25 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22

Figure 3: Maximum amplification rate −k∗

MAX as a function of natural frequency of

the bubbles ω∗

b0 for different values of void fraction α. Left: current work, Right:

d’Agostino et al., JFM, (1997). In all cases R0 = 0.01.

as ω∗

b0 decreases, −k∗ MAX reduces and flow stabilizes as observed before.

ω∗

b0 also decreases as α increases hence stabilizing the flow for higher void

fractions.

Suhas S Jain Stability of bubbly flows 19 / 23

slide-20
SLIDE 20

Introduction and formulation Stability analysis

Effect of bubble resonance

Numbers for ω∗

b0 in previous two cases were picked that are relevant to

practical applications.

0.1 0.2 0.3 0.4 0.5

  • 0.05

0.05 0.1 0.15 0.2 0.25

Figure 4: Left: current work, Right: d’Agostino et al., JFM, (1997). In all cases α = 0.003.

doesn’t match the paper exactly, since R∗

0 values are missing in the paper,

but the results are close (I use R∗

0 = 0.32).

as ω∗

b0 decreases, flow again stabilizes.

at resonance ω∗

b0 ≈ ω∗, flow is more stable (a local minimum can be seen).

Suhas S Jain Stability of bubbly flows 20 / 23

slide-21
SLIDE 21

Introduction and formulation Stability analysis

Reasons for stabilizing effect of bubbles

Compressibility effect: “a certain amount basic flow energy must be used to do work against the force due to the elasticity of the medium, before it becomes available to initial instability” (Blumen et al. (1975)). Bubble dynamic damping: this provides another source of energy absorption, which at resonance is significant due to the large amplitude of bubble response.

Suhas S Jain Stability of bubbly flows 21 / 23

slide-22
SLIDE 22

Introduction and formulation Stability analysis

Conclusion

summary Governing equations of bubble-liquid mixtures along with the closure were formally derived. Disturbance relations were derived making appropriate assumptions. Inviscid stability analysis of a bubbly inviscid shear flow was performed. Stabilizing effect of presence of bubbles was studied. Stability as a function of natural frequency of bubbles and void fraction was also studied. Stabilizing effect of bubbles at resonance was also investigated. Work of d’Agostino et al., JFM, (1997) was successfully reproduced.

Suhas S Jain Stability of bubbly flows 22 / 23

slide-23
SLIDE 23

Introduction and formulation Stability analysis

THANK YOU

References: d’Agostino, L; d’Auria, F & Brennen, C (1997), ’On the inviscid stability of parallel bubbly flows’, Journal of Fluid Mechanics, Vol. 339, pp. 261-274. d’Agostino, L. & Brennen, C. E. (1989), ‘Linearized dynamics of spherical bubble clouds’,Journal of Fluid Mechanics, 199, 155–176. d’Auria, F., d’Agostino, L. & Brennen, C. E. (1995), ‘Inviscid stability of bubbly jets’,AIAA. Blumen, W., Drazin, P. G. & Billings, D. F. (1975), ’Shear layer instability of an inviscid compressible fluid. Part 2.’,Journal of Fluid Mechanics, 71, 305–316. Keller, B, J & Miksis, M, (1980), ’Bubble oscillations of large amplitude’, The Journal of the Acoustical Society of America, 68, 628. Michalke, A. (1965), ’On spatially growing disturbances in an inviscid shear layer’, Journal of Fluid Mechanics, 23, 521–544.

Suhas S Jain Stability of bubbly flows 23 / 23