• Nie Znaleziono Wyników

Unsteady particle accumulation in wall turbulence

N/A
N/A
Protected

Academic year: 2021

Share "Unsteady particle accumulation in wall turbulence"

Copied!
2
0
0

Pełen tekst

(1)

15TH EUROPEAN TURBULENCE CONFERENCE, 25-28 AUGUST, DELFT,. THE NETHERLANDS

UNSTEADY PARTICLE ACCUMULATION IN WALL TURBULENCE

Dmitrii Ph. Sikovsky

1,2

1

Institute of Thermophysics of Siberian Branch of Russian Academy of Science, Novosibirsk, Russian

Federation

2

Novosibirsk State University, Novosibirsk, Russian Federation

Abstract We propose the asymptotic theory of unsteady accumulation of inertial particles in the viscous sublayer of wall-bounded particle-laden turbulent flow. We derive the diffusion equation for the particle concentration in the viscous sublayer and find the self-similar exact solution of this equation at large times. It is shown that near the wall the maximal concentration grows as t1/2, while the distance from the wall to the concentration pike as well as its width decay as t-1/2. The obtained solution is corroborated by the results of stochastic Lagrangian simulations.

In [1] it was shown that the steady-state solution for the kinetic equation for the probability density function of particle position and velocity (PDF) exhibits the near-wall singularity. It was assumed that the infinite time is needed for the point particles to reach such steady solution with the singularity of particle concentration. The aim of the present paper is to study the unsteady accumulation of particles in the viscous sublayer of wall-bounded turbulent flow.

In the thin viscous sublayer the intensity of wall-normal fluid velocity decays to zero as vf¢ =qy2 +O y( )3 , where y is the distance to the wall and wall units are implicit throughout for all variables. Above the viscous sublayer, in the buffer zone the particles interact with the near-wall coherent structures such as the low-speed streaks and energetic streamwise vortices which radius is of order of viscous sublayer thickness. Based on the observations from DNS/Largangian tracking numerical data [2] and the results of asymptotic analysis [1], we can divide the particles on two groups. Particles of the first group are captured by the near-wall energetic eddies and slung toward the wall with the wall-normal velocities of order of one, which are much larger than local f . Thus, they are not influenced by the fluid velocity fluctuations and can be referred to as ballistic, or free-flight particles. Particles of the second group are trapped by the low-speed streamwise streaks and have velocities of the order of f and can be referred to as diffusional particles. In [1] it was shown that the near-wall build-up of particles is created by the slow diffusional particles.

Consider the motion of diffusional particles in the limit y ® 0

2 ( )3

dv v

qy O y

dt +t = + (1)

where t is the particle relaxation time, q =q x t z t t( ( ), ( ), ) x t z t( ), ( ) are the streamwise and spanwise positions of particle.

During the time interval of order of decorrelation time of fluid turbulence near the wall – Lagrangian timescale –TL, the diffusional particle shifts in wall-normal direction on a distance Dy ~qy T2 L, which is much less than its initial position y t( ), the shifts in stream- and spanwise directions are small as well Dx z, ~y. Thus we can assume the particle wall-normal position y t( ) in (1) to be a slow variable as compared to a fast variable q x t z t t( ( ), ( ), ), in which in turn we can assume the particle position to be almost fixed: q x t z t t( ( ), ( ), )»q t( ). Then from (1) it follows the Kramers-Moyal expansion for particle PDF P v y t( , , )= ád(v-v t( )) (dy-y t( ))ñ[1]

2 2 1 1 4 5 0 2 1 ( ) ( ) ( ) vP vP P P P y O y t y t v t l v l v y - - æ ö ¶ ¶ ¶ + - = çç+ ¶ ÷÷+ ÷ ç ÷÷ ç ¶ ¶ ¶ è ¶ ¶ ¶ ø (2) where 0 1 0 (0) ( ) exp( / ) q q t t dt l t t ¥ -=

ò

á ñ - , 1 0 (0) ( ) [1 exp( / )] q q t t dt l t ¥ =

ò

á ñ - - .

Multiplying (2) on vn and integrating over all v gives the transport equations for the particle concentration Pdv

F =

ò

, flux j =

ò

vPdv, and energy á ñF =v2

ò

v Pdv2 0 j t y ¶F ¶ + = ¶ ¶ , 4 2 1 ( ) 1 v y j t y y l t t æ¶ ö÷ ¶ á ñF ¶F ç + ÷ + = ÷ ç ÷ è¶ ø ¶ ¶ (3)

(2)

4 4 3 2 ( ) 0 1 2 v v 2 y 2 y j t y y l l t t t æ¶ ö÷ ¶ á ñF ¶ ç + ÷á ñF + = ÷ ç ÷ è¶ ø ¶ ¶ (4)

At large times t >> t, the quasi-equilibrium approximation [1,3] of (3),(4) gives the following diffusion equation 4 Dy t y y y a é æ öù ¶F ¶ ê ç¶F F ÷ú = ê çç + ÷÷÷ú è ø ¶ ¶ ëû, (5) where 2 0 (0) ( ) L D q q t dt q T ¥ =

ò

á ñ = á ñ , 0 0 1 4 4 L T tl t a tl l t = » + + (cf. [1]).

We found the exact self-similar solution of (5) for the case of inertial particles with t >TL / 3 2 Dt F y Dt(2 ) F = , F( )h =Ch-aexp(-h-2) (6) where 1 0 1 2 2 C a dy ¥ - æç - ÷ö = G çç ÷ F÷÷ è ø

ò

.

The solution (6) describes the unsteady accumulation of inertial particles near the wall at large times, when all particles becomes diffusional. The latter is the consequence of the irreversibility of the transition from free-flight particles to diffusional one expressed by (5), since it is almost impossible for the former free-flight particle retarded near the wall to increase their velocity up to order of one values again. Particle concentration profile has the near-wall pike with magnitude growing as t1/2 and the thickness collapsing as t-1/2. According to (6) after the infinite time all particles will deposit the wall, so that F~ ( )dy at t = ¥. The real particles have the finite radius r , such that the solution (6) is valid at times t <<D r-1 2. Beyond these times t>>D r-1 2 the regularized steady-state singularity solution [1] is reached. At large distances y >>( )Dt-1/2 the particle concentration obeys the power law F~ y-a, which is similar to the steady solution obtained in [1].

We perform the stochastic Lagrangian simulations using the model [3] of turbulent channel flow at Ret =180 seeded with the point particles. The results of simulation corroborate the expression (6) (Fig.1). As seen from Fig.1, at the initial stage of accumulation the maximal concentration grows linearly with time, which can be explained with the help of another self-similar solution of (5) having the form F ~t F y Dt1(2 ).

References

[1] D.Ph.Sikovsky. Singularity of inertial particle concentration in the viscous sublayer of wall-bounded turbulent flows. Flow, Turb. and Comb.

92: 41-64 ,2014.

[2] A. Soldati, C. Marchioli. Physics and modelling of turbulent particle deposition and entrainment: Review of a systematic study. Int. J. Multiphase Flow 35: 827-839, 2009.

[3] D.Ph.Sikovsky. Stochastic Lagrangian simulation of particle deposition in a turbulent channel flows. 10th International ERCOFTAC Symposium on Engineering Turbulence Modeling and Measurements, Marbella, Spain, 17-19 September 2014. DOI: 10.13140/2.1.4204.6722.

Cytaty

Powiązane dokumenty

Nowe, zmienione zasady wprowadzania do obrotu lub udo- stępniania na rynku krajowym wyrobów budowlanych nieobję- tych rozporządzeniem Parlamentu Europejskiego i Rady (UE) nr

The differences in the seismic image are generally discernible within all the structural stages present in the geological profile of the analysed region but are most emphasized in

The main criterion for the occurrence of bacterial or thermo- chemical sulfate reduction processes is the presence of sulfates and organic matter in rocks, in which these

189 of the Code of Civil Procedure (CCP), a limited liability company (LLC) [spółka z o.o.] has a legal interest to determine invalidity of the statement made by Supervisory

This article, introducing the legal regulation related to the introduced time limits, aims to draw attention to interpretation doubts regarding the issue of linking the time limit

Na podstawie zgromadzonej przez Radę Akredytacyjną dokumen- tacji, organ ten dokonuje oceny punktowej spełnienia poszczególnych standardów przez podmiot ubiegający

Dziecko w rodzinie dawniej i dziś, zorganizowana przez Katedrę Historii Ustroju i Prawa Wydziału Prawa, Prawa Kanonicznego i Administracji KUL, Zespół Szkół Ekonomicznych

The model of multiple criteria decision making of equipment maintenance in Haditha hydropower station can generally be divided into three levels: goals and