Engineering
and Architecture /5. Heat and Gas Supply and Ventilation
Vladimir I. Sokolov, Alim A. Kovalenko
Volodymyr Dahl East-Ukrainian
National University, Ukraine,
Inna K. Nasonkina
Donbass National Academy of Civil
Engineering and Architecture, Ukraine
The Stages
of Aerosols’ Diffusion in Turbulent Flows
The
estimation of diffusion process characteristics of aerosols in the ventilation
systems of industrial enterprises requires the research of discrete particles’
behavior in a turbulent flow. Such behavior depends on the concentration of
these particles and their size, on comparison to the scale of turbulence
working media[1-3]. This is a pressing issue at
measuring of aerosols’ concentration in the exhaust of nuclear power plants.
When
concentration is high the collision of particles and the influence of media in
the vicinity of particles cause the direct interaction of particles. In case
when concentration of particle is not high the interaction between the
particles can be neglected and each particle is considered as it is the only in
the turbulent flow. This is very common in practice.
There
are the logical assumptions for gas and aerosol exhaust in the channels of vent
systems:
· turbulence in a flow is
homogeneous and stationary;
· particle is spherical and it
is so small that it motion relatively of media submits to the Stokes law
resistance;
· particle is very small in
comparison to the smallest length of the wave in turbulent motion;
· any external force which
influences on a particle is related to a potential field, for example gravity.
Recognizing
these assumptions the behavior of spherical particle in a flow can be described
with equation of Basse [4]:
, (1)
where
; (2)
t - time; t0 - an initial moment of
time; index f - related to basic media, and index p - to a particle; m - a dynamic viscosity of
media; d - a diameter of particle; rp and rf are densities of
particles and gas stream; vp
- a speed of a discrete particle; vf
- a speed of particles of basic media in vicinity of discrete particle far
enough not to test indignations of relative motion of this particle.
The
coefficients of temporal Lagrange correlation for the interval of time t are used as
characteristics of turbulent motion
, (3)
and the Lagrange
functions of power spectrum are
,.
(4)
Here are the pulsation components of
speeds of discrete particle and gas flow and an index <…> designates that
pulsating value in turbulent flow becomes a mean value.
It be mentioned that coefficients of temporal
Lagrange correlation is related to the Lagrange functions of power spectrum
with the help of relation
, (5)
We must
add that the coefficient of the Lagrange correlation of basic flow is usually
presented as close to exponential dependence for implementation of quantitative
estimations [4]
, (6)
where is the Lagrange integral temporary scale taken as a measure of the
longest time interval during which a particle moves in a given direction .
Linking
the Lagrange variables we write down the relations for the Lagrange coordinates
with the help of the function of power spectrum
,. (7)
As the coefficient of turbulent diffusion is , then we obtain the following for a discrete particle and gas
. (8)
Under small time of
diffusion
, ,
from where there is
. (9)
Under
long time of diffusive process a main role belongs to the low frequency
components of motion
(10)
and
.
(11)
According to expressions (10) and (11) the
coefficient of diffusion is proportional to the part of kinetic energy in
turbulent motion with a zero frequency. But there is no difference between the
motion of a particle and the motion of media under condition of zero frequency.
Thus, it is very logical from physical point of view that coefficients of
diffusion for a discrete particle and particles of basic stream must be the
same. Therefore
. (12)
We must
underline that those notions of “small” and “long” time of diffusion are very
relative and must be defined in every particular case. The expresses (12) is
for infinitely large time of diffusion. As the channels of real systems have
boundaries, in most of cases there will be the difference in values of
coefficients of turbulent diffusion.
We estimate
the duration of induction period where the Stokes forces cause an acceleration
of particles to speed of basic flow. For this purpose in equation (1) we will
ignore the last term. Such assumption can be taken as fully legitimate because
forces of resistance and forces related to the gradient of pressure and to the
relative acceleration of mass become vital. Then we obtain
. (13)
In this
case the constant coefficients are defined according to equation (2). As for aerosol exhaust in air flow the value has an order 10-3 and
we can establish that b»0. That is why the
equation (13) can be presented as
(14)
We
integrate it under initial condition that for gas flow speed. There is a solution (14)
,
from where we
obtain
. (15)
And
.
Thus, recognizing (15) from expression (9) we obtain
. (16)
According to (16) we estimate the time of induction period. If the end
of induction period is counted by the value , taken in most of technical calculations, then the time of period
is . Taking to consideration the expression (2) we have for coefficient a
and taking into account that has an order 10-3 we
establish the following relation for the time of induction period
. (17)
After the induction period is over the power spectrums of particles and
of basic flow are not the same therefore the diffusive coefficients are
different. Thus, the diffusive period under variable coefficient of diffusion
of aerosol particle will take place. Taking the coefficient of the Lagrange
correlation for gas flow as exponential dependence (6) we obtain the
corresponding power spectrum in dependence on frequency of pulsations w.
.
Considering equation (13) the power spectrum for discrete particles is
.
Then
and according to
(5)
By the
coefficient of the Lagrange correlation we define the coefficient of diffusion
((3),(4) and (8)).
As
there is
(18)
then we obtain
.
As
mentioned above << 1 then we are supposed b » 0 and . From here we obtain
. (19)
Let us
estimate the Lagrange integral scale of time. Based on (18) where t®¥
,
and taking to
consideration that degree of turbulence is , we obtain
. (20)
where u0 is an average
speed of flow in a channel.
The
following transformations are done
,
where is the Reynolds number; is kinematics viscosity of basic flow, dã - a hydraulic diameter of
channel; is a scale of time equal to the time of passage of a basic flow particle
through the way which is equal to hydraulic diameter of channel.
Examining the
Peclet number
,
we obtain
. (21)
The calculations
have shown that for a degree of turbulence e = 0,05 in diapason of the Peclet diffusion numbers 300...400 considered
for typical of ventilation the Lagrange value of integral scale is . Thus, it approaches the time -of passing of a basic flow particle
through the way equal to hydraulic diameter of channel.
It
substitute (21) to (20) and put into operation dimensionless time .
Considering that ; è ,
. (22)
As it is seen when it is completely correspond to
the result (12).
According to (22)
maximum dimensionless time of the transitional period in a diffusive process
under condition that aerosol particle’s coefficient of diffusion approaches the
basic flow coefficient of diffusion is
. (23)
This relation shows
that time of transitional period depends on different factors and it must be
estimated in every particular case.
Thus, the process
of aerosols’ diffusion in turbulent flow is very compeck and can be divided
into 2 initial parts: the induction period, when the Stokes forces cause an
acceleration of particles to speed of basic flow, and the transitional period,
when the process of diffusion takes place under variable coefficient of
diffusion. The diffusive process in the induction period is almost absent and
the duration of this period is determined by equation (17). Most of transitional period is set by relation (23)
and the coefficient of aerosols’ diffusion is estimated by relation (22).
Literature
1.
Rutherford A. Introduction
to the Analysis of Chemical
Reactors. Departments of
Chemical Engineering University of Minnesota. Prentice - Hall. Inc., New
Jersey, 1967. - 340 p.
2.
Ñîêîëîâ Â.È.
Àýðîäèíàìèêà ãàçîâûõ ïîòîêîâ â êàíàëàõ ñëîæíûõ âåíòèëÿöèîííûõ ñèñòåì/ Â.È.
Ñîêîëîâ. – Ëóãàíñê: ÂÓÃÓ, 1999. – 200 ñ.
3.
Õèíöå È.Î.
Òóðáóëåíòíîñòü/ È.Î.Õèíöå. - Ì.: Ôèçìàòãèç, 1963. - 680 ñ.
4.
Êàëþæíûé Ã.Ñ. Äèôôóçèÿ ãàçîâ è àýðîçîëåé â
òóðáóëåíòíûõ ïîòîêàõ/ Êàëþæíûé Ã.Ñ.,
Êîâàëåíêî À.À., Ñîêîëîâ Â.È., Ìèíèí Ñ.À. – Ëóãàíñê, 1999. – 100 ñ.