Beruflich Dokumente
Kultur Dokumente
Jet with variable fluid properties: Free jet and dissipative jet
Manisha Patel a,n, M.G. Timol b
a
Department of Mathematics, Sarvajanik College of Engineering & Technology, Surat 395001, Gujarat, India
b
Department of Mathematics, Veer Narmad South Gujarat University, Magdalla Road, Surat 395007, Gujarat, India
art ic l e i nf o a b s t r a c t
Article history: An analysis of the laminar jet of an incompressible Newtonian fluid emerging from a narrow slot or a
Received 13 February 2015 circular hole, where the physical properties like viscosity and thermal conductivity depends upon the
Received in revised form temperature, is given. Both the cases: the case of In the absence of viscous heat dissipation and the case
4 June 2016
of In the presence of viscous heat dissipation are considered. The governing partial differential equations
Accepted 4 June 2016
of the flow problem are transformed into the ordinary differential equations by group theoretic tech-
Available online 7 June 2016
nique. The Runge–Kutta method is applied to obtained numerical solution of the transformed ordinary
Keywords: differential equations.
Jet flow & 2016 Elsevier Ltd. All rights reserved.
Thermal conductivity
Heat dissipation
Newtonian fluid
Non-homogeneous BVP
Runge–Kutta method
1. Introduction fluid at rest is a classical problem in jet flows. This has been dis-
cussed in most of the books on the fluid mechanics. In the study of
When a fluid moving in a pipe or tank crosses a slot or a nozzle jets one has to face some inherent difficulties; (i) satisfying
there is a sudden decrease in cross section area and consequently homogeneous boundary condition at infinity, (ii) physical condi-
there is a considerable increase in velocity. It gives rise to the flow tion of constancy of flux of momentum or of heat or both.
of a jet. A fluid flows in the absence of rigid boundaries and, The flow of a jet is divided into two main regions [1–3], the core
therefore, it is a free flow. These flows are important in technology, and mixing region as shown in the Fig. 1. The core region is si-
metrology etc. tuated at the center of the jet and near the exit nozzle. In this
Fluid properties like viscosity, thermal conductivity cannot in region, viscous effects are negligible and the fluid may be con-
general, be regarded as constant. In many physical situations such sidered to be inviscid. The mixing region begins at the edge of the
as liquid metal injectors and highly heated arcs etc. there is a large core. In this zone, there is a large variation in velocity and density
temperature difference between the temperature of jet and that of and the effect of viscosity and heat conduction must be
surrounding. Hence it becomes necessary to consider the variation considered.
of viscosity and thermal conductivity etc. with the temperature in The first study was recorded in literature by Schlichting [4]
case of jets as well. Density of a liquid changes with the tem- when he obtained closed form solution of laminar jet. Bickly [5],
perature but the rate of decrease of density with increase of Squire [6] have discussed laminar jets of an incompressible fluid
temperature is much smaller than that of viscosity and thermal with different geometries of the orifice, Pai [7] studied two-di-
conductivity. Therefore in all the analysis which follows density mensional and axisymmetric jet mixing of a compressible fluid.
will be taken as constant while viscosity and thermal conductivity Kapur [8], Gutfinger et al. [9] introduced jets of a non-Newtonian
will be assumed to be temperature dependent. However, density fluid. In most of the above study it is assumed that the physical
variations when coupled with forces due to gravity will give rise to properties are constant with respect to temperature but it is valid
buoyancy forces. It will therefore, be desirable to take into account only when the temperature differences are very small. The study
such forces in vertical jet. of jet with variable fluid properties started in the year 1950. Os-
A jet issuing from an orifice and mixing with a surrounding trach [10] studied it in the case of flat plate problem. Extensive
work in this area is carried out by many researchers. Among those
n it is worthy to quote here Spalding's et al. [11], Christiansen et al.
Corresponding author.
E-mail addresses: manishapramitpatel@gmail.com (M. Patel), [12] in flow through channels and tubes. Pai [7] suggested em-
mgtimol@gmail.com (M.G. Timol). pirical relations for incompressible Newtonian and Non-
http://dx.doi.org/10.1016/j.ijnonlinmec.2016.06.001
0020-7462/& 2016 Elsevier Ltd. All rights reserved.
M. Patel, M.G. Timol / International Journal of Non-Linear Mechanics 85 (2016) 54–61 55
Newtonian fluid flow. A jet with variable fluid properties was flow structures and distribution of the Nusselt number, and
simply analyzed by Kalathia [3]. Here, we shall solve it completely pressure coefficient along the flat plate. Lee et al. (2012) have
by the method of converting homogeneous Boundary value Pro- experimentally measured the Nusselt number distribution over
blem (BVP) into non-homogeneous initial value problem. The de- isothermal flat plate. This study focused on describing unsteady
tailed analysis of the problem and method of solution is discussed modes for each height ratio and Re number. Rady and Arquis [16]
in this paper. have numerically compared local and stagnation Nusselt number
Jets are used in various industrial applications because of pro- of a multiple impinging slot jets over an isothermal flat plate in
ducing high heat and mass transfer coefficients. The flow struc- laminar regime. They analyzed the effect of exhaust ports and
tures of unconfined impinging jets are divided to three main re- spent flow on heat transfer rates, and they suggested that locating
gions: free jet region, stagnation region and wall jet region. The surface protrusions before exhaust ports can be enhanced the
maximum Nusselt number is observed at the stagnation point or Nusselt number and reduce adverse effect of spent flow in mul-
the surrounding point due to the formation of thin thermal tiple impinging jet systems. Kubacki and Dick [17] have numeri-
boundary layer in that region. Confined impinging jets have these cally investigated the turbulent confined plane impinging jets
three global regions, but the upper wall that is in the same level using different turbulent models.
with nozzle's exit restricts entrainment of jets, so the length of Kayansayan and Kucuka [20] both experimentally and nu-
potential core in confined jets is more than free jets; therefore the merically have investigated the flow structure and parameters on
Nusselt number has experienced greater amount related to free jet concave impinging plate. They analyzed turbulent and laminar
in the same condition for laminar impinging jets. Several in- regimes for confined plane impinging jets that are issued from the
vestigations have been done on the impinging jets for many years symmetry line of semi cylindrical channel. Measurement of pres-
that have been discussed the important factors on such flow sure and Nusselt number distribution over impinging plate has
structures. Chiriac and Ortega [13] has numerically studied the been reported. They simulated the laminar flow by a numerical
heat transfer, flow structures and transitional behaviors of con- model too. It has been indicated the amount of the Nusselt num-
fined slot jet at a fixed nozzle to surface distance and different ber over target plate has been enhanced compared to flat plate due
Reynolds numbers. Park et al. [14] have numerically studied two to surface curvature. Rahman and Hernandez [21] have numeri-
dimensional laminar and turbulent impinging slot jets using cally studied the conjugate heat transfer between a round water
Simple-base segregated streamline upwind Petrov- Galerkin finite jet and convex surface. In this analysis, a constant heat flux is
element method. It has been indicated that this method is more imposed on the bottom edge of the surface, and consideration of
accurate than other upwind numerical methods that used the ar- different solid material and plate thicknesses for the convex sur-
tificial diffusion. Lee et al. [15] has numerically investigated the face have been given generality to this investigation. Choi et al.
[22] have experimentally measured the distribution of mean ve- Kalathia et al. [3] are given by:
locity and fluctuating velocity over a concave surface for turbulent
∂u̇ ∂v ̇
impinging slot jet. They interpreted the measured local and stag- + =0
∂ẋ ∂ẏ (1)
nation Nusselt number over the concave plate by turbulent char-
acteristics of flow around stagnation region and wall jet region. ⎡ ∂u̇ ∂u̇ ⎤ ∂ ⎡ ∂u̇ ⎤
ρ ⎢ u̇ + v̇ ⎥ = ⎢μ ⎥
Yang et al. [23] have numerically investigated the Nusselt number ⎣ ∂x ̇ ∂ẏ ⎦ ∂ẏ ⎣ ∂ẏ ⎦ (2)
distribution and velocity field over concave surface for turbulent
⎡ ∂θ ̇ ∂θ ̇ ⎤ ∂ ⎡ ∂θ ̇ ⎤ ⎡ ∂u̇ ⎤2
impinging jet at relatively small curvature. The k–ε model was ρCp ⎢ u̇ + v̇ ⎥ = ⎢K ⎥ + μ⎢ ⎥
used for simulation of turbulent jet. Validation with experimental ⎣ ∂ẋ ∂ẏ ⎦ ∂ẏ ⎣ ∂ẏ ⎦ ⎣ ∂ẏ ⎦ (3)
result has been proved the accuracy of this model. Elcoot [24] The boundary conditions are
carried out the analysis of modulation instability of hydrodynamic
waves on liquid jet in the presence of electric field. Saulo Orizaga ∂u̇ ∂θ ̇
ẏ = 0; v ̇ = = =0
et al. [25] has been discussed the problem of viscous axisymmetric ∂y ̇ ∂ẏ (4)
jets in the presence of uniform and non-uniform electric field. ẏ = ∞; u̇ = θ ̇ = 0 (5)
Desai and Kalthiya [26] have been discussed the similarity so-
lutions of the heated jet of non-Newtonian fluid. The effect of Here u and v are components of velocity in the x and y direc-
variable fluid properties on the flow is also discussed in their pa- tions. ρ is a mass density, θ is the temperature.
per. They have applied systematic group theory approach to obtain For such a flow the momentum flux across any cross-section
similarity solutions. This Brief discussion on the effect of variable perpendicular to the jet axis is constant i.e.
viscosity and thermal conductivity of a nonisothermal, in- ∞
β1 + β2 = 0 (28)
B2 ¯′ 2 ¯ ¯ ′′ C ΓB ∂ ¯ Γ ¯
Solving above equations, one gets Γ = δ for β2 ≠ 0 (otherwise (F + FF ) = − (3 + Γ ) 3 (G F ′′)
trivial solution) and A2 A ∂η (38)
BC ¯′ ¯ ⎡ Γ+ 1 ⎤ ∂ Γ
α1 −α1 2α1 ¯ ¯ ′) = − (3 + Γ ) ⎢ C
(F G + FG ⎥ (G¯ G¯ ′)
β1 = , β2 = , α2 = (Γ ≠ − 3) A σ ⎣ A2 ⎦ ∂η (39)
3+Γ 3+Γ 3+Γ (29)
Subject to the boundary conditions
Therefore we choose proper invariants for assumed transfor-
mation group follows: η = 0; F¯ = F¯ ′′ = 0; G¯′ = 0 (40)
y η = ∞; F¯′ = G¯ = 0 (41)
η= ; ψ = F (η) x1/ (3 + Γ ) , θ = G (η) x−1/ (3 + Γ )
x2 / (3 + Γ ) (30)
From the above equations the conditions of invariance of
Using (30) into set of Eqs. (15)–(18) equations for dissipative transformation should be
free jet will be simplified as (42)
AB = CΓ
2 ∂
F ′ + FF ′′ = − (3 + Γ ) (G Γ F ′′) From the given boundary conditions it is seen that F′ starts from
∂η (31)
a certain constant value at η = 0 and approaches zero as η → ∞. F′
(3 + Γ ) ∂
F ′ G + F G′ = − (G Γ G′) representing velocity cannot have infinite value between
σ ∂η (32) η = 0 and η = ∞ . Hence if ( ∂f /∂η) is integrated from 0 to ∞ a finite
∞
∫0 2
F ′ dη = 1 constant can be obtained,
(33)
∞
∞
∫0 F ′Gdη = 0
(34) ∫0 F′dη = Constant
(43)
(Here dash denotes the differentiation with respect to η).
with the boundary conditions Using Eq. (37) into Eq. (43) we obtain,
η = 0; F = F ′′ = G′ = 0 (35) ∞
η → ∞, F ′ = G = 0 (36)
B ∫0 F ′dη¯ = Constant
P = ⎡⎣ F * (∞) ⎤⎦
(1 − Γ ) /2
It is clear from boundary condition that F (0) = 0. (60)
Hence F (∞) = 1 (45) Value of λ2 is determined from the relation (55) and from these
values the solution of Eqs. (31) and (32) is obtained. Using the
The transformation in Eq. (42) reduce the Eqs. (33) and (34) to
relation (47) constant B and C are calculated from the relations
B2 ∞ ∞ (46) and (42). Thus values of η , F and G are found from the relation
∫0 ¯ ¯ η¯ = 1 and BC
FGd ∫0 F¯ ′ 2dη¯ = 0
A (46) (37).
Hence we have to obtain the solution of Eqs. (31) and (32) with
3.2. Result and discussion
the boundary conditions (40) and (45). Again the boundary con-
dition at the second point is transformed to non-homogeneous
Reduced system of Eqs. (56) and (57) into initial value problem
form. To transform the boundary value problem with boundary
is solved for different value of Γ and σ using the initial conditions
conditions (40) and (45) into initial value problem, we define
(58) and (45). Integration is carried out using the Runge–Kutta
group transformation as follows:
method for first increment and afterwards Adams–Moulton
F = P λ1F * scheme of integration is employed. Step size of integration is taken
η = P λ2 η* to be 0.05. Last value of F̄ gives infinite values of F̄ using this value,
the parameter of transformation P, is obtained from Eq. (57) values
G = P λ3 G* (47)
of λ1, λ2 and P are:
where P is the real parameter of the group transformation and
Γ = − 0.5, σ = 1.0, P = 0.01
λ 1 , λ 2 , λ 3 are real constants. Form Eq. (47), when the transfor-
mations (47) is introduced into Eqs. (38) and (39), we obtain λ 1 = 0.25
∂ λ 2 = − 0.75
F * ′2 + F *F * ′′ = − (3 + Γ ) (G*Γ F * ′′)
∂η* (48) Solution of Eqs. (38) and (39) are obtained by the relation (47)
as
3 +Γ⎡ ∂ ⎤
F * ′G* + F *G* ′ = − ⎢ (G*Γ G* ′) ⎥
δ ⎣ ∂η* ⎦ (49) F = (0.01) 0.25F *
With the initial conditions η = (0.001) −0.75η *
η* = 0; F * = F * ′′ = G* = 0 (50) Once the value of F and η are known the value of integral
obtained using
The condition for above equations to be free from P is,
∞
λ1 + λ2 + Γλ3 = 0 (51) I= ∫0 F ′ 2 ∂η
To integrate Eqs. (48) and (49) two more initial conditions are Is numerically calculated by trapezoidal rule with IBM's scien-
required (two missing initial condition introduced for the first tific subroutine package. The value of I is
time) i.e. the value of F * ′ at η = 0 and the value of G* at η = 0. In
I = 0.0798
order to fine these missing initial conditions it is assumed that,
From these values, solution of Eqs. (31) and (32) is obtained by
F¯ ′(0) = P and G (0) = P (52)
the relation (37).
Under the transformation of Eq. (47), the Eq. (52) will turn to Similarly, for other values of Γ and σ , values of η , F and G can
be obtained.
F * ′ (0) = P ⇒ λ1 − λ2 = 1 (53)
G* (0) = P ⇒ λ3 = 1 (54)
Therefore 4. Case II: dissipative jet (in the presence of viscous heat
dissipation)
λ1 = ( Γ + 1)/2
λ2 = ( 1 − Γ )/2 Group transformations: consider one-parameter liner group
transformations as
λ3 = 1 (55)
x¯ = a α 1x, ψ¯ = a β 1 ψ
Thus finally we have following set of equations
y¯ = a α 2 y , θ¯ = a β 2 θ (61)
∂
F * ′ 2 + F *F * ′′ = − (3 + Γ ) (G*Γ F * ′′)
∂η* (56) where a, α 1, α 2, β 1, β 2 are certain constants.
The dependence of f and ϕ on θ may be assumed, in the first
3 +Γ⎡ ∂ ⎤
instance of the form
F * ′G* + F *G* ′ = − ⎢ (G*Γ G* ′) ⎥
δ ⎣ ∂η* ⎦ (57)
f = θ Γ and ϕ = θ δ (62)
Subject to conditions
Γ and δ being some constants under the above group (61).
η* = 0; F * = F * ′′ = G* = 1; F * ′ = 1 (58)
We have
η* = ∞, F * ′ (∞) = G* = 0 (59)
2α 2 + α 1 − 2β 1 = 3α 2 − β 2 Γ − β 1 (63)
Subject to the initial condition (58) Eqs. (56) and (57) are in- α 1 + α 2 − β 1 − β 2 = 2α 2 − β 2 δ − β 2 = − β 2 Γ + 2 (2α 2 − β 1) (64)
tegrated numerically until F * ′ and G * remain unchanged. These
α 2 − 2β 1 = 0 (65)
M. Patel, M.G. Timol / International Journal of Non-Linear Mechanics 85 (2016) 54–61 59
2α 2 − 2β1 + β2 (1 − Γ + δ ) = 0 (69)
B ∫0 F¯ ′dη¯ = Constant
Then the set of (Eqs. (15)–17) reduces to The transformation in Eq. (82a) reduce the Eq. (75) to
∂ B2 ∞
F′ 2 + FF ′′ = − (3 + 2Γ ) (G Γ F ′′) ∫0 F¯ ′ 2dη¯ = 1
∂η (73) A (86)
Hence we have to obtain the solution of Eqs. (73) and (74) with
(3 + 2Γ ) ⎡ ∂ ⎤ the boundary conditions (81) and (82). Again the boundary con-
2F ′G + FG′ = − ⎢ (G Γ G′) + σG Γ F ′′ 2⎥
σ ⎣ ∂η ⎦ (74) dition at the second point is transformed to non-homogeneous
And the condition will be form. To transform the boundary value problem with boundary
conditions (81) and (82) into initial value problem, following
∞
∫0 F′ 2dn = 1
(75)
transformation is defined:
F = P λ 1F *
Along with the boundary conditions:
η = P λ 2 η*
η = 0; F = F ′′ = G′ = 0 (76) (87)
G = P λ 3G *
η → ∞, F ′ = G = 0 (77)
where P, λ1, λ2, λ3 are constants.
Thus the boundary value problem is transformed to an initial
value problem. The system of equations is;
4.1. Solution technique (in the presence of viscous heat dissipation)
∂
F * ′ 2 + F *F * ′′ = − (3 + 2Γ ) (G *Γ F * ′′)
∂η * (88)
If we look at the set of boundary conditions obtained in the
present case one finds that the boundary condition at second point 3 + 2Γ ⎡ ∂ ⎤
2F * ′G* + F *G* ′ = − ⎢ (G*Γ G* ′) + σG*Γ F * ′′ 2⎥
i.e. η = ∞ is homogeneous. This condition is to be transformed to σ ⎣ ∂η* ⎦ (89)
non-homogeneous form. For this let the transformations be
With the initial conditions
η = Aη η* = 0; F * = F * ′′ = G* = 0; F * = G* ′ = 1 (90)
F = BF
(78) The condition for above equations to be free from P is,
G = CG
λ1 + λ2 + Γλ3 = 0 (91)
where A, B, C are constants, Now Eqs. (73) and (74) reduce to;
2 ′ ′ ′′ = − (3 + 2Γ ) ∂ (G¯ Γ F¯ ′′) To integrate Eqs. (88) and (89) two more initial conditions are
F¯ ′ + FF
∂η¯ (79) required (two missing initial condition introduced for the first
time) i.e. the value of F * ′ at η = 0 and the value of G* at η = 0. In
order to fine these missing initial conditions it is assumed that,
⎡ ⎤
¯ ¯ ′ = − 3 + 2Γ ⎢ ∂ (G Γ G¯ ′) + σG Γ F¯ ′′ 2⎥
2F¯ ′G¯ + FG
σ ⎣ ∂η ⎦ (80) F ′(0) = P and G (0) = P (92)
With the transformed conditions; Under the transformation of Eq. (87), the Eq. (92) will turn to
η̄ = 0; F¯ = F¯ ′′ = G¯ ′ = 0 (81) F * ′ (0) = P ⇒ λ1 − λ2 = 1 (93)
G* (0) = P ⇒ λ3 = 1 (94)
η = ∞; F¯ ′ = G¯ ′ = 0 (82) Therefore
And From the above equations the conditions of invariance of
transformation should be λ1 = ( Γ + 1)/2
AB = C Γ (82a) λ2 = ( 1 − Γ )/2
λ3 = 1 (95)
From the given boundary conditions it is seen that F′ starts
from a certain constant value at η = 0 and approaches zero as Thus finally we have following set of equations;
60 M. Patel, M.G. Timol / International Journal of Non-Linear Mechanics 85 (2016) 54–61
[27] Ahmet Pinarbasi, Coskun Ozalp, Selim Duman, Influence of variable thermal 4 (6) (2013) 407–418.
conductivity and viscosity for nonisothermal fluid flow, Phys. Fluids 17 (3) [29] A. Abbassi, H. Ben Aissia, Forced-mixed convection transition of a buoyant
(2005) 1–4. axisymmetric jet with variable properties, FDMP 10 (1) (2014) 115–147.
[28] G. Borah, G.C. Hazarika, Effects of variable viscosity & thermal conductivity [30] R. Seshadri, T.Y. Na, Group Invariance in Engineering Boundary Value Pro-
and magnetic field effect on the free convection and mass transfer flow blems, Springer-Verlag, Germany, 1985.
through porous medium with constant suction/heat flux, J. Comput. Math. Sci.