The complexity of the
experimental diagnostics of various liquid flows is mainly caused by their
three-dimensional and unsteady nature, as well as the presence of significant
gradients of all parameters that are of interest for practice. Under natural
convection in heated bodies, the main thermophysical processes develop in a
thin boundary layer several millimeters thick.
In certain cases, physical properties of the liquid
are also subject to significant changes in this layer [1], which must be
considered when processing experimental data.
The small thickness of the layer complicates the use of contact methods
of experimental investigation. Consequently, the development of contactless
optical methods for convection investigation and the comparison of experimental
results with calculated data is highly relevant thermophysical task.
Contactless
refraction methods for measuring temperature inhomogeneities in liquid [2 – 4]
are characterized by comparative simplicity and visibility. Temperature
gradients in liquid are visualized due to the deviation of the ray trajectory when
passing through a medium with a variable refractive index.
The direct task of
calculating the refraction of light on a temperature inhomogeneity in liquid is
identical to calculating the transmission of a ray through an optical medium
with a gradient change in the refractive index [5].
Whereby, the profile of the change of the refractive
index along the ray trajectory differs from the standard optical models of
gradient optics and is defined by a functional or numerical dependence on the
spatial coordinate.
The inverse
task of reconstructing the three-dimensional distribution of the temperature
profile from refraction images, or refractograms can be solved within the
framework of the adopted model of the temperature dependence on the refractive index
of liquid.
In the
present paper, smoothly inhomogeneous temperature fields are considered, in
which the function of dependence of the refractive index on the coordinate does
not have abrupt jumps and discontinuities.
For
cases when vortex flows arise in the medium [6], the solution of the inverse
task of reconstructing the temperature profile is very difficult and in most
cases impossible.
To solve the direct
task of refraction on a temperature inhomogeneity in liquid, it is also
possible to apply a numerical calculation method based on solving the equations
of geometric optics. It allows investigating inhomogeneous media with a random functional
temperature dependence.
The
present paper demonstrates the possibility of numerical calculation method’s
application for solving thermophysical tasks, as well as shows the possibility
of solving the inverse task of reconstructing the temperature profile in an
inhomogeneity with radial symmetry.
The principle of operation of computer-laser refraction
(COLAR) method [2] is based on probing the flow with a laser sheet beam - an
astigmatic beam with an elliptical cross-section, the size of which along one
of the axes is significantly larger than the other. Such beam is also called a
"laser sheet" (LS).
The setup (fig. 1) consists of a laser
1,
an optical system for the formation of a LS
2, a cuvette with liquid
3,
a heated body
4, a diffusing screen
5, a digital video camera
6,
a personal computer
7
and a special software.
Fig. 1.
Configuration of experimental setup of COLAR method
LS is directed under a heated object situated in a
cuvette with liquid. During the LS propagation through an optically
inhomogeneous flow, due to refraction, its particular areas deviate from the
plane surface, i.e. a complex curved surface, changing both in space and in
time, is obtained. LS projections at the exit from the optical inhomogeneity
(refractograms) are observed on the screen, recorded by a digital video camera
and entered into a computer for further processing.
Fig.
2 demonstrates examples of
refractograms
when a laser sheet beam passes through liquid near heated bodies of various
shapes.
The
example of an optical system for calculating a positive temperature
inhomogeneity under a heated sphere is represented in fig. 3; the rectangle
shows the liquid volume. The laser beam in the first approximation is replaced
by a beam of rays, the deflection of which is shown in a vertical plane passing
through the middle of the sphere in a vertical section. In the represented
case, the temperature of the sphere
Òb
was assumed to be
higher than the temperature of liquid
Òl. The calculated refraction
curves conform with the coordinates of the beam on the right wall of the liquid
volume, introduced according to the coordinate binding [7]. In the considered
model of positive (Òb
>
Òl)
spherical temperature inhomogeneity, the dependence of the refractive index
n
on temperature
Ò
was calculated by the formula [2, 7]:
.
|
(1)
|
Heated
object under investigation
|
Experimental
refractograms
|
|
Sphere
|
|
Parallelepiped
|
|
Cylinder
|
Fig. 2.
Experimental refractograms of a laser sheet beam for various heated bodies in
liquid
1 – laser
beam, 2 – cuvette with liquid, 3 – heated sphere, 4 – screen
Fig. 3.
Optical
scheme for calculation of refraction in a temperature inhomogeneity under a
heated sphere
For
the heated sphere model the temperature
Òb
=
Ò
was determined relative to the water temperature
Òl
=
T0,
depending on the distance to the center of the sphere
r
[7]:
,
|
(2)
|
where
T0,
ΔT, ΔR, a
– temperature field model parameters. The parameter
T0
is determined by the temperature of the walls of the cuvette with
liquid, the value of
T
(R) is equal to the temperature on the
surface of the sphere, and the ratio Δ
T
/
a
corresponds to the
gradient of the temperature field in the boundary layer with the thickness
à.
The value of the temperature gradient at
r
=
R
is determined by the
shift
Δ
R,
whereby
the thermal conductivity of the sphere is not equal to the thermal conductivity
of the liquid, there is a jump in the temperature gradient on the sphere
surface and grad[
T
(r)]≠0. However, the investigation does
not exclude the possibility of equality or closeness to zero of the temperature
gradient in the boundary layer (adhesion of the boundary layer), which
determines the choice of the Gaussian model of the temperature field. Fig. 4 represents
dependencies
T
(r) and
n
(r), plotted for the
parameters
T0
=20 °Ñ,
Δ
T
=80 °Ñ,
Δ
R
=0,
R
= 12 mm,
a
=1 mm at a section
R
<
r
<
rmax,
where
rmax
= 14,5 mm, an approximate radius that is
close to the asymptotic value
n
(T0).
Fig. 4.
Temperature profile and refractive index profile of water for the model:
R
=12 mm,
T0
=20 °Ñ
,
Δ
T
=80 °Ñ,
Δ
R
=0,
a
=1
mm
(approximation by the 5th degree polynomials)
The
calculated refraction patterns at a distance of 155 mm from the center of the
sphere in the direction of beam propagation, obtained from analytical expressions
and using a numerical algorithm, are demonstrated in fig. 5.
Fig. 5.
Refraction curves for a heated sphere:
●●● – analytical calculation,
– calculation result CodeV
Refraction methods for investigating
optically transparent inhomogeneous media, based on measuring the deviation
value of the probing radiation elements, allow high-quality visualization of
the area under investigation. However, obtaining quantitative characteristics
in most cases is difficult due to the complex spatial dependence of the
refractive index of the medium. The analytical solution of both direct and
inverse tasks can be obtained only for a plane-layered and radially-layered
medium. In other cases, the solution of the direct task is possible only with
the numerical methods application.
If
the dependence of the refractive index on the coordinates
n
(x,
y,
z)
is known, then to find the equation of the ray trajectory, one can use the
well-known differential equation [5], which makes it possible to determine the unitary
vector
l
(x,
y,
z) tangent to the ray at
each point of its trajectory:
.
|
(3)
|
In [8], a numerical method for solving the
reduced equation (3) is considered. The method is based on a step-by-step
calculation of the trajectory of a ray passing through the area under
investigation, and consists in finding a ray vector specified by the direction
cosines in each section.
The
computational region is divided into cells of size Δ
zk,
Δ
xk
and Δ
yk
(fig. 6), within
which the value of the refractive index is considered constant:
. The section of the trajectory within each
cell is represented by a straight line. The method for choosing the size of the
computational mesh cell is introduced in [8] and is based on solving test tasks
that have an analytical solution.
The
boundary condition of the problem is determined by the ray vector
l
0
at the entrance to the inhomogeneous medium, namely, by its direction cosines (cos α0, cos β0, cos γ0)
and the coordinate of the entrance to the inhomogeneous medium (x0,
y0,
z0).
The
gradient of the refractive index in the kth cell is determined by the proportion
. Knowing
l
k, one
can calculate the angle φ
k
between the unitary vector
l
k
and the vector grad
k
n
based on the
scalar product:
.
Fig. 6.
Computational domain cell
The
equation for the angle of rotation Δφ
k
of the
vector
l
k
in the cell has the form [8]
.
The
vector
l
k+1
in the next cell will equal
, where
ν
k
is
the principal normal vector to the trajectory,
.
As a result, at the exit from the
computational region, we obtain the coordinates (xexit,
yexit,
zexit)
and direction cosines (cos αexit, cos βexit, cos γexit)
of the ray vector
l
exit. The refractogram at the exit from
the optical inhomogeneity can be calculated based on specifying a family of
rays at the entrance to the inhomogeneity.
The calculated refraction pattern of a laser sheet
beam passing under the bottom of the cylinder is demonstrated in fig. 7. The
optical inhomogeneity contains a qualitative edge effect observed in the
experiment [9, 10], as a result of which the refraction pattern will have a
strongly pronounced kink near the edge of the cylinder. The following
parameters are accepted in the model: a cuvette with a thickness of 125 mm, a cylinder
in the middle, cylinder radius:
R
= 17 mm, Δ
R
= –0,14
mm,
à
= 0,7 mm,
T0
= 20°C, Δ
T
= 60°C.
The calculated curves are in satisfactory agreement with the test calculation
using analytical expressions [9].
Fig. 7.
Refraction
curves for positive cylindrical inhomogeneity: ●●● –
numerical calculation,
–
calculation result CodeV
The
boundary layer in cold water around a heated body with a spherical structure
(for example, a sphere or a cylinder with a hemispherical bottom) has a complex
structure; however, in the region near the lower point of the hemisphere, the
temperature field is spherical-layered with a high degree of accuracy. The deviation
value of individual LS elements from the initial position depends on the
temperature gradient along the line of propagation of the LS, while the
deflection of the LS element occurs along a line passing through the center of
spherical symmetry.
Let
LS initially pass at a distance
h
from the body. Let us consider an
algorithm for processing refractograms in accordance with the scheme shown in fig.
8.
Fig. 8.
Algorithm
for processing refractograms
Let
us briefly describe the stages [11]:
1)
All refractograms are preliminarily processed to plot the dependence of the
maximum deviation of the LS element from the initial one along the
Y
axis (point
G) on time and body temperature.
2)
An arbitrary point
A
with coordinates (xt,
yt)
of the refractogram is considered and the point of intersection of the straight
line containing the segment
OA
with the level
Y
=
R
+
h
is found. It will be point
B
with coordinates
xp
= (R
+
h)
xt
/
yt,
yp
=
R
+
h, which are
obtained from considering the triangle
OAF.
3)
The value of deflection of the LS element along the selected direction equals
to the length of the segment
AB
= Δ
r
= [(xt
–
xp)2
+ (yt
–
yp)2
]1/2.
4)
The value of Δ
r
is used to determine the temperature of the medium
in a given direction at a distance
OB
=
r
from the body
surface according to the dependence constructed at stage 1 of the present
algorithm. As a result, the dependence
T
(r) is plotted.
It
should be noted that in the experiment the temperature of the body surface is
recorded, and not the temperature of the medium at a distance
h
from the
body. Therefore, to reduce the inaccuracy in the obtained distribution of the
liquid temperature in the boundary layer, it is recommended to choose the value
of
h
as the smallest possible, which, as a rule, is determined by the
size of the LS in the waist, which is located under the lower point of the
hemisphere.
Fig.
9 demonstrates graphs of the dependence of the water temperature on the
coordinate in the boundary layer based on the processing of refractograms (fig.
8) according to the mentioned algorithm for various temperatures
Tc
of the bottom of the cylinder.
Fig. 9.
Dependency of water temperature from coordinate in the boundary layer of the
cylinder at the surface temperature:
1 – 75 °Ñ
, 2
– 66 °Ñ
, 3
– 56 °Ñ
, 4
– 46 °Ñ
The
results of calculating the refraction patterns of laser beams in thermal
inhomogeneities in liquid, performed using a numerical algorithm, are in
agreement with test calculations performed analytically and with a laboratory
experiment. The presented computational algorithms are a reliable tool for
solving direct and inverse tasks of natural convection by the laser
refractography method. The appliance of the laser refraction method in the
investigation of thermal fields in liquids also makes it possible to visualize
graphically the structure of the temperature field.
1.
Polezhaev V.I.
Convection
and heat transfer in the near-critical area: research results in terrestrial
and space conditions // Proceedings of RNKT-3. 2002. Ò.1. P.94-99.
2.
Evtikhieva O.A., Raskovskaya
I.L., Rinkevichius B.S.
Laser refractography. – M.: FIZMATLIT, 2008. – 176
p.
3.
Evtikhieva O.A., Raskovskaya
I.L., Rinkevichius B.S., Tolkachev
A.V.
Fundamentals of laser
refractography. // Modern optical methods for flows investigation:
Multi-authored monograph. – M.: Overlay, 2011. P. 12 – 24.
4.
Artemov V.I., Yankov G.G., Evtikhieva
O.A., Lapitskiy
K.M., Raskovskaya I.L., Rinkevichius B.S., Tolkachev
A.V.
Numerical
and experimental investigation of natural convection in liquid near a heated
cylinder // Proceedings of the Fourth Russian National Conference on Heat
Transfer. – M.: Publishing house MPEI, 2006. Ò.3. P. 42 – 46.
5.
Kravtsov Yu.A., Orlov Yu.I.
Geometric optics of
inhomogeneous media. – M.: Science, 1980.
6.
Varaksin A.Yu., Romash M.E., Kopeytsev V.N.
Physical modeling of air
tornadoes: issues of generation, stability and control. // Modern optical
methods for flows investigation: Multi-authored monograph. – M.: Overlay, 2011.
P. 252 – 274.
7.
Evtikhieva O.A.
Refraction of the laser sheet in a spherically
inhomogeneous thermal boundary layer // Measurement equipment. – 2006, ¹ 5. P.
35 – 39.
8.
Lapitskiy
K.M., Raskovskaya I.L., Rinkevichius B.S.
Algorithm for
calculating refractograms of a laser sheet in an optically inhomogeneous medium.
// Measurement equipment. – 2009, ¹5. P. 36 – 40.
9.
Tavrov A.V., Lapitskiy
K.M.
Calculation of the refraction of
laser beams in liquids applying optical design programs
// Proceedings
of the 10-th International Scientific and Technical Conference "Optical
Methods for Flows Investigation". / Edited by Yu.N. Dubnischev, B.S. Rinkevichius.
– M.: Publishing house MPEI, 2009. P. 198 – 201.
10.
Lapitskiy K.M., Raskovskaya I.L., Rinkevichius B.S.
Calculation of refractograms
in three-dimensional inhomogeneous media with edge effects // Proceedings of
the 10-th International Scientific and Technical Conference "Optical
Methods for Flows Investigation". / Edited by Yu.N. Dubnischev, B.S.
Rinkevichius. – M.: Publishing house MPEI, 2009. P. 210 – 213.
11.
Lapitskiy
K.M., Rinkevichius B.S., Tolkachev A.V.
Quantitative
visualization of natural convection near a heated body using a laser sheet // Proceedings
of the 17-th School-seminar for young scientists and specialists under the
guidance of. Academician of Russian Academy of Sciences A.I. Leontiev
"Problems of gas dynamics and heat and mass transfer in aerospace technologies".
– M.: Publishing house MPEI, 2009. Ò.1. P. 115 – 118.