Nusselt number, while 4 ¼ 270 (¼3p/2) the minimum in a range of
Ra < 1400. Compared with the natural convection of a single cylinder
in an infinite large space [8,9], the averaged Nusselt number is
about two times smaller, which is mostly due to that the driven
force is coming from one side of the hot or cold cylinder.
4. Comparison with the experimental results
An experimental loop was set up in order to make a comparison
with the obtained numerical results. The experimental system includes
a test section, a cold and a hot water loop, a flow rate control
system and a data acquisition system.
In the test section, a plexiglass circular pipewas used as the tube
shell, which is 300 mm in length, and the inner diameter and
thickness are respectively 20 mm and 2.5 mm. Two stainless steel
microtubes are circulated respectively with cold and hot water,
both with outside diameter of d (¼1.0 mm), wall thickness of
0.125 mm, and total length of 352 mm (effective heat transfer
length of 281 mm). The cold and hot microtubes were placed
horizontally and in parallel with the axis of the plexiglass circular
pipe. In the cross section view, two microtubes have a tube pitch of
1.5 mm and are placed in the center of the pipe shell.
The whole experimental section was thermally insulated by a
type of aerogel material having a thermal conductivity about
0.02 W/(mK) in room temperature. The deionized water was used
for all of the experiment fluid, i.e. filling up the pipe shell, and
circulating through the two microtubes. The inlet temperatures of
water flowing opposite into the two microtubes are controlled by
two water baths, respectively. The adjustments of hot and cold
water flow rates were realized, respectively, by two vessels, in
which the pressures were maintained by connecting with a common
nitrogen cylinder.
All of the temperatures were measured by T-type thermocouples
with 0.1mmin diameter, which have an uncertainty of 0.1 C.
The upstream or import pressures of hot and cold water were
measured by two pressure sensors with a full range of 0.5 MPa and
an accuracy of 0.1%, and the exits are free flow. The mass flow rates
through the hot and cold tube were measured, respectively, by two
electronic balances, which have an accuracy of 1 mg. All of the
measured data were recorded in every 3 s for approximately 5 min
after confirming that the experimental system reached a hydrodynamic
and thermal steady state.
It is found that the numerically calculated Nusselt numbers
were in general agreement with the experimental results, as shown
in Fig. 10, and the differences becomes obvious at low Rayleigh
number, which is believed that the two dimensional limitation
might be the brief reason. In fact, the natural convection in axial
direction may contribute a fraction especially with low flow rates
and a large temperature difference. A further comparison of numerical
simulation in three dimensional with the experimental
data is needed for our next step work. The experimental data fit
well with the numerical simulated at relative large Rayleigh
number, because the large flow rates can reduce the temperature
difference from the inlet to the outlet that the three dimensional
effect is weakened. A correlation was proposed based on the numerical
results, which is as follows,
Nusselt number, while 4 ¼ 270 (¼3p/2) the minimum in a range of
Ra < 1400. Compared with the natural convection of a single cylinder
in an infinite large space [8,9], the averaged Nusselt number is
about two times smaller, which is mostly due to that the driven
force is coming from one side of the hot or cold cylinder.
4. Comparison with the experimental results
An experimental loop was set up in order to make a comparison
with the obtained numerical results. The experimental system includes
a test section, a cold and a hot water loop, a flow rate control
system and a data acquisition system.
In the test section, a plexiglass circular pipewas used as the tube
shell, which is 300 mm in length, and the inner diameter and
thickness are respectively 20 mm and 2.5 mm. Two stainless steel
microtubes are circulated respectively with cold and hot water,
both with outside diameter of d (¼1.0 mm), wall thickness of
0.125 mm, and total length of 352 mm (effective heat transfer
length of 281 mm). The cold and hot microtubes were placed
horizontally and in parallel with the axis of the plexiglass circular
pipe. In the cross section view, two microtubes have a tube pitch of
1.5 mm and are placed in the center of the pipe shell.
The whole experimental section was thermally insulated by a
type of aerogel material having a thermal conductivity about
0.02 W/(mK) in room temperature. The deionized water was used
for all of the experiment fluid, i.e. filling up the pipe shell, and
circulating through the two microtubes. The inlet temperatures of
water flowing opposite into the two microtubes are controlled by
two water baths, respectively. The adjustments of hot and cold
water flow rates were realized, respectively, by two vessels, in
which the pressures were maintained by connecting with a common
nitrogen cylinder.
All of the temperatures were measured by T-type thermocouples
with 0.1mmin diameter, which have an uncertainty of 0.1 C.
The upstream or import pressures of hot and cold water were
measured by two pressure sensors with a full range of 0.5 MPa and
an accuracy of 0.1%, and the exits are free flow. The mass flow rates
through the hot and cold tube were measured, respectively, by two
electronic balances, which have an accuracy of 1 mg. All of the
measured data were recorded in every 3 s for approximately 5 min
after confirming that the experimental system reached a hydrodynamic
and thermal steady state.
It is found that the numerically calculated Nusselt numbers
were in general agreement with the experimental results, as shown
in Fig. 10, and the differences becomes obvious at low Rayleigh
number, which is believed that the two dimensional limitation
might be the brief reason. In fact, the natural convection in axial
direction may contribute a fraction especially with low flow rates
and a large temperature difference. A further comparison of numerical
simulation in three dimensional with the experimental
data is needed for our next step work. The experimental data fit
well with the numerical simulated at relative large Rayleigh
number, because the large flow rates can reduce the temperature
difference from the inlet to the outlet that the three dimensional
effect is weakened. A correlation was proposed based on the numerical
results, which is as follows,
การแปล กรุณารอสักครู่..