How to calculate the temperature rise in a

sealed enclosure
Often times electrical or electronic components are housed in sealed enclosures to prevent the
ingress of water, dust or other contaminants. Because of the lack of ventilation in these
enclosures all of the heat generated by the internal components must be dissipated through the
walls of the enclosure via conduction then from the external surface of enclosure to the
environment via radiation and natural convection as shown in figure 1.

Figure 1. Heat transfer from a sealed enclosure with heat generating components
Accurately calculating the temperature rise of each component housed inside the enclosure is a
complicated task that is best accomplished using computational fluid dynamics and heat transfer
software. However in many cases being able to estimate the average air temperature within the
enclosure based on the dimensions and the material of the enclosure is sufficient information to
allow you to develop a design that can be further refined through testing.

The path of the heat flow out of the enclosure is represented by the general thermal resistance
network shown in figure 2.  The four vertical walls of the enclosure can be considered a single
wall by combining the areas of each wall. The top and bottom walls of the enclosure have to be
evaluated separately since the heat transfer coefficient of each is different.
Figure 2. Sealed enclosure thermal resistance network
The heat must first be transferred from the air inside the enclosure to the internal surface of the
enclosure walls. Rvi, Rti and Rbi represent the thermal resistance associated with the heat transfer
to the internal vertical, top and bottom surfaces respectively. The internal thermal resistances are
highly dependent on the number of heat sources, arrangement of the heat sources and location
of internal structures that affect the air flow within the enclosure. Because of the variability of
these parameters a highly accurate estimate of this thermal resistance using simple hand
calculations is not possible. However by making some assumptions regarding the layout of the
components within the enclosure a reasonable estimate of the internal thermal resistance can be
calculated. The following assumptions will be made:
1. The heat sources are evenly distributed throughout the enclosure
2. The structures in the enclosure do not significantly obstruct the movement of air flow throughout
the enclosure

If there are no fans or other type of forced convection within the enclosure the heat transfer from
the internal air to the walls is via natural convection.  Correlations used to calculate the heat
transfer from horizontal and vertical plates [1] will be used to estimate the heat transfer from the
top/bottom surface and vertical surfaces respectively. The equations used to calculate the
thermal resistance associated with the convective heat transfer from the air inside the enclosure
to the walls of the enclosure are as follows:

Convection to Internal Bottom Surface

1
where:
2

3
average air temperature inside the enclosure
average enclosure internal wall surface temperature

Convection to Internal Top Surface 5

where:

 
Convection to Internal Vertical Surfaces 7
where:

8
9

Equations 3,6 and 8 are developed in [1] and are only applicable with air as the flow medium and
laminar flow. For most electronic applications the flow within the enclosure will be laminar.

Radiation can account for a significant percentage of the heat transfer in situations involving
natural convection as is the case with a sealed enclosure. The radiation heat transfer from the
heat generating components to the internal walls of the enclosure is represented by R radi. Since
our simplified analysis allows us to estimate the internal air temperature only we will use that
value to calculate Rradi. In most situations the internal air temperature will be lower than the
component temperatures. In these cases the calculated heat transfer due to radiation internally
will be understated. Since our goal is to provide an initial estimate of the performance of the
enclosure that will be refined later in the design process this error in the internal radiation heat
transfer is acceptable. Rradi is calculated using the following equations:
Radiation to Internal Surfaces
10
where:
11
12
is the surface emissivity of the enclosure
 (Stefan-Boltzman constant)
Rcond is the conduction thermal resistance of the wall of the enclosure and is given by equation 13
13
where:
is the thermal conductivity of the enclosure material
is the thickness of the enclosure material
The heat generated inside the enclosure is transferred to the surrounding atmosphere from the
external surface of the enclosure via natural convection and radiation.  As with convection to the
internal surfaces the convection heat transfer from the bottom, top and vertical external surfaces
will be evaluated separately. The same convection equations used for the internal surfaces will
be used for the external surfaces with  Ti-Tis replaced with Tes-Tamb where Tamb is the ambient
external temperature and Tes is the external surface temperature of the enclosure.

The thermal resistance representing the radiation heat transfer from the external surface of the
enclosure to the atmosphere is given by 14.

14
where:
15
In order to determine the internal average air temperature you must first determine the surface
temperature of the enclosure Tes.  Tes cannot be solved for directly and will have to be determined
using a numerical solver available in any mathematical software or using the “Goal Seek”
function in Microsoft Excel.   By performing an energy balance at the surface of the enclosure
equation 16 is developed. The value of Tes is determined by finding a temperature that satisfies
equation 16.
16
where:
17
 is the convection thermal resistance from the external bottom surface (reference equation 1
and substitute Ti-Tis for Tes-Tamb)
is the convection thermal resistance from the external top surface (reference equation 5 and
substitute Ti-Tis for Tes-Tamb)
is the convection thermal resistance from the external vertical surfaces (reference equation 7
and substitute Ti-Tis for Tes-Tamb)
is the total heat generated by the internal components
With Tes known Tis can now be calculated using equation 18.
18

Note: The thickness of the enclosure walls are assumed to be sufficiently small such that the
internal and external surface areas are approximately the same.

You are now finally able to calculate Ti using the internal thermal resistances and Tis. As with the
calculation of the external wall temperature the internal average air temperature cannot be
calculated directly. Ti is calculated numerically by finding a temperature value that satisfies the
internal energy balance using equation 19.
19
where:
20

[1] R Simons, “Simplified Formula for Estimating Natural Convection Heat Transfer Coefficient on
a Flat Plate”, in: Electronics Cooling, Issue: August 2001
 

How to calculate the temperature in a

ventilated electronics enclosure
Electrical and electronic components are sometimes housed in enclosures with ventilation openings to
allow ambient air into the enclosure to more efficiently cool heat generating components. The most
common ventilation configuration is to have openings at the top and bottom of the enclosure as shown in
figure 1.

 
Figure 1. Ventilated enclosure cooled via natural convection

The air flow through the enclosure is driven by the difference in air density of the cooler air outside the
enclosure and warmer air inside due to the heat transferred from the heat generating components. As the air
flows vertically through the lower ventilation openings and across the heat generating components its
temperature increases. As a result components mounted at higher locations in the enclosure will be
subjected to air temperatures above the ambient air temperature. Determining the internal air temperature
surrounding these components will allow you to identify and relocate components that have a maximum
ambient operating temperature lower than the surrounding air temperature. The size or number of the
ventilation openings can also be adjusted to lower the internal ambient temperature.

If detailed thermal analysis needs to be conducted on components that have attached heat sinks, the heat
sink can be modeled as if it were sitting in open air and the calculated internal air temperature at the heat
sink location can be used as the ambient temperature.
As the air flows into, through and out of the enclosure the flow restrictions caused by the inlet ventilation
openings, internal components and outlet ventilation openings produce a corresponding pressure drop
denoted by ΔPinlet, ΔPcomp and ΔPoutlet respectively.
The difference in density between air flowing into the enclosure and that exiting the enclosure produces a
pressure difference ΔPbouyancy. This pressure difference must be balanced by the sum of the pressure drops
through the enclosure as expressed in equation 1.
1
where:
2
density of the ambient air
average density of the air in the enclosure
acceleration due to gravity
vertical distance between the bottom of the heat sources and the outlet ventilation openings (see figure 1)

The average density of the air inside the enclosure can be approximated using equation 3.

3
where:
thermal expansion coefficient of air at T∞
average air temperature in the enclosure
Substituting equation 3 into 2 produces equation 4 that relates the average internal and ambient air
temperatures to the pressure difference due to the change in air temperature.

The pressure drop due to the inlet and outlet ventilation openings (equation 5) is related to the speed of the
airflow through the openings and the number of and size of each opening. A representation of a typical
ventilation openings is shown in figure 2.

5
is the air speed through each ventilation opening
is the fluid resistance coefficient

Figure 2. Ventilation openings

The calculation for the fluid resistance coefficient, k is the same as that used in the blog article Estimating
Forced Convection Cooled Heat Sink Performance in a Ventilated Enclosure (Part 1). It will be repeated
here for convenience.
for Re>105
6
for 30<Re<105

7
for Re<30

8
where:
is the total area of the ventilation
is the total open area formed by the ventilation holes
9

10
11
is the kinematic viscosity of air
12
is the area of a single ventilation opening
is the perimeter of a single ventilation opening

Note:   can be calculated for openings of arbitrary shape. For circular openings   the diameter of the
circle.

Assuming the heat generated by the components is uniformly distributed, equation 13 can be used to relate
the heat transferred to the air inside the enclosure via convection Qconv to the temperature difference of the
air leaving and entering the enclosure through the ventilation openings.
 13
where:
is the volumetric flow rate of air passing through the enclosure
is the specific heat of air at constant pressure
is the temperature of the air exiting the enclosure

With the assumption that the heat generating components are distributed uniformly in the enclosure the
temperature of the air increases linearly in the vertical direction through the enclosure. The average air
temperature can then be calculated with equation 14.

14

Substituting the average internal air temperature with equation 14 in equation 4 and combining that with
equation 13 results in equation 15, a new expression for the buoyancy pressure drop.

15

The volumetric flow rate is used to calculate the speed of air through each ventilation opening using
equation 16.

16
In most electronics applications the pressure drop due to the ventilation openings is much larger than that
due to the internal electronic components since they are not densely packed. As such ΔP comp can be ignored
for most applications. Combining equations 5 and 16 and substituting the new values of inlet/outlet
pressure drops and buoyancy pressure difference from equation 15 into equation 1 gives equation 17. This
equation has only one unknown, the volumetric flow rate of the air flowing through the enclosure since
ΔPcomp is assumed too small for consideration.
17

The variables with the subscripts inlet and outlet reference the corresponding variable values for the inlet
and outlet ventilation openings. The value of the volumetric flow rate can be calculated using an equation
solving function available in most mathematics software or using the goal seek function in MS Excel.

Because there is a linear increase in the temperature of the air along the height of the enclosure equation 18
can be used to determine the air temperature as a function of the height location in the enclosure.

18
Note that the total heat loss from the components is equal to the sum of the heat losses due to convection
Qconv and radiation Qrad. A conservative approach is to assume all the losses are due to convection. However
if the component temperatures are high and the ventilation openings are small then radiation may account
for 30% or more of the total heat losses.
ArcEnergy (Electrical)

(OP)

26 Aug 15 15:54

Hello all,
I have already gone through the threads in this forum related to Solar gain, Temperature rise. I have got
enough Information but some how stuck with unrealistic values through my calculations. I hope some you
you can help me in this.

My question is ,to estimate the amount of temperature rise inside the enclosure due to solar loading.
I have a Switchgear enclosure (made of aluminuim) of LxWxH of 2.4mx1.15mx1.97m , Thickness would be
around 0.01m
The internal power Dissipation is 900W .. it is because of a current of 12.5kA flowing in a conductor of
resisrance 6x10^-6 Ohms
The ambient temperature (Tamb) is 40°C
I am considering the Solar Radiation of 1100W/m^2. The enclousure is painted with light Grey colour with
Absorption coefficient of 0.5 and emittence coeff of 0.87 (correct me if iam wrong)
h for the Aluminium as 2 W/m^2K (not sure if this value make sense)

I would like to know how much would be the temperature rise inside the enclosure (Tenc).

i have been using the formule mentioned in other threads like

Sun Radiation * Absorption coeff = emittence coeff * 5.67x10^-8 * (Tenc^4 - Tamb^4) + h * (Tenc-
Tamb)

By using this formulae i am getting the enclousure temp as 95°C .. that means a temp rise of 55°C .. which
i feel is too high.

My end aim is to calculate the amount of flowing current that i should reduce when I place my switchgear
in outdoors.

Since i am an electrical engineer this temperature stuff is Kind of new to me. I hope some one will throw
some lighjt in this Topic.

Thanks in advance
Regards,
ArcEnergy

Replies continue below

Recommended for you

3

racookpe1978 (Nuclear)26 Aug 15 16:42

Look very, very closely at the "assumptions" made in your radiation equilibrium formula!

Your formula is a valid approximation for a "black body" of emissivity = .87 (to correct it to a grey body
assumption) ... with no other heat losses but long-wave thermal radiation.

First: Your solar absorbtion is just over one wall of 2.4 m x 1.15 M (worst case), and then you need to add
the internal resistance heat gain.

But the thermal losses are going to occur from ALL 6 walls of the outside of the box using natural
convection into the air AND conduction losses AND long wave thermal radiation back into the basic 40
degree atmosphere.

So, to simplify things, assume no solar radiation, and just the 900 watts/box heat load trying to get out of
a closed Al box into 40 degree C air. Determine what the approximate temperatures will be for the 6 walls
of the box: Top will be hottest (internal air rises up the sides, cools at the top, then falls down to the
bottom, then rises again; bottom outside will be "cooled" by the outside air slightly differently than the top
and side walls.) So, use three approximate convection equations that total into the 900 watts being
dispursed into 40 deg C air.

From that approximation of bottom, sides, and top wall temperature, determine the radiation losses into 40
deg C (273 + 40) deg air. You can probably ignore the conduction losses down the supports of the box.

So, you have one internal heat input (900 watts) and 12 approximate heat losses (the radiation and
convection losses from each of 6 different wall positions) into 40 degree air. Subtract the 6 radiant heat
losses from the 900 original watts, and re-calculate the 6 wall temperatures. Re-caclculate the radiant heat
losses (again) - The differences should quickly become near-zero for this coarse an approximation.

Then, calculate the single heat gain at each hour of the day from the sun - it will NOT be 1100 watts/m^2
except at noon on the highest day of the year (probably June 22) for your latitude! More likely, it will be
substantially less every other day of the year and every other hour of the year. Give me your latitude - I
can get you a spreadsheet that gives approximate (theoretical clear sky, direct radiation) solar gains per
exposed sq meter for each hour, each day of the year at any latitude.

Add that extra radiation heat gain to the 900 watt internal load for the hours around noon. Then re-
calculate the previous equilibrium wall temperatures to approximate the 9:00 AM, 10:00 Am, 11:00 AM,
noon, 13:00 PM, and 14:00 PM and 15:00 PM air temperatures and solar heat gains. Note that actual
outside air temperatures will maximize between 13:00 and 15:00 each (summer and winter) day, but will
not be (on average) ever highest at noon on any given day.

Unless you are desert near-equator conditions, there will only be a few weeks of the year with high daily
solar radiation and high air temperatures at the same time. You still have to design for worse case
conditions of course, but the air will not be 40 deg C all day and all night. And the sun will not be shining
at 1100 watts all day either.

IRstuff (Aerospace)26 Aug 15 17:51

MIL-HDBK-310 has tables for solar loading; my recollection is that 1100W/m^2 is basically the max for
about 3 hrs during the hot day. Note that the ambient temperature is usually modeled to lag the solar
peak.

You should be able to put racook's approach into Excel and configure an energy balance equation that you
can goal seek, and just have Excel do the iterating.

TTFN
FAQ731-376: Eng-Tips.com Forum Policies
Need help writing a question or understanding a reply? forum1529: Translation Assistance for Engineers

Of course I can. I can do anything. I can do absolutely anything. I'm an expert!

There is a homework forum hosted by
engineering.com: http://www.engineering.com/AskForum/aff/32.aspx

dbill74 (Mechanical)27 Aug 15 13:46

First you appear to be forgetting that the sun moves across the sky and will impart solar radiation at
different angles at different times of day. This also means that 4 sides of the enclosure will receive solar
radiation (up to three at one time) throughout the day. Also local weather conditions will change ambient
temperature throughout the day as well. 95°C appears to be a reasonable temperature to reach inside the
enclosure you described.
Then I get to your 'end aim'; and I have to ask why? 
The switchgear is providing power to something, if you reduce the power going to ...whatever... it is not
going to be able to do what it needs to do. If you want to keep the temperature in the enclosure under a
certain point, you need to know net energy going into the enclosure, solar and internal so you can have an
HVAC engineer calculate how much ventilation to provide.

racookpe1978 (Nuclear)27 Aug 15 18:13

Well, it would be (at maximum) only two sides that could be exposed to direct radiation, not three.

But you do have a good idea to reduce maximum heat load.

1. Get rid of the gray paint - keep ALL of the Al walls shiny. This will reflect more incoming solar energy,
because not only will the sun-lit side be reflecting more (absorbing less) but the secondary radiation from
the ground and skies on the back sides be reflecting more.

2. Rotate the box so only the EDGE of the southern side (if northern hemisphere!) actually faces south.
True, when the sun is at 9:00 Am or 15:00 PM, that one side will face more directly towards the sun, but
the edge will face due south at noon (when sunshine is highest) and the two "flat" faces will be at 45
degree angles from the sun. It will be interesting to determine if the combined two faces - both angled
AWAY from the sun - collect more radiation than a single flat face aimed due south.

3. Even simpler, why not mount a single flat 1/8 (or 1/4 inch) Al plate about 2 inches from the top and
south wall as a sun shade? Use the north wall (if northern hemisphere) for the maintenance access
opening/door. There will be some (not much!) secondary radiation from the exposed sunscreen, through
the sunscreen plate, to the shaded electronics box, but much less. 2 inches of air space are big enough to
serve as insulation, and is large enough for natural convection air flow to be unimpeded.

IRstuff (Aerospace)27 Aug 15 19:45

I'm not convinced that #2 with the tilted faces is going to make that much difference, since the projected
area of the two tilted faces will still equal that of a single, untilted face.

TTFN
FAQ731-376: Eng-Tips.com Forum Policies
Need help writing a question or understanding a reply? forum1529: Translation Assistance for Engineers

Of course I can. I can do anything. I can do absolutely anything. I'm an expert!

There is a homework forum hosted by
engineering.com: http://www.engineering.com/AskForum/aff/32.aspx

dbill74 (Mechanical)27 Aug 15 20:27

Max number of sides getting solar radiation is three sides. Top, south and east/west.
Assuming one side does face south (northern hemisphere), there is a limited amount of time during which
only the top and the one side will see the sun, however the time is limited. The other sides will receive
sufficient solar radiation as to not be negligible.

Shading is a good idea, however 900W of heat generated in the enclosure is significant and ventilation is
still required.

ArcEnergy (Electrical)
(OP)

27 Aug 15 21:25

Thanks alot for the responses.

I want to explain a little more clear about my calculations.

The switchgear is tested with 12.5kA of rated current at 40°C of ambient temperature. according to
standards we are allowed to have a 65K of rise inside the switchgear cubicle. this value is mainly because
the current carrying conductor has only tested to withstand 105°C. as you see the allowed temperature
rise is 105°C-40°C = 65K

I have then calculated the effect different ambient temperatures on conductor current by keeping the
conductor temperature of 105°C constant using the formula

Inew/Irated = sqrt[(105°C - Tamb new)/Temp rise allowed]

Eg: If the Tamb new is 50°C, then Inew = Irated x sqrt [55/65) gives us the new Current of 11.5kA .. That
means when the out side temp is 50°, you are only allowed to pass 11.5kA in the conductor.

Now considering the situation of placing the switchgear in outdoors. Now the additional effect is solar
radiation. According to the standard "C37.24-2003 - IEEE Guide for Evaluating the Effect of Solar Radiation
on Outdoor Metal-Enclosed Switchgear"

for grey painted enclosures, the absorption is 0.50 which increases a temperature rise of 7.7K when the
solar radiation is 1044 W/m2... So in my new calculation, i have added this 7.7K to my allowed 65K temp
rise and calculated the Inew for the amb temp of 40°C. I get the value of 11.64 as allowed current in the
conductor when placed my switchgear in outdoors at 40°C.

THE WHOLE ABOVE CALCULATIONS ARE FINE TO ME IF I DONT GO DEEP INTO THE SUBJ OF SOLAR
RADIATION.

But in the other thrds that posted here about solar radiation, i have read the explanations by IRSTUFF and
IONE and realised this is not as easy as i thought. i have to consider convection radiation concepts which i
am not aware of.

so I have posted my question in terms of Watts

according to racooks explanation i have to perform

Q1 x area of wall 1 = htc x area of wall 1 (Tenc1-Tamb) for 6 sides ..

But for every case i have two unknowns, Q and Tenc .. but as he said sum should be 900W, I will add Q1,
Q2 ...Q6 = 900W.
So My equation would be

900W = 2x[5.52 (Tenc1-313) + 9.46 (Tenc2-313) + 4.54 (Tenc3-313)] ------> 2.76m2, 4.73m2 and
2.27m2 are the areas of 2x3 walls and htc of Al 2W/m2K.
Here for simplicity i have considered Tenc1 =Tenc2 =Tenc3 as 338K

the results didnt make sense to me. I am convecting more than the 900W .   ..I am lost here even before
considering the radiation effect.. 

I am pretty much sure i messed it up somewhere. I was kind of hoping a regular formula to measure the
temp rise inside the enclouse based on the solar radiation. didnt see this coming.  I would be really
thankfull if anyone walk me through this stuff.

Regards
Arcenergy

IRstuff (Aerospace)28 Aug 15 05:06

I think the issue is that you've not iterated on what is supposed to be a system of equations that are
solved simultaneously. The basic premise is conservation of energy flow, i.e., gizintas = gizoutas. When it's
done correctly, you will get a self-consistent set of temperatures. Your guesses are fine, but they are not
cast in concrete. You must iterate on the answer until everything converges, or use a math program that
can do that for you.

TTFN
FAQ731-376: Eng-Tips.com Forum Policies
Need help writing a question or understanding a reply? forum1529: Translation Assistance for Engineers

Of course I can. I can do anything. I can do absolutely anything. I'm an expert!

There is a homework forum hosted by
engineering.com: http://www.engineering.com/AskForum/aff/32.aspx

kjoiner (Mechanical)28 Aug 15 19:25

You may also want to check out some of the enclosure manufacturers such as Hoffman or Haewa. They
have some guidelines for determining heat loads. Below is a link to a Hoffman brochure.

http://www.pentairprotect.com/wcsstore/AuroraStore...

Kyle

denniskb (Mechanical)29 Aug 15 06:30

ArcEnergy,

I got your email so will try to help out.

Wouldn’t it be nice if things were so simple that we had a formula to solve everything? The situation you
describe is in reality complex and so is the solution unless we make some significant assumptions to
simplify the solution without losing rigour.

I guess that a box of this size is supported on a concrete slab with an air gap on the underside of 75-
100mm?

I also guess that the box is ventilated to allow air to enter near the bottom and exit near the top?

Given the equipment type I guess that even short periods above 105°C may not be acceptable?

I also guess that the unit will have a low thermal mass so the temperature will reflect the short term
midday maximum conditions?

The indicated solar radiation level of 1100 W/m2 is extremely high. The sun radiates to the upper
atmosphere at around 1300 W/m2. The atmosphere reduces this intensity to a maximum possible of
~1065 W/m2 direct summer sun on a clear (i.e. not hazy) cloudless day at noon and probably on top of a
mountain. More typical are 850 W/m2 – direct summer sun clear sky maximum, 660 W/m2 - Australian
desert direct summer sun, clear sky maximum, 400 W/m2 - Australian summer mean. As noted in the
responses the maximum solar radiation intensity varies with your location on the planet, the time of day
and atmospheric conditions. My guess is that you could safely design for a worst case of 850 W/m2 and
perhaps lower?

The angle of the surface to the sun directly affects the solar intensity and the angle of the box to the sun
directly affects the area exposed. I believe you can assume direct solar radiation on any one face only so
you should assume the largest face which is the top of the box which also fits with the maximum mid-
summer, midday conditions. This assumption does not imply that there is then no radiant transfer from the
other surfaces of the box; simply that this is not direct solar radiant heating. There will still be radiant
transfer with the surroundings.

For solar radiant heat transfer (and you must remember that this includes two way heat transfer and
always between two surfaces) the wavelength is short and largely penetrates the atmosphere therefore the
associated heat rejection is to the sky and not the local air which is at 40°C. When you fly in a plane the
airline often displays the external air temperature which is often around -35°C. Radiant transfer heat
rejection from the surface will on a clear day be to an assumed surface of between -43°C and -35°C. Yes
on not clear days this radiation may be to a much higher temperature however under these conditions the
incoming solar intensity will also be much lower. If you are using a design solar radiant intensity of 850
W/m2 then you should assume a sky temperature of 35°C.

The values you suggest for the grey paint absorption and emittance are credible but note that you should
have two absorption values – one for Solar (short wavelength) and one for Ambient (longer wavelength). I
would suggest you use 0.5 for both. Alternative paint colours may help while everything is new and shiny
but with time and a layer of dust the improvement will not last. Using bare Aluminium is not a good idea as
its radiant properties are typically worse than paint.

For the sides radiant heat transfer will occur with the surroundings so what is located near to the box will
affect this. It is typically reasonable to assume that the surroundings are at ambient temperature and
located nearby (e.g the ground) and with radiant properties similar to soil. I would suggest you might
ignore the bottom surface of the box as the wall temperature will likely be close to the concrete
temperature.
All real world heat transfer problems involve i) conduction, ii) convection and iii) radiant heat transfer. It is
only in the text books that you can solve just for one. Often one or two are dominant and in this case
conduction can likely be assumed to be not significant. However care is needed because the dominant
mode can vary as conditions approach equilibrium. I manage this by solving for all three from which the
dominant ones become obvious.

Convective heat transfer with the surroundings has its own set of complexities largely due to the huge
number of variations in film coefficients associated with various flow and heating/cooling conditions. I
manage this by solving a large number of film coefficients from which I can choose while noting others
with similar values.

Convective heat transfer through ventilation of the box is almost a different topic again and one that I do
not have much experience with (particularly if it is natural rather than forced) so for now I plan to ignore
this and keep is as a safety factor for the design. Others may know how much benefit is available from
ventilation?

OK that was the background so how solve your problem?

Given the assumed conditions above fix the internal temperature at 40°C and calculate the heat transfer
rate per m2 for solar and convective modes through the top surface and the walls. Multiply the rate per m2
by each surface area to calculate the overall heat inflow/outflow rate in kW.

Assuming a nett heat inflow fix the internal temperature at 60°C and do the same.

When the heat inflow equals or is close to the heat outflow you have found the equilibrium temperature for
the empty box.

When the net heat outflow matches the 900W heat load from the equipment you have found the
equilibrium temperature for the enclosure. I would not be at all surprised to see a 95°C result.

Note that at this point you have not found the actual temperature of the device but the temperature of the
air in the box. However, I am guessing that this is the intended limit for the switchgear?

I will try to run some cases and provide the heat transfer rates for your surfaces and post these shortly.

TTFN suggestion 3 is however a very effective way to improve the outcome but I would suggest one
improvement. On the underside of the shield attach an insulating layer (e.g. a glued on closed cell foam
rubber mat – make sure the glue is suitable) of around 5-10 mm thickness. Without the insulation the box
will experience radiant heat transfer direct with the shield at elevated temperature. With the insulation the
rubber surface temperature will be much lower. Unfortunately this solution makes the heat transfer
calculations even more difficult as you must now determine the natural draft convective heat transfer in
this gap. As I said at the beginning “if only things sere so simple …”

Dennis Kirk Engineering

www.ozemail.com.au/~denniskb

racookpe1978 (Nuclear)29 Aug 15 20:17

Concur with the recommended sun shield, vents, and correcting his assumed (too high!) 1100 watt/m^2
solar insolation loading.

It is not clear why louvers would be rejected/not used - though sand infiltration could be a factor.

Rather than "gluing" rubber on the back of the sun screen, I'd recommend a wood-backing: A plain sheet
of 3/8 plywood on the back of Al sunscreen would reduce the re-radiation without any threat of glue or
rubber deterioration over time. Simpler to bolt on, rigid, low emiissivity for LW radiation, and an adequate
through not perfect insulator so the hot Al sheet on the hot side has a lower thermal radiation on the
cooler, shaded side.

But he DOES have to go through the math sequence of iterated equations for each of the six walls of the
enclosure. No single formula works.

ArcEnergy (Electrical)

(OP)

30 Aug 15 16:24

Hello Dennis,

Thanks alot for acknowledging my email and for your clear explanation.
Most of your assumptions are correct. The switch gear will be installed on a concrete slab. doesnt know the
exact gap though but your assumption seems reasonable.

Your assumption of boy being ventilated is not true. It is completely closed with out any ventilation. i guess
it comes under protection degree IP55 . But i thing you have suggested to ignore the effect of ventilation
which fits to the situation.
and yes even the shorter periods above 105°C would be critical. That is why i am considering a worst case
scenario of very high radiation of 1100W/m2 . even the standards use around 1044W/m2 i guess.

I dont know exactly what a thermal mass means. but based on the explanation from google, the low
thermal mass scenario would fit the application.

With the angle of sun, Since i am considering the worst case situation, I would say your assumption is
good that the largest face area(4.68m2 )is directly exposed to the sun light.

about the solar radiant heat transfer, it is not 100% clear to me. did you mean that the radiation from the
enclosure is of two types? one is into the surrounding atmosphere which is 40°C and other is to sky which
is at -35°C ? Am i correct?

Now coming to the solving,

Only Box. no internal heat dissipation

You suggested me to assume the internal temperature of the enclosure is 40°C and told me to calculate
heat transfer from solar and convective modes.

Heat transfer from solar to internal enclosure (direct sun so largest face) 1100W/m2 x 0.5 x 4.68 = 2574W
is the amount of heat energy resulted through sun light into the enclosure.

Now the convection through the walls to the outside atmoshphere using
q = htc (Tinside-Toutside) x each face . But since our assumed internal temp and outside temp is 40°C,
isnt that the heat convection is 0? 

Solar radiation increases the heat inside to 60°C.

now using convection formulas q= htc (Tinside-Toutside) x each face ..
= 2 (333-313) x 4.68x2 +
x 2.76x2 +
x 2.25x2 ... The total Equals to 774W ..
radiation emission from all sides = 0.5x 5.67x10^-8 (333^4-313^4)[4.68x2 + 2.76x2 + 2.25x2] = 76.5 x
19.38 = 1482W

Total emission is 2256W which is not exactly equal to 2574 but i guess we could assume at 62°C internal
temperature the solar heat inflow would be equal to heat out flow through convection and radiation.
This gives us a rough idea that due to solar radiation of 1100W/m2 shining directly on the top of the
enclosure a temperature rise of 22K (62°C-40°C) is possible.

Please let me know if these above calculations make sense to you are not. And you are right i am mostly
interested about the temperature of the air inside the box.

Regards,
ArcEnergy
Tunalover (Mechanical)30 Aug 15 22:58

ArcEnergy,
I have a well-documented Excel spreadsheet that solves your exact problem. How can I get it to you??
TL

Tunalover

denniskb (Mechanical)31 Aug 15 07:18

ArcEnergy,
“if only things sere so simple …”
You will most likely not be able to make this work with 1100 W/m2 and should change to a more realistic
maximum of 850 W/m2

A. Your simple calculations have not accounted for all heat transfer paths nor included for absorptivity and
emissivity.
Thermal radiation is emitted by all bodies that are at a temperature above absolute zero with an ideal
intensity of H = B x Tabs^4 - where B is the Stefan Boltzman Constant of 5.67 x 10^-8 W/m2.K.
The emission levels are directly affected by the emissivity of the source body E1
The level received by another body in its path are affected by the absorptivity of the receiver body A2.
Therefore the transmitted energy from body 1 to body 2 is H = A2 x E1 x B x (T1 + 273)^4
However surface absorptivity is dependent on wavelength so you need two A2 values A2s and A2b for
solar and background
With two bodies this can lead to five values E1, A1, E2, A2s, A2b but if you assume body 1 is a black body
then E1 and A1 are = 1.00
So for the enclosure (body e) consider only Aes, Aeb, Ee
Note that while you have these values for your grey paint these will be for the new clean condition but will
degrade as the paint oxidises and gets a coating of dust. In my experience you would be better using
values of 0.75, 0.75, 0.85 respectively for Aes, Aeb, Ee.
Three radiant heat transfer paths can affect the enclosure;
 Absorption from the sun for which the intensity (Is) is known - Hs = Aes x Is
 Absorption from the background - Hb = Aeb x 1.00 x B x (Tb + 273)^4
 Emission from the enclosure – He = Ee x 1.00 x B x (Te + 273)^4
You must account for all three of these for the enclosure;
For the top surface;
 Solar absorption
 Absorption from the sky at -35°C – [Yes absorption NOT emission]
 Emission from the enclosure surface
For the four sides;
 Assume solar load is nil
 Absorption from the background at ambient surface temperature (say ambient + 10-15°C)
 Emission from the enclosure surface

B. The convective heat transfer is primarily governed by the inside and outside film coefficients and not the
conductivity of the plate.
The solution requires algorithms to calculate the Nusselt Number (Nu) and requires various fluid properties,
geometric data for the enclosure and fluid flow conditions. Further these must be determined at both
inside and outside faces of the enclosure. Trust me this is no easy task. The heat transfer is directly related
to Nu which can easily vary by a factor of 1 to 10. Therefore if you get this wrong your answer can be way
off.
For example Nu will vary significantly with flow velocity (i.e. Reynolds Number) which for natural draft is
unknown (I have assumed 0.15 m/sec). This is particularly important inside the enclosure because low Nu
will lead to the enclosure wall being at a lower temperature and so reject less heat to outside. One option
is to install fans inside to blow the air along the walls? Another is to weld ribs on the inside of the
enclosure to increase the surface area and so again to increase the wall temperature.

C. Some of the heat transfer paths can have very high H values (as you have found) but some are
absorption (+ve) and some are emission (-ve) and the balance, being the sum of all three, can be quite
small (particularly near the equilibrium temperature). Since you need to reject an additional 900 W heat
load using just the small balance you can get widely varying results from just minor variations.
For example solar absorption 638, background absorption 137, emission -689, convection -100 the balance
is just -14. A change in the emission calculated of just 2% will either delete or double the heat balance!
What then becomes important is to able to test the sensitivity of the outcomes to possible changes in the
assumptions and so to get a “feel” for the reliability of the result. For me I can do this because I have
automated the simultaneous calculations for the three radiation paths and the convective algorithms and
calculations and can re-run cases in less than a second. The problem then is that this opens up all of the
issues noted above which in reality are much more important than the relatively “simple” calculations.
I have run some calculations with the following results (Rad is the sum of Solar + Backg + Emiss, Nett is
Conv + Rad, Load is the total heat load)
Paint Weathered + Dust (assumed 0.75, 0.75, 0.85)
Natural Convection inside and out (assumed 0.15 m/sec)
Top Surface - 4.7 m2, Four sides – 10.1 m2
Inside Skin Conv Solar Backg Emiss Rad Nett Load
Top Surface °C °C W/m2 W/m2 W/m2 W/m2 W/m2 W/m2 W
40 69.4 -87.1 637.5 136.8 -662.4 111.9 24.8 117
50 70.1 -89.7 637.5 136.8 -667.8 106.5 16.8 79
60 70.8 -92.4 637.5 136.8 -673.2 101.1 8.7 41
70 71.5 -95 637.5 136.8 -678.4 95.9 0.9 4
80 72.1 -97.5 637.5 136.8 -683.6 90.7 -6.8 -32
90 72.8 -100 637.5 136.8 -688.7 85.6 -14.4 -68
100 73.4 -102.5 637.5 136.8 -693.7 80.6 -21.9 -104

Side Walls °C °C W/m2 W/m2 W/m2 W/m2 W/m2 W/m2 W

40 40.2 -0.1 0 463.7 -463.5 0.2 0.1 1
50 41.2 -1.6 0 463.7 -469.7 -6 -7.6 -76
60 42.9 -3.4 0 463.7 -475.6 -11.9 -15.3 -154
70 43.1 -5.3 0 463.7 -481.3 -17.6 -22.9 -230
80 44 -7.3 0 463.7 -486.8 -23.1 -30.4 -306
90 44.9 -9.4 0 463.7 -492.1 -28.4 -37.8 -380
100 45.7 -11.4 0 463.7 -497.4 -33.7 -45.1 -453

You will see that at 100°C the total heat rejection load is -104 + -453 = -557 W so has not yet achieved
the desired load of 900 W.
As noted above doubling the inside velocity with fans or doubling the inside surface area would
significantly increase the total heat rejection and likely achieve the desired outcome.

Even with the tools I have developed this is no simple task and needs expertise and time. Amazing
considering this is just a simple box out in the sun but as I keep repeating “if only …..”
Dennis Kirk Engineering
www.ozemail.com.au/~denniskb

ArcEnergy (Electrical)

(OP)

31 Aug 15 07:26

Tunalover,

Can you kindly forward that sheet to my email address

somebodyinnobody@ gmail .com.

Thanks in advance mate :)

regards
ArcEnergy

chicopee (Mechanical)31 Aug 15 20:55

Since solar radiation appears to be a problem regardless of the numbers that you crunch, put a canopy
over your switch gear and be done with this