# heat transfer question

• posted

We are trying to determine the heat transfer for a system. the details are: enclosed vertical cylinder, 7" diam x 10 tall, 1/8" wall borosilicate glass, 250W heater element immersed in oil, anodized Aluminum top cover, 75 deg room temp.

Questions.

1. How to determine outside wall glass temp?
2. How to determine top cover temp?
3. by increasing cylinder volume how much will the glass/aluminum temp drop?

the aluminum top cover has fins that extend into the oil. If we remove the cover completely, how hot will the glass/oil get.

reply directly if possible

clay

Loading thread data ...
• posted

Dear clay:

Find the emissivity of the borosilicate glass as a function of temperature. Find the transmissivity of the borosilicate glass as a function of wavelength. Find the emissivity of the oil as a function of temperature. Find the emissivity of anodized aluminum as a function of temperature. Solve for convection from a vertical cylinder, convection of oil to the top cover, and radation from all surfaces.

Or hire someone to do this.

Or build it, and measure.

I'd *guess* close to 200 degF. Based on the skin left on a light bulb of roughly equivalent dimensions...

Not what newsgroups are for.

David A. Smith

• posted

Hm, yes, you could add radiation in there if you really wanted to. At those power levels over that type of surface area, you're not going to see high temperatures (and accordingly, low heat transfer via radiation). Radiation will have a very slight impact on the overall surface temperature of the cylinder.

An easy quick conservative calc (neglecting radiation) is:

Surface area = 297 in^2 = 2.06 ft^2 Convection coefficient = 2 BTU/hr/F/ft^2 (ASHRAE handbook, still air)

so:

Q=SA*h*dT

250 W = 2.06 ft^2 * 2 BTU/hr/F/ft^2 * (Tw - 75F) Tw = 136F

Hot, but not too bad.

Dave

• posted

message news:...

thanks guys,

In response to Dave #1. We already have built it. Someone else "supposedly" did the calcs and came up with the preferred volume. I fabricated the parts "based on their specs" over a year ago, didn't hear again from the customer until last week. according to them, it was getting REALLY hot. How hot?, did they measure it? no. but it was "really" hot.

Now they want some help. need to determine if increasing cylinder height (oil volume) will do as much to reduce glass temp as increasing the surface contact area between the top cover fins and the oil. The aluminum cover is 200X more conductive than the glass. I can also increase the top cover thickness AND surface area significantly. 2nd question, do I really want to get involved again with these guys?

the oil is a dielectric mineral oil, and is around 2X less conductive as water. So it is working "somewhat" as an insulator, in other words it has a heat capacity. so it is going to take some time for the oil to heat to an equilibrium correct? is there a formula for calculating the time?

The top cover is basically a big heat sink, and is transferring heat in all three modes, radiation likely the most effective mode correct? the top cover is isolated from the glass by an o-ring. how do I determine the difference between when the cover is off, and making contact with the oil?

clay

• posted

Dear clay:

...

A "really" firm number!

1) will they pay? 2) will they allow you to measure? 3) will they tell you how hot is hot enough? (perhaps the load can be brought up to incremental levels, even if just for a test)

Solutions:

1) increase surface area. Height would be the first/best avenue. 2) force convection. Blow on the thing with a fan. Or pump water through tubes in the oil convection space. 3) Dave Harper is just too smart for his own good (;>)), so increasing emissivity would be a distant third alternative. Look into dark black coatings on the aluminum. If the system will function with carbon black in the oil, do that.

Does the system run only for short periods of time? If not, then you are simply wasting your bullets establishing a time constant.

Probably not unless it is "too hot". ;>)

It won't matter, really. The heat will primarily convect up off the inner source, flow radially outwards, then down along the vertical walls, to be reheated. So the top will always be the hottest spot, even if it takes a little longer to get there. And metal always "feels" hotter.

Unless you have a flame or explosive hazard, consider "guards". If there is such a hazard (and oil that is "too hot" can be), then stick to 1 and 2 above. Also there are nice silicone oils that have a high dielectric constant, and reduced ability to burn (look for "flash point", "vapor point, and "toxicity").

Ultimately, unless the customer is part of the solution (measurements, access, and "ownership"), then he is part of the problem. You decide if you can do the job *in spite* of the customer... and any lawyers he/she/they might bring to bear.

David A. Smith

• posted

Excellent. thanks David S.

Some things I hadn't considered. Customer will pay, they are just somewhat difficult to work with, and are a few hours away. And have very little engineering experience. I can certainly set up for testing etc.. just wanted to know what my expectations should be up front. before I make the commitment and trip. forced convection is an option if only it can be quiet, so I don't know how effective that will be. the water cooling is interesting. the max temp is a subjective number, it was just a lot higher than they expected.

On the time issue, I was trying to establish a relative time frame. will I have to watch it rise for 6 hours to reach level, or as fast as a half hour. actual run time is in the 8-10 hour range.

in your solutions, you didn't include the aluminum top cover and increasing its surface area, increasing fins in oil surface contact area. Increasing the height of the cylinder is an option, but the simplest (short term for testing) is the top cover. Have any thought on that? see below..

right now there is air space between the oil and top cover (a lot actually ~ 1/4"), with only minimal contact between the oil and top. I can increase that contact area dramatically (80-90%) Do you have a guess if this will be significant?

silicone oil is something we are looking at, but it is significantly more expensive, and generally only available in 55gal drums. We only have a need for 4-5 gallons/year.

clay

• posted

Looks like you are not the expert they need.

Major cooling method will be convection.

My advice, if they paid promptly for the last job, offer to adapt or build fresh anything they want, but stay away from any desine work.

Temperature labels will tell how hot it's getting, then you need to find out how hot they think to hot is... again if they can see the temp labels as they run it up and judge when it's at the limit they want to work to you have enough information to start making intelligent guesses at what is required ( a little tinkering like a couple of fins on the lid and a fan, through to replacing the entire apparatus { preferred option if they are good payers ? :-) }.)

• posted

Dear clay:

...

Free advice is always worth what you pay for it.

A contact temperature measurement device would be a good start. An optical or IR thermometer would not be terribly useful on "transparent" surfaces (the quartz might be an issue, but I'm not sure). Confirmation from the customer as to "how hot is too hot" is next. Find out if they can run the load at less than 250 watts (or whatever).

Quiet is easy, but "quiet" is another subjective measurement. The velocities associated with natural convection for objects that are not "too hot" are quite low. Directing a stream of high velocity air, from a squirrel cage type blower should make marked improvements.

You'll still have water to cool down, but it can be applied to a surface area elsewhere.

I'd guess that 10 minutes would get you pretty close to 90% of final surface temperatures. If you wanted to, you could do a plot of temperature vs. time...

Top convection is small (especially on so small a diameter with heated perimeter), and radiation, as I was reminded, is also small. Yes, you could be sure that the aluminum was intimately in contact with the oil, and drop the net temperature by 5 degF or so.

Not terribly significant. The top will have difficulty "attracting" air to convect its heat away, since the sides are already heating its source. But I could be wrong, and often am.

Amortize the expense on two or four systems, and keep the extra cash. It should be safer...

David A. Smith

• posted

I am basically with Jonathan on a cooling approach. An aluminum top with fins or fingers dipping into the oil, and a fan ontop of the aluminum top piece will provide almost any desired temperature reduction.

And that's an illustration of just how much you are in the dark. Why heat the oil? For what purpose? How hot does it need to be? Can less than 250 watts be applied? If you really pull down the oil temperature with a redesigned top/fan arrangement, will that defeat the whole purpose?

It would be better to know the purpose of the whole thing, but you can earn an honest crust without. Better remember your disclaimers when tendering your bid.

Brian Whatcott Altus OK

• posted

thanks for all the suggestions & help. Here is what it is:

the heaters are RF transmitter vaccum tubes. the whole electronics PCB is submersed in dielectric fluid, inside a clear glass cylinder, capped with an aluminum plate/oring combination on either end. No cooling fans. If we add fans they will have to be very very quiet. The aluminum top plate can be finned all over the place. currently it is not. The purpose of the oil is to keep the transmitter tubes cooled of course. We are trying to determine which of the cooling options is going to provide the greatest effect. Increasing the fluid volume/glass surface area, or increasing the convective surface area of the aluminum top cover (adding fins, cooling fingers etc... Since the TC of aluminum is 200X the glass I was thinking that would be where to put the effort. but I wanted a basis to start from.

thanks, clay

• posted

Dear clay:

Change the glass to aluminum. Anodize it black. Make it taller (see note). Without a fan, thermal lift will have to do the cooling (heated air rises). Brian is correct, as this (fan) is how CPUs are cooled. But metal fingers into the oil will have to be arranged to draw heat from the "rolling donut" of fluid in the chamber, but not reduce the size of the flow or its velocity. This is *not* simple, and may not go as planned on first pass.

(note: if fingers are inserted into the oil, the aluminum *top* could be made taller-sided to good effect, without increasing the amount of sidewall/oil volume)

Fans can be had that are "very quiet". You cannot achieve a definite result, without definite numbers.

David A. Smith

• posted

Unless one has high voltage insulation expertise and is aware of the these particular aspects of the design, I'd be extremely leery about changing material from an electrical insulator to a conductor. ________________________________________________________ Ed Ruf Lifetime AMA# 344007 ( snipped-for-privacy@EdwardG.Ruf.com)

formatting link

• posted

Dear Ed Ruf:

Good point. Dielectric oil is only good to so many volts per mil... and it can absorb moisture from the air (something silicone oil can be less prone to do).

David A. Smith

• posted

Anodizing it will help, but the majority of the heat transfer will be convective. I'm not sure what the cost of anodizing is now, but it might be cheaper to look at increasing the convection instead.

In response to an earlier post, adding fins inside the container won't do anything for the surface temperature. It will help decrease the internal temperature of the liquid, but won't change the heat flux leaving the container. The external convection coefficient, surface area, and emissivity (to a lesser extent) are the primary factors that need to be increased in order to cool the surface temperature.

Dave

• posted

message news:...

Dave,

The top cover is isolated from the liquid oil by 1/4" of air space. So at the moment, the top cover is doing nothing (effectively) for cooling. Here is the part that is confusing about your comment about adding fins. If the internal liquid temp is lowered (by adding conducting fins to the aluminum top) then why does that NOT decrease the surface temp. Are you referring the glass surface temp, or the aluminum cover surface temp or both. If the surface area of the (highly) conductive aluminum top is increased substantially, (by adding fins) and conductive fingers into the oil added, wouldn't the percentage of the heat convected off of the top cover increase substantially, thus lowering the glass surface temp?

clay

• posted

Dear clay:

The efficiency with which heat is delivered to the inside surface, does not change the temperature the outer surface must get to to release the heat (both by convection ~95% and radiation ~5% or less).

If you are not further than this, clay, then you might need to hire someone to help you with this.

Problem: Outer surface too hot.

Given: Heat_to_transfer = constant * area * temperature_difference

Constants: Heat_to_transfer, constant (unless you change something drastically)

Variables: area, temperature_difference

Ideal solution: minimize temperature_difference

So it seems clear to me that area must go up if temperature_difference is to go down.

Drastic changes:

- add a fan

- increase emissivity (change that 5% to 6% ;>))

- invoke cooling from the top (may not be that drastic without a fan)

David A. Smith

• posted

Clay:

Sorry, I didn't remember that part. Taking that into account, adding fins will help a little. But the surface area of the top is the minority of the surface area. However, it won't eliminate the excessively hot sides by itself. The ratio of the heat flux through the top vs. the sides won't change too much.

Dave:

Yep, bingo. The same amount of heat has to leave the surface, regardless of what's going on inside. It will lower the liquid temperature, but the higher convection coefficient causes a zero net change in heat flux through a given wall.

Clay, in regards to Dave's suggestions:

...I agree that they will work. I would also suggest maybe increasing the surface area. If you're worried about fan noise, you can always locate the fan somewhere else and "pipe" in air... although I'm not sure that's an option.

Increasing the surface area and adding a fan are the two most effective things you could do.

However, let me really screw things up by throwing out an alternate solution... :-)

Increase the thickness of the sides SIGNIFICANTLY with a good insulating material (glass, lexan, etc). Go back to the fins on top, and use a fan on just the top. The top with stay hot (and the internal temperature might go up some), but the side temperatures will be reduced. The total heat leaving the container will be the same, but by doing this, you're significantly changing the ratio of heat flux from the top vs. the sides. To get some realistic numbers, you'd probably need a math model to play around with.

Hope that helps. Dave

• posted

message news:...

I think Dave might reconsider his take on heat flow rate. What nobody will contradict is that if the electrical input exceeds the heat transfer from the container, its temperature rises without limit! If the heat transfer from the container exceeds the electrical input, the temperature would fall without limit.

So what always happens is this: after an initial transient, the temperature of a container settles at the level where heat off take EXACTLY equals electrical input. If radiative convective and conductive paths are improved, the excess temperature needed to balance the heat flow in this way is reduced.

This is your interest. Can you more easily improve the means to radiate, conduct or convect, you ask. Radiation is not a major contributor at temps you are considering: Conduction and convection are worth considering.

A blown finned head that dips fingers in the oil would be suitable. Or if you prefer a lower fuss method, you could buy some heat pipes and fix them to the existing head. The upper end of the pipe would ideally be blown or fixed to a big thermal mass. Or if you want to show a new device, a head with an oversized diameter and fins like a motobike cylinder head would work without being blown.

Brian W

• posted

Once again I'm going to state, more bluntly this time, if you have no high voltage insulation design/analysis expertise you need to be extremely leery in making system changes, material wise and geometry wise. There may direct path and surface tracking concerns. The OP has given absolutely zero information in regard to the electrical side of this device. To propose to make any such changes is unwise at best, imo. I did my graduate work in high voltage gas blown switches and then 4+ years in high power gas laser design. This is not a discipline to be taken lightly. It may not be an issue, but that needs to be an informed engineering decision. Ignorance is not bliss. ________________________________________________________ Ed Ruf Lifetime AMA# 344007 ( snipped-for-privacy@EdwardG.Ruf.com)

formatting link

• posted

Just one addition... Incorporating additional conductive (hence a capacitive link to ground) and more paramagnetic substances in the vicinity of this circuit may well change how it operates, including adding modes of ambient electrical interference that make the circuit unstable/unusable.

I think clay needs to hire *you*!

David A. Smith

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.