Free engineering help needed

[model 10]

I am trying to build a spare heat exchanger. I can't afford to buy one. I've built 2 from copper and could not control the leaks, so copper is out for me. Now I have the material to build one from stainless and I should be able to tig weld it up leak free.
Question: Copper has thermal conductivity of 400 and stainless is only 14. Does that mean the stainless heat exchanger needs to be 28x bigger than the copper job? I know this very simplistic but any WAG?

[Dennis.G]

Although you are correct on the comparative thermal conductivities of SS versus copper, the conclusion of 28x heat exchanger surface required is not correct. Most of the issue of heat transfer between the two fluid steams has to do with the properties of the fluids themselves and their flow conditions. Yes, a SS heat exchanger will require more surface area than a copper one, but not nearly as much more as you suggested.

Maybe play with this: Chemical Engineering Now Online Calculations

[2Hulls]

Originally Posted by Guy

Question: Copper has thermal conductivity of 400 and stainless is only 14. Does that mean the stainless heat exchanger needs to be 28x bigger than the copper job? I know this very simplistic but any WAG?

Better double check your 400 value for copper. A quick google check returned me values more like 200-225 for pure copper. Alloys of copper are quite a bit lower and are in the range of the SS. I assume by "size" you mean surface area of the tubes.

Dave

[Cheechako]

Yes, the water that flows thru cooling the hot water is most important. Heat transfer is required though, not sure how that works out in the end.... but I don't think my perkins heat exchanger had copper tubes running thru it... cupro nickel?
OP: have you checked used marine places? Those places used to have a bunch of universal type heat exchangers around... not sure now...

[2Hulls]

Originally Posted by Dennis.G

Most of the issue of heat transfer between the two fluid steams has to do with the properties of the fluids themselves and their flow conditions.

Correct, but in this comparison the fluids and flows would be the same.

Dave

[exMaggieDrum]

Good quality heat exchangers are expensive. But losing your engine, or your boat, is more expensive. Copper is one of the easiest materials to seal with solder or brazing but building a heat exchanger with internal and external tubing, and tubes going into the sides or other tubes is not the easiest project to do. I would save up my nickels and buy a used one if you can. If used get it inspected and serviced and put it in your spares. Make sure it fits. And cupro-nickel is a better material. But if you DIY good luck. And you may have to put in a zinc holder for Cu or Cu-Ni. I wouldn't recommend stainless but I'm not an engineer so FWIW.

[Boatguy30]

One of the Perkins bowman style should be cheap to make or even buy a spare core for.

[Dennis.G]

Originally Posted by 2Hulls

Correct, but in this comparison the fluids and flows would be the same.

Dave

Yes, but my point is that in the total net heat transfer between the two streams the conductivity of the tube walls has only an series type resistance effect, and is not to be considered a proportional factor of heat transfer.

An electrical analogy (the fluid thermo being much more complicated of course):

R(total) = R(fluid1) + R(tubewall) + R(fluid2)

If R of the fluids is much higher than the tube wall R, differences in tube wall R will have a small effect on R (total).

Look at the this linked graphic. I think is kind of shows what I am saying. Even if the temperature slope across the wall is changed quite a bit, there is little effect on the overall temperature profile (and rate of heat exchange).

http://nptel.ac.in/courses/103103032...0/images/1.png

[Frankly]

Dennis is close to hitting the nail with the hammer. The limiting resistance in this heat transfer process is the thermal resistance of the liquid film against the heat exchanger material. Actual thermal conductivity of the HE material is not a big part. Turbulent liquid flow helps a lot in this regard.

What is important in the material selection end is corrosion resistance. SS is probably not real high on the list of suitable materials. Most HE tubes are probably made from a copper nickle alloy.

I am in the camp of buy a heat exchanger and replace the cost by recovering aluminum cans alongside the roadway.


____________________


"someday you are the dog and some day the tree"

[2Hulls]

Originally Posted by Dennis.G

Yes, but my point is that in the total net heat transfer between the two streams the conductivity of the tube walls has only an series type resistance effect, and is not to be considered a proportional factor of heat transfer.

I don't want to belabor this, but "series type resistance effect" has no meaning in heat transfer, from what I remember. I don't understand electricity so whatever analogy to heat transfer that might exist is lost on me.

Q dot = m dot X cp X dT

The only variable in this particular comparison between two choices of materials of the heat exchanger tubes is cp, the heat transfer coefficient which is made up of the material properties of the tubes (thermal conductivity) and surface areas. The desired heat transfer rate (Q dot), flow rates (m dot), and temperature difference across the tubes (dT) are the same in each case. So, in fact, the thermal conductivity of the tube walls is exactly a proportional factor of heat transfer.

Dave

[model 10]

I am running our spare factory heat exchanger now. Can't see spending another couple hundred bucks for a used that could fail right away. I don't see any great quality in the factory job anyway but it is sized correctly. I want a spare that is as good or better than the factory in durability anyway. If I had not botched the first 2 tries , I would only be in this $30 us but I'm starting to loose it now.
In the end it will be about how many RPM I can use. Min for our boat would be about 1300rpm which is about 3 knots. If it works as good as the factory job that would be something.

[hellosailor]

No offense meant, but if you can't solder copper, I would expect leak-free welding in stainless to be way less likely to happen. In the oil and pipeline industries, they won't even interview for pipeline welders who have less than ten years experience on the job--because getting welds, even in those big pipes, without leaks, is so damned hard. And all you need is one weak spot that pinholes to ruin everything.


It also isn't just a matter of metal conductivity rates. "Good" heat exchangers are designed to create turbulent flow in each pipe, rather than the normal laminar flow. If you don't accomplish that, with laminar flow you get hot fluid passing down the center of each tube, without transferring heat in or out, and only a "skin layer" of fluid in each tube actually being cooled. Part of this depends on the diameter of the tubing, and I've no idea what they do internally to make the fluid "non-laminar" as it flows.
Just FWIW.


With the price of stainless (stainless anything!) it could be an expensive experiment, versus trying to find a secondhand cooler.

[Terra Nova]

Originally Posted by Frankly

...What is important in the material selection end is corrosion resistance. SS is probably not real high on the list of suitable materials. Most HE tubes are probably made from a copper nickle alloy.

I am in the camp of buy a heat exchanger and replace the cost by recovering aluminum cans alongside the roadway...

Gotta go along with this.

Stainless is perhaps the worst choice for a heat exchanger. And this is a case where trying to save less than $200 could cost you your engine or even your boat.

[model 10]

Re soldering vs tig welding; I can tig weld which is why I'm want to use stainless steel. I don't know if you can solder up a whole bunch of little copper tubes, roll up a body out of copper sheet, solder the whole mess together and have it not leak, I couldn't do it in 2 tries. Maybe the third but I'm moving on.
All this has me thinking back to fancy oil coolers in aircraft. Even the insides of the tubes or passages had little fins. It made them impossible to clean if you had a component failure because you could not get the dirt out. I can make little twisted vanes for each tube in mine..

[Dennis.G]

Originally Posted by 2Hulls

.. So, in fact, the thermal conductivity of the tube walls is exactly a proportional factor of heat transfer.

Dave

It is only a proportional factor of heat transfer across the tube wall itself, in fact is proportional to conductivity times delta T across the tube wall. But the delta T across the tube wall is small in comparison to the total delta T from one fluid stream to the other.

The formula Q = M x Cp x DT just gives the overall heat transfer between two streams when given flow, and delta T for a fluid (M being mass flow and Cp being liquid specific heat (NOT thermal conductivity), and DT is delta T). Has nothing to do with calculating heat flow from the conductivity properties of fluid or exchanger walls.

Anyway, look at real world heat exchangers that use different tube materials. You will find that they do not differ to any huge degree size-wise according to materials when designed to the same values of M, Cp and DT (i.e., for same Q rating in kW or BTU/hr). Back to OP's question, NO the tube surface area is not proportional to tube material thermal conductivity in overall heat exchanger design, but is effected by it to a lesser extent.