heat transfer copper tubing

[emesine]

What heat transfer can I expect from a copper tubing immersion type heat exchanger? How many BTU/Ft2-hr-sqft?

Will this change depending on the diameter of the pipe?

Thanks!

 

[Gooserider]

What heat transfer can I expect from a copper tubing immersion type heat exchanger? How many BTU/Ft2-hr-sqft?

Will this change depending on the diameter of the pipe?

Thanks!

Lot of variables, including the differential in temperature, flow rates through the HX, etc. (the greater the difference the more heat will Xfer) I think if you look in that "engineering toolbox" link I put in the tidbits sticky, there are some equations for that (or if someone else wants to post a better link, or the actual formulas...)

It doesn't change directly with the diameter of the pipe, but for a given cross sectional area, several small pipes will perform better than one larger pipe, both because the surface area of the resulting HX will be larger, and because a small pipe will give more turbulent flow, thus bringing more of the hotter water into contact with the pipe walls...

i.e. a 1' length of 3/4" tube will have a cross section of 0.44in2, for a volume of 5.28in3 and a surface of 28.2in2. A 1/2" pipe has a cross section of 0.196in2, so use 2 parallel lengths to get about the same cross section - 0.392in2 for a volume of 4.71in3, and a surface of 37.7in2 - or about 25% more surface area....

From what I've seen, it appears the best / most common configuration for a coil HX seems to be multiple 50-100' lengths of 1/2" tubing in parallel - with the number dependent on the needed flow volume and heat transfer rate. 1/2" seems to be about the best balance point between material cost, construction simplicity, and performance... As with all these projects there are a lot of "right" answers that will work, but this seems to be where most folks end up.

Tom from ME, and some our other experts have also argued that there does not seem to be a lot of practical benefit from paying more to get finned tubes, putting in turbulators, or other approaches aimed at getting more surface area or turbulence to theoretically improve the heat transfer rate...

 

[julien]

http://www.engineeringtoolbox.com/pressfit-piping-friction-loss-d_1167.html

19 gpm in a 1.25" pipe pressure drop is 1.54 psi/100ft

18gpm in six 1/2" parallel pipes, 3 gpm/pipe, pressure drop is 2.3 psi/100ft

In 8 pipes the drop is approx 1.4 psi/100ft

 

[Rick Stanley]

I have recently come across a section of 3" copper pipe 56" long. Am thinking about trying to build a large heat exchanger to make dhw.

Has anyone tried winding 1/2" soft copper around a 1" pipe or something like that to make a long coil to insert into a large tube? I'm trying to post a link to a pic of the configuration I mean. Think it will coil or would it kink or flatten? Any thoughts? Suggestions? Thanks.

http://cgi.ebay.com/Tarm-Wood-Boile...emQQptZLH_DefaultDomain_0?hash=item33572b8312

 

[Gooserider]

I have recently come across a section of 3" copper pipe 56" long. Am thinking about trying to build a large heat exchanger to make dhw.

Has anyone tried winding 1/2" soft copper around a 1" pipe or something like that to make a long coil to insert into a large tube? I'm trying to post a link to a pic of the configuration I mean. Think it will coil or would it kink or flatten? Any thoughts? Suggestions? Thanks.

http://cgi.ebay.com/Tarm-Wood-Boile...emQQptZLH_DefaultDomain_0?hash=item33572b8312

Nice idea, but I'm not sure how tight you can bend 1/2" copper tube - I know it has a minimum bend radius, but I'm not sure what it is. It is also a question of what sort of equipment you have, the people that make commercial coils push the bend limits to the max, but they do it with special machinery designed for the job.... You might be able to do it with a bunch of smaller diameter tubes wound in parallel, or possibly if you can do something like one of our other users did recently and added a bunch of extra fins onto the central straight pipe (not sure how much difference that made, but...)

Gooserider

 

[in hot water]

It can be wound, even finned tube. But it takes a machine built for the job. This is 1/2" inside a 4" copper tube. Several companies still wind coils for tankless boiler HX, Heat Transfer Products is one, you might check there for a wound HX coil.

Get some 1/2" 180 bends and go back and forth the long way inside the 3".

Here is one of my ideas to improve the efficiency of those heat pipe HXers. I just cut up some scrap copper and soldered it onto the center tube to get more surface area. Cheap and easy. Not sure how much it added to the HX efficiency.

hr

 

[in hot water]

Another attempt.

hr

 

[emesine]

I've never personally tried, but I have heard you can bend copper easier if you fill it with sand (keeps it from kinking).

So, I can run 18 gpm through 6 100' copper pipe lengths...... total cost of around $1200 plus about 20 hours of my time in building the manifold, soldering the pipe, fixing burn holes in my tank lining where the solder dripped, etc. When I replace the lining of my tank in 10 years I get to do it all over again. Frankly, sounds like a nightmare.

Or I can buy a big plate HX for $600.

600 ft of 1/2" copper comes out to 78 sq feet of area. A 5 X 12 X 40 HX has around 13 sq feet of area. Even if my transfer is 5X worse with copper tubing sitting in still water, it would still be enough.

Thanks for the info.

 

[rowerwet]

you could also get a tubing bender that is a long spring that fits exactly over the outside of the pipe, I don't know where else to get them but aircraft tooling companies like aircraft spruce or brown tool sell sets in 1/2", 3/4", and 3/8" that work great, also being a spring you just keep sliding it down the tube as you bend it. they are made for bending aluminum tubing so copper would be easier.

 

[julien]

I've never personally tried, but I have heard you can bend copper easier if you fill it with sand (keeps it from kinking).

So, I can run 18 gpm through 6 100' copper pipe lengths...... total cost of around $1200 plus about 20 hours of my time in building the manifold, soldering the pipe, fixing burn holes in my tank lining where the solder dripped, etc. When I replace the lining of my tank in 10 years I get to do it all over again. Frankly, sounds like a nightmare.

Or I can buy a big plate HX for $600.

600 ft of 1/2" copper comes out to 78 sq feet of area. A 5 X 12 X 40 HX has around 13 sq feet of area. Even if my transfer is 5X worse with copper tubing sitting in still water, it would still be enough.

Thanks for the info.

I presume 6 X 50' are sufficent to make a job, and you can use thin refrigeration tubing. Sure surface in hx tranfer is smaller because the water move each side of transfer wall, and speed water is higher. But the price for this efficiency is more maintenance and additional circulator and electricity.


And i suggest a stainless steel tank for several decades of use without care.

http://equipementsderabliere.elapierre.com/default_en.asp?no=101

 

[DenaliChuck]

I have successfully bent copper tubing into very tight coils using the sand method. It was 3/8" tubing in a 4" diameter coil.

 

[hkobus]

What heat transfer can I expect from a copper tubing immersion type heat exchanger? How many BTU/Ft2-hr-sqft?

Will this change depending on the diameter of the pipe?

Thanks!

http://www.engineeringtoolbox.com/overall-heat-transfer-coefficients-d_284.html

Hope I did the link thing right, but this seems to be the answer to the first question. I t will change with the diameter of the the pipe, the larger the diameter of the pipe the less heat transfer surface per unit of fluid to transfer heat from. This should also help deciding between the cost difference between the flat plate and the copper coil.
There are many more links on the above site relating to efficiency of HX coils etc.
Another site to check is Nofossil's site, he has a spreadsheet to aid in the calculations.

http://www.engineeringtoolbox.com/heat-load-steam-pipes-water-d_287.html

Added this link later, may help with heat HX design.


Henk.

 

[emesine]

Yes, I've seen that. How about this:

According to Engineering toolbox, water-steel-water heat transfer is around 60-80. This is STILL water, essentially no movement on either side of the exchange. Numbers for copper are comparable.

On the other hand, it is easy to get heat transfer over 1000 using a steel plate heat exchanger, where there is movement on BOTH sides of the metal plate.

In my situation, there is still water on one side (the storage tank) and moving water on the other side (water moving through the copper tube). It seems reasonable to me to split the difference for this situation. I am guessing I can expect around 500 btu/(ft2-F-hr).

So...... if I need a 12X5 by 40 plate HX with a square footage of about 13 (conclusion from two companies that make plate HXs), then I will need about double this surface area for a coil HX.

Does that sound reasonable?

Andrew

 

[hkobus]

Yes, I've seen that. How about this:

According to Engineering toolbox, water-steel-water heat transfer is around 60-80. This is STILL water, essentially no movement on either side of the exchange. Numbers for copper are comparable.

On the other hand, it is easy to get heat transfer over 1000 using a steel plate heat exchanger, where there is movement on BOTH sides of the metal plate.

In my situation, there is still water on one side (the storage tank) and moving water on the other side (water moving through the copper tube). It seems reasonable to me to split the difference for this situation. I am guessing I can expect around 500 btu/(ft2-F-hr).

So...... if I need a 12X5 by 40 plate HX with a square footage of about 13 (conclusion from two companies that make plate HXs), then I will need about double this surface area for a coil HX.

Does that sound reasonable?

Andrew

http://woodnotoil.googlepages.com/home

You may want to check this site out. I am currently building my tank but still on the fence on the HX system. I think you are right on the capacity of these flat plate HX's, and they may have as much as 1250 Btu/*F/Hr, but the flow rates are really high and you my find the Delta T hard to keep up to get the total Btu transfer. I am also concerned with the turbulence and the disruption of stratification in the tank.
Maybe woodnotoil will chime in and update the change of HX.

Henk.

 

[DaveBP]

If you are building your own nonpressurized tank and want to use a flat plate that needs to use high flow to get the BTU exchange required don't LET the turbulence bite you. In an open tank it's simple to bump up the size of the pipe in the tank so the same GPM comes out at a much slower speed.

Say you are coming off your boiler with 1-1/4" pipe at 10GPM. That's about 2.5 feet-per-second. If you added an elbow and went into a couple feet of 2" pipe to open into the tank that water entering the tank is only moving about 1 FPS. Use 3" pipe and you're only crawling at a few inches-per-second. That will cut your turbulence issues down considerably, probably to a non-issue. This will work on a pressurized tank as well but requires welding.
Going into the larger pipe with an elbow will help slow the water down (turbulence can be a good thing) and not punch through the center (laminar flow) of that larger diameter body of water without slowing down to the volume of the larger pipe.

This will help if you need to run higher flow rates to squeeze those last BTUs out of the burn with a lower temp difference across the HX than the first part of the burn. And I think everyone will want to do that, especially if you are running high temp baseboards.
When in doubt, if you're going flat plate HX, go for the larger pipe size ports. They offer less resistance to the higher flow rates. That way if you want to push your envelope you have the "head room" to use a larger circ and get the flow rate up.

 

[emesine]

So... DaveBP,

You would use a plate HX over a copper coil? I see benefits to both, and I am very much on the fence. Plate HX has simplier installation and serviceability, plus better performance. Copper immersion has questionable performance (meaning no one can tell me exactly what the heat transfer will be), but less moving parts, probably more reliable than worrying about fouling in the HX, pump corrosion, etc.

Still on the fence,

Andrew

 

[Gooserider]

OK, first thing to realize, is that regardless of it's type, figuring out the performance of a heat exchanger is a hairy and nasty problem... There are a bunch of variables, and most of them keep changing... The biggest of these is the Delta-T, or temperature difference between the fluid on one side, and the fluid on the other... Not only is this going to change over the course of a burn as the tank heats up, it will even change from one part of the HX to another at any given moment... Really the best that can be done is a series of approximations to try and make an educated guess as to the actual performance.

Cribbing a lot of info from the Engineering toolbox links above, we start out with a "still water" figure for a water to water HX, with copper as a separating medium, of 60 - 80(Btu / ft2 / hr / *F) or in metric 340 - 455(W / m2 / *K) In essence this says that you will see 60-80 BTU's per hour transferring across each square foot of HX area for each degree of difference in the two temperatures... I'll split the difference and use 70 (Btu / ft2 / hr / *F)

A nominal 1/2" tube will have a surface area of about 18.8in2 per foot of length, or 7.65' of tubing per square foot of surface area... Call this "one HX unit"

Since the overall performance gets worse as the tank heats up, and you get less and less difference between the boiler temp and the tank temp, we need to figure for something close to the worst case performance to get reasonable results, but going for the absolute worst case would need a near infinitely large HX as the temperature difference approaches zero, so we should probably solve for 3 cases just to see what difference it makes, say 5*F, 10*F and 20*F.

Per NoFo and other posters, we could assume that our boiler output is about 75% of rated power, but we also know that our peak tank heat should be happening towards the end of the burn as the fire is entering the coaling stage and the boiler is starting to cool slightly. Thus I'm going to assume a 150K BTU/Hr class boiler, with an actual heat to transfer of 100K BTU/Hr. (it makes the math a little easier!)

The formula above, assumes "still water" - how much difference does it make that this water is moving? I don't think it makes much, as if one takes a small enough time slice, the water is essentially still... Since the conditions are constantly changing, the number we are going to get is only going to apply for an instant in any case. Pumping water through will have the effect of increasing the temperature difference as there will be less time for the pumped water to transfer it's heat, obviously the faster the flow, the greater the difference...

So to figure for each temperature range, we get 70 Btu's of transfer per unit of HX, per degree of temp difference, so

CAP / (dT) (BTU) = HXU x (Sz) = L
where:
CAP = total amount of heat to transfer (BTU / hr)
dT = Difference in temperature (*F)
BTU = BTU transfer rate (BTU / hr / ft2 / *F)
HXU = 1 square foot of HX area
Sz = Length per HXU
L = Total length of HX tubing required

100,000 / (dT ) (70) = (HX units) x 7.65 = feet of 1/2" tube needed

100,000 / 5 x 70 = 285.7 x 7.65 = 2,186 ft at 5*F difference
100,000 / 10 x 70 = 142.9 x 7.65 = 1093 ft at 10*F difference
100,000 / 20 x 70 = 71.4 x 7.65 = 546 ft at 20*F difference

It seems fairly obvious from this, that we will probably be best off working with a 20*F difference in order to keep the HX an affordable size. If we assume a 190*F maximum boiler output temp, that would in theory give us a 170* max tank temp, but since as stated above, this is an "average" temp, the tank top is likely to be higher. Also, while the BTU/hr capacity goes down as the temp differential decreases, it will still transfer SOME heat as long as the boiler is hotter than the tank, so presumably one could approach the boiler output temperature although more slowly...

This equation can also be worked in the other direction - assuming the 546' HX from above, what sort of heat can we expect to transfer with a cold tank? Assume the same 190* boiler output temp, with protection to limit the boiler input to 140*. That gives a 50* dT. Reworking the equation, we get

(HXU) (dT) (70) = CAP

71.4 x 50 x 70 = 249,900 BTU / Hr - or considerably more than the nominal boiler output...

Gooserider

 

[DaveBP]

You would use a plate HX over a copper coil? I see benefits to both, and I am very much on the fence. Plate HX has simplier installation and serviceability, plus better performance. Copper immersion has questionable performance (meaning no one can tell me exactly what the heat transfer will be), but less moving parts, probably more reliable than worrying about fouling in the HX, pump corrosion, etc.

When I was planning my tank last year the price of copper skyrocketed. So I switched to consider flat plates. Then I went back and forth for a spell. I don't know why they can't make a fence with a comfortable seat but I still don't know which I would have gone with. I think my choice would be heavily effected by the current price of copper. But a flat plate ain't cheap either if you need to stuff a lot of water through it. Regardless of the type, with any heat exchanger I would advise that you choose the most expensive one you can afford and get the next bigger size.
When it comes to engineering a solution, when I get to 3 consecutive assumptions I usually lose faith in the process let my checkbook (or a chance Ebay bargain) decide.

Honestly, not being able to come to a clear conclusion was what drove me to pressurized propane tanks when a couple of 500s came my way.

 

[in hot water]

Here are some formulas if you want do do some long hand plate HX calcs, from Caleffi I-dronic 6. It is an issue about solar combi systems, but the same concepts apply to wood fired transfer and storage. www.caleffi.us has the issue as a free download, along with 1-5.

hr