Computing Head Loss in Gravity conversion

JeffGuy

I have found some good info on computing head loss to figure out a good circulator pump to use (most of it found on this website). But I have four questions that I can't find answered anywhere. I would really appreciate some help here.



1. When computing head loss for a gravity conversion system, I understand that I should look only at the longest radiator pipe run and ignore the rest. But  since this is a gravity conversion, there are several different sizes of pipe in that run, with 4", 3", 2", and 1-1/2". What should I do to compute head loss when different sizes of pipe are being used? Should I figure the head loss for each pipe size (times its length), and add these together? Or should I use the smallest pipe size and figure what head loss would be if the whole run was piped at this size? Or should I compute head loss for the smallest pipe size only and ignore the rest? Or should I ...???



2. Is it possible or necessary to include head loss for the radiator which is part of this run? Or can the radiator be ignored?



3. I have two zones in my house - the radiator gravity conversion zone, and a slantfin baseboard zone. These will be piped separately from secondary T's and will have separate circulator pumps, sharing only a small amount of the new secondary pipe. I assume I should figure the heat loss, target flow rate and head loss for these two zones completely separately. Is this correct?



4. I assume I can ignore the primary (boiler) zone piping completely when figuring head loss for these secondary zones? Is this also correct?



Man, it's nice to have somewhere to ask a question and expect a reasonable reply. THANK YOU!!!

Jamie Hall

Some answers...

hopefully useful.



1 -- it's a pain in the neck, but the correct way to do it is to figure the head loss for each different size and length, and add them up.  One does need to be a little wary -- if there is another loop, for example, which might be close it never hurts to add that one up too, and check -- the total length may be less, but if it has a lot of smaller pipe, the head loss could be more, and that's what counts.



2 -- you can ignore the radiators



3 -- Yes, that is correct.  Figure them separately



4 -- I believe so.

JeffGuy

Not just one ...

But all questions answered, and in two hours!



Thanks, I've got some math to do!



Did you help figure out the specs on your heating system and pumps when you replaced the boiler in your house museum? I'de appreciate any thing you learned the hard way and could pass on (it sounds like I have a somewhat similar problem to the one you had - in fact I had two Smith/Mills coal to oil boilers from 1945 in my basement when I started this process). Thanks again



Jeff

Steamhead

On a gravity conversion

there's very little head loss, since all they had was the difference in weight between hot and cold water to make the water move. That's why the pipes are so big.



When B&G addressed this in their 1940's-era handbooks, they suggested we size pumps at 3-1/2 feet of head. This has worked well for me.



Take a look at this chart, in which I ran the numbers so they would work for any brand of circ. All you have to do here is know the system's total EDR, and select your circ from the charts:



http://www.heatinghelp.com/article/343/Circulators/238/Sizing-Circulators-for-Hot-Water-Heating-Systems

JeffGuy

Excellent article

Thank you for the ref to the article. Very helpful, but it raises a big question for me.



If I count the EDR of my radiators I get 1188 sq ft (using the radiator sq ft estimate paper on this website). A couple radiators were removed from the kitchen and small bath (Grrr...) by a previous owner so EDR would have been a bit higher. Using the B&G table in your paper I see that the GPM in my conversion should be about 30 gal/min.



But when I measured the heat loss for the radiated zone of my house, I got 134,000

BTU/hr which would say that I need 13.4 gal/min for a 20 deg heat loss

(using the Taco Selecting Circulators paper and a 20 deg delta T). This is a big difference, and I was actually hoping to run the pumps even slower to get a larger delta T - say 25 deg delta T.



So is there a need to run such a high GPM through my radiators? After all they ran fine with no pump, using gravity alone. Has anyone ever measured the speed that water runs through gravity pipes when there is no circulator? I would have guessed that it was quite slow - in which case wouldn't it be fine to use a slow speed and extract as much heat from the water as possible before sending it back to the mod-con? Or is this wrong - and is the Taco paper not applicable here?

Steamhead

Size the circ to the radiation

and size the boiler to the heat loss. 30 GPM at 3.5' head could be served by a Taco 0010, Grundfos UP-26-64F or B&G NRF-33, these would all provide a gentle circulation not unlike when the system ran on gravity.

JeffGuy

But

Does this mean that a gravity system with no circulators is moving 30 gallons per minute through its pipes?



And does this also mean that the Taco TD10 document "Selecting Circulators" should only be used for new installs and should not be used to figure out which circulator to use when it will be used in a gravity conversion?

Gordy

Volume

 Not answering for Steamhead. Your trying to move a large volume of water with low headloss piping. So you need a circ with a steep pump curve such as the 0010, or the others listed. 



 You want to slowly move the btu train through the system so the btu passengers can get on at the boiler, and off at the radiator.



Gordy

Mike Kusiak_2

Gravity system design

As originally configured, your system would have probably been designed for an average radiator temperature of 170F and a delta T of 20 F. The flow needed to accomplish this would have been achieved by proportioning the piping to insure a flow rate constent with the requirements and the difference in water density, which resulted in the large pipe sizes characteristic of a gravity system.



In your present system, you have an EDR of 1188 sq ft. which at the 170F design temperature would result in a heat emission of 150 BTU/sq ft. times 1188 or 178200 BTU/hr. Under those conditions, a flow of 17.9 gal /min would have been required to meet requirements, and if the piping was properly designed that flow rate would have actually resulted by gravity convection.



Your current heat loss is only 134000 BTU/hr, so you are effectively over-radiated by a factor of 1,33. This will allow you to meet design conditions at a lower temperature than 170F, which is a good thing for condensing. If you install a boiler producing 134000 BTU/hr. your delta T will only be 20F if the flow is 13.4 gal/min as you previously mentioned. If you increase the flow to 30 gal/min the delta T will drop proportionally,



I believe Steamhead's charts increase the flow rates by 50% to compensate for problems inherent in 100 year old systems. The piping is no longer smooth inside and probably doesn't allow flow as well as when new. From experience installers found that by increasing the pumping by 50%. better flow was achieved and more even heating resulted. Increasing the flow resulted in a lower delta T, but this was of no consequence before condensing boilers need for low return temperatures.



Why don't you install a multi speed pump, whose maximum output is around 25 or 30 GPM. That way you can then reduce the flow rate at lower speeds to find the best compromise between flow and delta T to insure both maximum condensing and even heating.

Steamhead

In a large enough system

you would certainly move that amount of water. 30 GPM is not that much with all that radiation, and remember that's what's moving thru the boiler. If you divide 30 by the number of rads you have, you might come up with an average flow rate thru each rad of something like1- 2 GPM.



For example, my own system has 550 square feet distributed among ten rads. I have a Grundfos UP-15-42F circ, which will pump 13 GPM at a 3-1/2' head. This works out to an average flow rate of 1.3 GPM at each rad. I noticed that when I tried to use a smaller circ than the 15-42 (specifically the B&G NRF-9F/LW) the system didn't circulate well to all the rads. So it is possible to pump these things too slowly, though over-pumping them is much more common.



Now, some rads are bigger than others and are fed from bigger pipes, so these would naturally have higher flow thru them than the smaller ones. But you get the idea.