How to calculate GPM for this system head loss?

sunlight33

Say there is a pump connected to three parallel heating circuits each with 2' of head, and the pump itself is 4' below the heating circuits which is on the second floor, would I use 6' to calculate the total GPM then divided by 3 to find the flow through each circuit?

Zman

The vertical distance is irrelevant in a closed loop system. Think of a Ferris wheel...
Check out Taco's Flow pro university and watch some of the webinars. I think this will make more sense.

CMadatMe

Basic rule of thumb to calculate ft of head and its very conservative. Head is not communalitive. So if you are zone with zone valves you only need to calculate the longest run. If a zone has an air handler or coil you must add that pressure drop to the ft of head of the run. 1PSI = 2.31' of Head

Longest run * 1.5 * 04 = Ft of head

Example 100 * 1.5 * .04 = 6' of head

Longest run - From the outlet side of the circulator thru the heating system back to the inlet side.

To calculate GPM - GPM = Btu/hr / (Delta T * 500)

Example: 100,000 / (20 * 500) = 10 GPM

hot_rod

Here is a good in depth looks at pump sizing and choices.

http://www.caleffi.com/sites/default/files/coll_attach_file/idronics_16_na_0.pdf

Jamie Hall

And furthermore -- if the heating circuits are in parallel, the head doesn't add (the flow does). So your head loss is just 2 feet. The pump curve will give you the gallons per minute for that head loss (which, incidentally, sounds a bit low...). That flow will divide between the three circuits so that the head loss in each circuit is the same (two feet?). If the circuits are identical, then the flows will be identical, and each one third of the total flow.

MikeSpeed6030

There are two ways to calculate flow in a complex piping system. One requires an advanced degree in mechanical engineering. The usual way is to size zone/loop pipes for the heat load, and to install a common circulator pump, such as a Taco 007, 95% of the time. If you want to get fancy, install manual balancing valves in each loop.

sunlight33

If all three zones are baseboards without TRV, is there any benefit using a Grundfos Alpha?

Gordy

Jamie Hall said:

And furthermore -- if the heating circuits are in parallel, the head doesn't add (the flow does). So your head loss is just 2 feet. The pump curve will give you the gallons per minute for that head loss (which, incidentally, sounds a bit low...). That flow will divide between the three circuits so that the head loss in each circuit is the same (two feet?). If the circuits are identical, then the flows will be identical, and each one third of the total flow.

From what I'm picking up from the op's post is confusion on what head is in a hydronic system. The OP is using vertical elevation distance in calculating head loss.

Head in a hydronic system is the friction of the water against the walls of the piping, and restrictive components. Valves, boiler passages, Mixing valves, tees, 90's ect. Not vertical elevation.

The pressure in the system is what gets the water to the highest point in the system. All the circulator has to do to move the water is create a large enough pressure differential to overcome the restrictive components in the system.


Chris's formula is a short hand method that gets you close.

Gordy

MikeSpeed6030 said:

There are two ways to calculate flow in a complex piping system. One requires an advanced degree in mechanical engineering. The usual way is to size zone/loop pipes for the heat load, and to install a common circulator pump, such as a Taco 007, 95% of the time. If you want to get fancy, install manual balancing valves in each loop.

Really? So you design all your heating systems around a 007 circulator 95% of the time.

sunlight33

According to SlantFin specs, heat output at 4 GPM is only about 5-7% higher than that at 1 GPM, so if I know the head loss of a zone I can set the speed of 15-58 to lower speed to save electricity and not suffer much in heat output, right?

Gordy

How do you know what gpm your getting? To directly answer yes you want to run the slowest speed that will give the needed output to the load. Will 5-7% be enough to not meet the load don't know with out a heatloss.

Rich_49

@MikeSpeed6030 .

It takes a fifth grade education to calculate head loss in a circuit to a gnats **** . A second grader can use the formula Chris posted .
Inherently mechanical engineers suck by the way .

bob_46

Three parallel circuits each with 2' of head the total head will always be less than the smallest in this case .2924 feet total.
I used "Siggy's" equation 6.20.
Sorry Rich I had to use a scientific calculator.

CMadatMe

sunlight33 said:

If all three zones are baseboards without TRV, is there any benefit using a Grundfos Alpha?

Yes. The pump will react in Auto Adapt mode to the changing pressures in the system as zones open and close and will provide electrical savings. Only issue I've ever had with Alpha's is when running in Auto Adapt with Air Handlers.

Jamie Hall

Rich said:

@MikeSpeed6030 .

It takes a fifth grade education to calculate head loss in a circuit to a gnats **** . A second grader can use the formula Chris posted .
Inherently mechanical engineers suck by the way .

Oh really?

CMadatMe

Jamie Hall said:

Rich said:

@MikeSpeed6030 .

It takes a fifth grade education to calculate head loss in a circuit to a gnats **** . A second grader can use the formula Chris posted .
Inherently mechanical engineers suck by the way .

Oh really?

Unless its common core math, then we have a problem Houston.

ced48

Seems like nobody has answered sunlight33's original question, does he calculate based on 2, or 6 pounds of head? This might be a simple answer for most of you guys, but it can be confusing for some of us with limited experience.

CMadatMe

ced48 said:

Seems like nobody has answered sunlight33's original question, does he calculate based on 2, or 6 pounds of head? This might be a simple answer for most of you guys, but it can be confusing for some of us with limited experience.

Think numerous guys gave him the answer. Even gave him the basic rule of thumb formulas and some explained that you don't count vertical distance as head.

Paul48

Mechanical engineers don't suck. The problem, is, that bad ones have the same authority, as good ones, to the unknowing. I just read, that in a poll of recent college graduates....ten percent thought that Judge Judy sits on the U.S. Supreme Court. Some of them may have been engineers. You need a good memory to get a college degree. You don't have to have any common sense.

Zman

ced48 said:

Seems like nobody has answered sunlight33's original question, does he calculate based on 2, or 6 pounds of head? This might be a simple answer for most of you guys, but it can be confusing for some of us with limited experience.

I couldn't figure out how he came up with 6 feet. Was it 3 loops at 2 feet or 2 feet for the loop and 4 feet for the elevation? Either is incorrect for different reasons.
I figured he would be better off doing some reading or watching a webinar since he seemed to be needing to get familiar with the basics.
The quick formulas folks have posted will work fine in a system with similar loop lengths. When the lengths vary, it would be better to use software and/or balancing valves.

ced48

I think his example has 3 circuits of equal head pressure all running in unison. Is the head equal to that of 1 circuit, or the total of all 3? Seems simple, but it can be confusing to some.

CMadatMe

ced48 said:

I think his example has 3 circuits of equal head pressure all running in unison. Is the head equal to that of 1 circuit, or the total of all 3? Seems simple, but it can be confusing to some.

We've answered that. Head is not commutative. You don't add them all up when having a single pump providing flow to multiple zones. Only the largest head counts. If the pump can overcome the largest it can over come all.

Jamie Hall

Paul48 said:

Mechanical engineers don't suck. The problem, is, that bad ones have the same authority, as good ones, to the unknowing. I just read, that in a poll of recent college graduates....ten percent thought that Judge Judy sits on the U.S. Supreme Court. Some of them may have been engineers. You need a good memory to get a college degree. You don't have to have any common sense.

You are so right...

Years ago we (I'm a Civil and Agricultural) had to have 12 years (twelve years!) of increasingly responsible project work under the authority of a senior registered person to even sit for the exam, never mind pass it. Nowadays? What is required is a degree, not experience in the field. Don't get me going...

rick in Alaska

MY uncle always liked the term "doctor at practice". Who wants a doctor who is practicing.
We could call ourselves "practicing heating professionals".
Sounds good anyway.
Rick

Gordy

Read the original post. It could be interpreted two ways. When someone states an elevation distance as though it reflects on head loss in a circuit. One has to question the manner in which the circuit head loss was calculated.

Brewbeer

For a fixed piping system, the total head of the system is a function of the flow rate of the fluid through the system. This is illustrated by the following equation, which comes from Taco's Technical Document 10 (TD10):

HL = k x c x L x (f^1.75)

Where:

HL = the head of the piping system (feet)
k = is a number based on tubing diameter
c = is a number based on fluid viscosity (which varies with temperature and glycol content)
L = total equivalent length of piping circuit
f = flow rate through piping in GPM

As an illustrating example, for a fixed piping system that uses 3/4 inch diameter copper pipe, water at 140F, and a system with an equivalent length of 100 feet:

Flow rate 1 GPM = Head loss 0.3 feet
Flow rate 2 GPM = Head loss 1.0 feet
Flow rate 4 GPM = Head loss 3.3 feet
Flow rate 8 GPM = Head loss 11 feet

Here is where the complexity starts to build: if you provide a fixed flow rate of 2 GPM to the described piping loop, the head loss is 1 foot. However, if you provide a fixed flow rate of 2 GPM to 2 identical parallel piping loops, the flow rate through each loop is 1 GPM. Since the head loss at 1 GPM is 0.3 feet, and individual head losses of parallel loops are not additive, the head loss of the system of two identical parallel loops at a total of 2 GPM (1 GPM per loop) is 0.3 feet. So far, the math is fairly straight forward.

Now to make it even more complicated, the relationship between flow rate and head loss for a given pump at a fixed speed is not linear. The Taco 007 will provide about 23 GPM (in theory) at 0.3 feet of head, and about 22 GPM at 1 foot of head. However, the "in theory" part is that if you try to flow 23 GPM through the example pipe system above, the head loss of that system at 23 GPM is 71 feet. So when ends up happening, is flow and head loss comes into an equilibrium condition. The math needed to calculate the equilibrium condition is getting complicated.

Now to make it even more complicated, if you change the example system (100 feet of 3/4 inch copper pipe) by adding additional loops in parallel having different lengths and sizes (and as a result different head losses), the flow rate of single pump multi-loop systems changes whenever a loop is added or dropped by its controlling zone valve. At this point, the math to calculate the numbers to "within a gnats ****" is highly involved, and this is where most people will defer to the mechanical engineer (assuming the ME knows what he is doing).

But in reality, for a typical home system which is almost certainly over-designed (probably by accident), it probably doesn't really matter.

ced48

Chris said:

ced48 said:

I think his example has 3 circuits of equal head pressure all running in unison. Is the head equal to that of 1 circuit, or the total of all 3? Seems simple, but it can be confusing to some.

We've answered that. Head is not commutative. You don't add them all up when having a single pump providing flow to multiple zones. Only the largest head counts. If the pump can overcome the largest it can over come all.

So, if I had 2000 circuits I would use the same pump?

CMadatMe

ced48 said:

Chris said:

ced48 said:

I think his example has 3 circuits of equal head pressure all running in unison. Is the head equal to that of 1 circuit, or the total of all 3? Seems simple, but it can be confusing to some.

We've answered that. Head is not commutative. You don't add them all up when having a single pump providing flow to multiple zones. Only the largest head counts. If the pump can overcome the largest it can over come all.

So, if I had 2000 circuits I would use the same pump?

That would all depend on the flow rate needed thru the 2,000 loops. 2' ft of head, is 2'ft of head.

Jamie Hall

ced48 said:

Chris said:

ced48 said:

I think his example has 3 circuits of equal head pressure all running in unison. Is the head equal to that of 1 circuit, or the total of all 3? Seems simple, but it can be confusing to some.

We've answered that. Head is not commutative. You don't add them all up when having a single pump providing flow to multiple zones. Only the largest head counts. If the pump can overcome the largest it can over come all.

So, if I had 2000 circuits I would use the same pump?

No, because if you had 2,000 circuits you would have -- at the calculated head loss for a single circuit -- 2,000 times the flow rate, in gallons per minute. Somehow, I doubt that the pump could handle that... there is a volumetric volume flow at a head loss of zero for every pump.

As to the calculations... oh yeah. Bad enough where you just have a few pipes in series or parallel. Through in two pumps in series or parallel and it gets messier. And where it gets really interesting is when you have multiple loops cross connecting in a system. I used to do fire flow estimation (and trust me, it was just estimation!) for municipal water systems. In the days before computers, one could use up a lot of paper...

SWEI

When you parallel circuits that have different head losses, the math does get a bit more complex thanks to a somewhat inconvenient fractional exponent. The longest loop method doesn't factor this in, but it gets you close enough in most cases.

sunlight33

If my calculation is correct, 15-58 at speed 2 will provide 4 GPM through each of the three zones, while speed 1 will provide 2.5 GPM. What would you estimate the heat output difference from baseboards to be from 4 GPM to 2.5 GPM? 2-3%?

My initial question has been answered, for some reason I was thinking about the typical pump that has to overcome air pressure, without realizing it's a closed loop system.

Zman

Well now you just started another one...

It really depends on how many feet of baseboard you have. The temperature drop from radiator to radiator goes up as the flow rate goes down. This means the the downstream radiators produce incrementally less heat. For most systems 2.5 GPM would work just fine.

It is not hard to calc the average water temp at each radiator at different flow rates using the universal hydronic formula.

Details?

sunlight33

For the sake of simplicity I measured the return temperature of the system loop and it was about 20-25 degrees lower than the supply initially when the boiler is on, after a couple of minutes the difference starts to become less and less, and it drops to about 10 degrees right around when the boiler turns off. So looks like it's all good.

Paul48

I'm not gonna get involved in a dispute. Look at the rated outputs supplied by the fin-tube manufacturers. The info defies the Universal Hydronic Formula, and it has been debated(hotly) here.

Zman

sunlight33 said:

For the sake of simplicity I measured the return temperature of the system loop and it was about 20-25 degrees lower than the supply initially when the boiler is on, after a couple of minutes the difference starts to become less and less, and it drops to about 10 degrees right around when the boiler turns off. So looks like it's all good.

Sound like it will work perfectly at a delta of 10 degrees.

Zman

Paul48 said:

I'm not gonna get involved in a dispute. Look at the rated outputs supplied by the fin-tube manufacturers. The info defies the Universal Hydronic Formula, and it has been debated(hotly) here.

The manufactures charts are for 1 heater at a specific AWT. They don't account for the AWT temp drop when heaters are piped in series.
There really is nothing to debate. If you have one 4' heater the charts are perfect.. If you tie 20-4' heaters together, the results will be quite different. The charts will be accurate for the specific heater at it's new adjusted water temp. That water temp will naturally be reduced at a different rate depending on system flow rates.

hot_rod

Here are some actual examples of how the math and reality match up.

As you see output in the fin tube example drops considerably at low, below 1 gpm flows, it is definable and measurable. That is where you need to be aware of what a fixed delta system will run up against.

I wouldn't agree out put is essentially identical at 1 vs 4 gpm. the question becomes is it worth providing the extra flow rate, for the extra BTU output. Assuming you stay below a 4fps velocity for noise concerns.

In your example if the difference is a one step speed change, no question the higher speed will provide additional flow and some additional heat output.

In the radiant loop example that flow change amounts to a 12% difference in output, if an additional 20 W of pumping power is required to accomplish that it may be worth it. As well as additional comfort from a tighter delta T in a radiant slab.
included is the simulation that provides that output difference.

Zman

Hatterasguy said:

The difficulty that most have with the baseboard outputs is that they have nothing to do with the hydronics forumula. Their output is based upon the temperature of the tube and that temperature doesn't vary much with flow rate until the flow rate drops below 1 GPM.

So, defying all logic, the output is essentially identical at a flow rate of 4 GPM as compared to 1 GPM. Most folks have a lot of trouble with this and the RWT doesn't change much with flow rate changes. The ΔT of the loop is based upon the length, not on the flow rate (essentially).

Many times we hear the suggestion to obtain a ΔT of 20F between SWT and RWT. However, on fin tube, one can drop the flow rate to an insignificant amount, and well below the capability of the baseboard, in a valiant effort to obtain that magic 20F ΔT. In reality, it's not struggle worth pursuing and the output of the baseboard drops off a cliff at flow rates below 1 GPM.

If no formula is involved, how would one determine the temp drop from radiator to radiator?
I respectfully disagree....

Paul48

Sorry for the confusion Carl..... My post was not related to your facts, and that was not the debated subject I was referring to. It was more in line with what Hat and hot rod are posting.

hot_rod

Hatterasguy said:

hot rod said:

I wouldn't agree out put is essentially identical at 1 vs 4 gpm. the question becomes is it worth providing the extra flow rate, for the extra BTU output. Assuming you stay below a 4fps velocity for noise concerns.

The difference in output is 6% using Slant Fin 30's output data.

No contractor or H/O could ever measure or observe this difference in the field.

Therefore, it is essentially identical.

But 6% could get a homeowner out of the weeds on an underperforming system, at or below design conditions. Or an underperforming loop or zone.

It shouldn't be dismissed, instead the knowledge to know how and where it comes from should be clear.

Manufacturers include various flow rates in their design guides for that purpose.

MikeSpeed6030

Gordy said:

MikeSpeed6030 said:

There are two ways to calculate flow in a complex piping system. One requires an advanced degree in mechanical engineering. The usual way is to size zone/loop pipes for the heat load, and to install a common circulator pump, such as a Taco 007, 95% of the time. If you want to get fancy, install manual balancing valves in each loop.

Really? So you design all your heating systems around a 007 circulator 95% of the time.

Many hot-water boilers are supplied with a circulator, typically a Taco 007. Weil-McLain is an example.