Head loss

rhl

The Taco Pump Sizing guide has a head loss formula, the text says that if you have a parallel runs then the head loss is computed for each parallel run. It separately discusses taking the maximum over all the parallel runs. Is the total head loss for the system the sum of the losses over these parallel runs? For example in a manifold.
Reference: https://www.taco-hvac.com/uploads/FileLibrary/SelectingCirculators.pdf

Jamie Hall

No. If the various runs are parallel, then the run with the greatest head loss is the maximum head loss the pump would see.

The flows in the various runs add.

At the risk of confusing you, if you have multiple parallel runs open at the same time, the flow will divide between the runs such that the head loss in each run is the same.

Voyager

rhl said:

The Taco Pump Sizing guide has a head loss formula, the text says that if you have a parallel runs then the head loss is computed for each parallel run. It separately discusses taking the maximum over all the parallel runs. Is the total head loss for the system the sum of the losses over these parallel runs? For example in a manifold.

Reference: https://www.taco-hvac.com/uploads/FileLibrary/SelectingCirculators.pdf

No, it is just what the Taco guide says: “In a system with multiple zones, the head loss needs to be calculated for each parallel loop. The loop with the highest head loss will be used in figuring the total equivalent length in Example 4.”

The loop with the highest head establishes the design head for the distribution system. The total flow, however, is the sum of the flows of each loop.

Brewbeer

Also, the head loss values will change in a system with zone valves, depending on which and how many zones are actively calling at the same time.

EBEBRATT-Ed

The reason they tell you to calculate all parallel runs is so you can find the head loss of the highest loop.
and see which zone has more resistance.

only the loop with the highest resistance is considered for head loss

SuperJ

You take the highest head requirement, but have to sum the flows to select the appropriate pump. Since the differential pressure a pump can create varies depending on flow rate. You may need balancing devices to keep the lower head loss parallel loops from "stealing" too much flow from the higher head loop.

nibs

I understand that for practical purposes, two parallel loops can be considered the same as one loop, but something tells my gut that there must be a small difference. Would like to know if the flow rate co-efficiency of friction in pipes is a straight line.
Spent some time last year working on head loss so am curious.

Jamie Hall

Yes, two or more parallel pipes can be treated as a single, larger pipe for some types of calculations.

The relationship of head loss to flow in a pipe is not a straight line. It's a power curve; the exponent is commonly taken as 1.83

hot_rod

This journal has a good explanation on pump sizing and circuit calculations.
https://www.caleffi.com/sites/default/files/coll_attach_file/idronics_16_na_0.pdf

rhl

Thanks guys. FWIW I’m a mathematician not a plumber, and Jamie it’s a polynomial function not a power function. Typically power functions mean exponential, i.e the formula is something like x^2 rather than 2^x, the latter is considered a power law.

Jamie Hall

uh huh. And when I said power function, that's what I meant, But whatever... head loss is proportional to Q raised to the 1.82 power...

Voyager

rhl said:

Thanks guys. FWIW I’m a mathematician not a plumber, and Jamie it’s a polynomial function not a power function. Typically power functions mean exponential, i.e the formula is something like x^2 rather than 2^x, the latter is considered a power law.

As a mathematician, you should have known that Jamie was correct. I prefer exponential function rather than power function, but same thing. I prefer to use power in the physics domain rather than mathematics to avoid ambiguity.

nibs

I remember looking it up in my Marks Handbook and the formulae were longer than three months in jail. Now I have time to ponder this, my Marks is at home.
Thanks @Jamie Hall

rhl

Yes and f(q) = q^1.82 is a polynomial not an exponential in q

Jamie Hall

I stand corrected as to mathematical terminology. A mathematician I am not, and don't pretend to be. I just know how to apply the blessed things...

Gordy

Parallel designed piping systems is the longest loop plus piping back to the boiler headloss for the pump sizing.

Head balances out through out the system. While flow rates will differ.

A little more discription of your particular example will shed more finite light.

rhl

That’s all I wanted to know. Therefore a manifold at the boiler room with parallel runs is likely no worse than running big mains in a direct/reverse return with respect to head loss