Fluid velocities for system design

SWEI

Quoted from another (somewhat agonizing) thread:

On the high end the excessive velocity point is defined by the FPS velocity. Exceed the hydronic guideline of 4 maybe 5 fps and you get to potential noise and excessive wear.
...
DHW piping sometimes allows 6 maybe 8 fps, but it is not a 24/7 flow like a hydronic system will experience.

Like most of us, I have been sizing pipe at 4 FPS for years, but does that actually represent current best practice when used with smart variable flow circulators? If a system design reduces flow at less than design conditions, and we have pretty much defined design conditions as 2.5% of the time on average, what flow velocities should we really be sizing to? Throw modern polymer piping materials into the mix and I think we have a case for re-evaluating our assumptions. As an example, Aquatherm suggests

Aquatherm has allowed engineers to design with velocities as high as 15 – 20 ft/sec (4.57 – 6.10 m/sec) depending on the job and design. This allowance comes with a caveat to ensure that there will not be any quick-acting valves or other sources of surge pressures in the system. In other words, it is permissible to design to the higher velocities for the pipe material, but the system may not be able to handle the higher velocities in terms of pressure surges, water hammer or noise issues.

I'm not suggesting that 15 FPS is a good idea (especially given the pumping energy that would require) but my <= 4 FPS systems are mostly running at < 2FPS the overwhelming majority of the time.

I'm interested in your thoughts...

bmwpowere36m3

And I've been struggling to get 2 fps… not the other way round. Whats' the downside of low velocity, air removal and larger pipe size required (cost)?

EBEBRATT-Ed

Yes air removal becomes a problem at low velocity. Can't put vents everywhere and the risk is air trapped that won't come out of the system. However the minimum velocity to entrain air is 1 1/2 fps (according to B & G) But SWEI has a good point. I have been stuck on 20 deg. td and 4fps & 4' hd/100' for so long and it works so I don't think to change it.

hot_rod

Information from the Amtrol Design manual indicated to move an air bubble 1/2" diameter thru a piping circuit requires 1.5- 2 fps. Along the same as the B&G info.

Vertical piping runs are the most critical areas as the power of the bouyancy, based on fluid temperature, can over come the flow velocity allowing the bubble to rise and trap at then high point.

Also for pipe sizes above 2" the velocity is not constrained by that 4 fps rule, but a pressure drop calculation, that does allow much higher velocities.

Does the Aquatherm suggest those velocities in all sizes, or 2" and larger?

This is also showing up in tangential, votex type air separators commonly used on large pipe sizes. If velocity slows to the point of loosing that cyclone effect, air removal suffers. One school of thought is the system be brought up to higher velocities for a period of time to assure air removal, then allow the pumps to ramp down, assuming you rid the system of then problematic air. Or occasionally ramp up the speeds to provide any additional air removal required.

Variable flow systems does present a new batch of questions and potential problems for both air removal and heat transfer.

SWEI

hot rod said:

for pipe sizes above 2" the velocity is not constrained by that 4 fps rule, but a pressure drop calculation, that does allow much higher velocities.

Does the Aquatherm suggest those velocities in all sizes, or 2" and larger?

Variable flow systems does present a new batch of questions and potential problems for both air removal and heat transfer.

That was really the point of this post. Once we remove erosion corrosion from the equation, and factor in operating hours at various percentages of full flow, what actually makes sense?

Paul48

SWEI.....What about components such as valves, made of the same old materials? I ask, because I saw brass ball valves in a picture on Aquatherm's home page.

hot_rod

I would think sizes 1/2 - 1" would be connected to some sort of heat emmiter, BB, panel rad, air coil convector, etc. So that velocity would need to be acceptable for every component or heat emitter in the circuit. That velocity in a 3/4 fin tube would be noticable.
Not sure any zone valve, TRV, balance valve, would be happy closing or regulating against that flow velocity. Water hammer arresters in hydronic piping?

Larger sizes for distribution or primary loops might be doable. Although even micro bubble type air purgers lose efficiency at high velocities.

Paul48

I get this picture in my head of 3 men adjusting a balancing valve.

Gordy

I think Hot Rod, and Paul sum it up best. Consider the other components, and the effect of driving flow rates into the upper edge of the envelope.for that 2.5% of the season.

Paul48

It begs the question.......Do you really want to distribute the heat slower, at anything other than design temps? Aren't you better off at the higher velocity, as far as AWT at the emitters, and even heat throughout the system?

SWEI

Paul48 said:

What about components such as valves, made of the same old materials? I ask, because I saw brass ball valves in a picture on Aquatherm's home page.

I've never used those -- just the PP ball valves in larger sizes. We transition to PEX at 1" and smaller most of the time.

Paul48

I suppose as long as all the components will handle the velocity, that wouldn't be a factor. I don't know at what point it would start to sound like a freight train. Probably anything beyond the old guidelines. Their velocities are a great example of what you can do with their product.They don't say how to best apply it though.

SWEI

My point was really that >4 FPS in materials like PEX and PP is not the issue it is with metallic pipe. Combine that with the fact that a well engineered variable capacity system will spend ~90% of its operating hours running at less than 50% of it's design capacity and I think we have a valid case for designing to a different standard. 6 FPS? 8 FPS? Siggy? Bean?

EBEBRATT-Ed

When you find the cost of those Aquatherm valves you probably won't be buying any! The price of this stuff in the larger sizes 2 1/2" & up is pretty steep. I can see a lot of water hammer in the future raising the fps, close off pressure etc

Paul48

It's almost impossible to get the answers. The ones with the resources to do the testing, the manufacturers, all have an agenda. They will only test in such a way as to promote their product. If their product is a POS, they will spend the money to re-educate you.

bmwpowere36m3

I agree with SWEI's premise, in that at design day if you design around 4+ fps, then the remainder of the time you could be pumping at velocities below 4 fps. That's assuming variable pumping and would need to flow less as heat demand goes down. Just like that whole debate on delta t… I think it depends on specific conditions and it would be interesting to see a scenario with some calculations at various BTU/gpm demands.

At best, and I’m only guessing, it would maybe allow you to run 1 pipe size smaller. So either you end up using smaller pipes to maintain a higher velocity at the expense of head loss or pump higher GPMs and run a lower delta T.

I know with PEX on domestic lines they allow much higher velocities... But even consider a tub faucet capable of supply 5 GPM being supplied with a 3/4" line, that's still under 4 fps.

Mark Eatherton

SWEI said "On the high end the excessive velocity point is defined by the FPS velocity. Exceed the hydronic guideline of 4 maybe 5 fps and you get to potential noise and excessive wear."

I think the key word here is POTENTIAL. I've read engineering reports from way back when from the Copper Development org. when they were testing pipes and flow and determining max. recommended velocities, and it was always a "potential". Never really proven and "heard", only theoretical. Some of their original testing was done on 1/2" copper pipe that was run for long periods of time, with no damaging effects from erosion corrosion, but these were straight runs of pipe. I don't remember the exact numbers, but it was something crazy like 100 GPM in a 1/2" pipe.

In my 39 years of bouncing around mechanical rooms, I've never heard velocity noises as an issue. Now cavitation noises associated with poor piping practices, sure. But straight forward hissing velocity noise has never raised its ugly head.

I think another thing that has changed significantly over the years is air removal. Back in the day, air was kept (as well as they could) in the system, recovered and kept in the compression tanks for allowing expansion and contraction. In todays world of rubber bladdered expansion tanks and micro bubble resorbers, if a person does a good job of purging initially, and has MBR's in their system, and their pumps are all pumping away from the PONPC, the chances of having ANY air noise in a good system are slim to none. There's a big difference in "free air" removal versus "entrained air" removal that needs more research to determine minimum required velocities. We're dealing with old school science here...

I like HR's idea of having a daily "Air purge cycle", except that it will require satisfied zones to be opened temporarily to purge their zones, which may contribute to energy waste.

As for sustaining velocities in the 15 foot range, and this is just me speaking, not the manufacturer, I only see this happening on loops that have no control valves in the main stream flow. Water source heat pumps, cooling towers, GSHP loops etc. Obviously, it WILL require substantial electrical horsepower to maintain those kinds of velocities. Nothing against the manufacturers, but that velocity looks more like something that was cooked up by the sales/marketing department than the engineering department. As Copper head Ken used to say, "Sales and marketing can override engineering any day of the week", to which I add, "Until a lawyer shows up and starts asking questions..."

I say stick with the proven (4 to 6 FPS) and stay on the efficient side of the curve.

Good question tho...

ME

SWEI

EBEBRATT-Ed said:

When you find the cost of those Aquatherm valves you probably won't be buying any!

I thought so too when I got my first set of quotes on them (40mm, 50mm, and 63mm) a few years back. Then I priced out the NPT transition fittings and decided it wasn't so bad. The overall cost of the job still comes out a lot less than copper in sizes above 1" nominal.

The price of this stuff in the larger sizes 2 1/2" & up is pretty steep. I can see a lot of water hammer in the future raising the fps, close off pressure etc

Not steep at all when you compare to large diameter copper. SDR11 Blue Pipe lists at $17.83 a foot in 90mm and 125mm at $33.19 a foot. Normal trade discounts apply (roughly the same as we get on Uponor.) Street prices on Type L copper are currently in the same range as list prices on Aquatherm.

Take a look at something like 3" x 1" and 4" x 1-1/4" tees. Aquatherm list prices for fusion outlets are $8.71 and $10.89 respectively (normal trade discounts apply, about the same as we get on Uponor.) Now look at this and this. Unless you own a tee drill, you're going to lose that bid to the guy doing Aquatherm.
I'm not even sure who makes PEX tees that large these days. When I looked a few years back (about the time the DZ issues came up) Everloc tees in those sizes had list prices north of $200 apiece.

EBEBRATT-Ed

SWEI, I am not against Aquatherm. We did a huge school with it took out 50 something gas fired roof tops and put in a boiler room and the whole job pretty much was Aquatherm 6" and down. One problem is the large # of hangers required, the spacing can be much, much less than your support structure, bar joists etc depending on temperature and pipe size so hanging is problematic along with plastic coated hangers that Aquatherm wants. We have problems getting fittings (I think there's only 1 real distributor in MA.) The other issue is making joints on the big stuff. 2" and down is ok, just hand paddles. For the bigger stuff you have to get that big heavy machine in the air or use the electro fusion couplings which are expensive. It's a good product, tough as nails I really don't think you can break that stuff.

I'm just not a big fan of it because of all the above. Just too old to change I guess.

Ed

hot_rod

I wonder why they don't show pressure drop graphs or tables to go with those velocities

SWEI

hot rod said:

I wonder why they don't show pressure drop graphs or tables to go with those velocities

They do (with quite a bit of detail) starting on the page after the one the snippets above came from. Page 3.10 in the 2014 catalog.

Support requirements for Blue Pipe and Fusiolen Greenpipe are more than copper but a lot less than for PEX. The SDR11 pipe flows a LOT more than SDR9 PEX does. I don't generally use it for near-boiler piping or at sizes below 1" nominal, but for many applications it is our go-to choice.

This was not actually intended to be a discussion of pros and cons of various piping systems, but so be it.

Gordy

Kurt,

A case of unintended consequences. So either you keep the min. FPS at 2, or do as Hot Rod suggested , and let the FPS drop in your design, and run a scheduled air maintenance program. But I agree with the on set of VS pumping things have changed.... Multi faceted.

Paul48

I'd think that down-sizing pipe size to increase velocity would just lead to a game of tail chasing, when paired with DT circulation.

SWEI

I'm not proposing we design to some near-limit condition of 10+ FPS and buy a bunch of big pumps. Rather, that we consider sizing to keep the range of expected flows such that as many hours as possible are flowing at 2-4 FPS. When we size at 4 FPS maximum, we frequently end up with a pipe size that results in something like 3.2 FPS at design conditions. The result is that during most of the hours of operation, the pipe actually ends up flowing < 2 FPS. If we sized at something like 5 FPS or 6 FPS (or maybe even 7 FPS), we might actually end up with better overall performance from the system. Obviously, overall friction loss and pumping energy would need to be considered here.