Question for the hydrodynamic gurus

[Amail]

I'm trying to measure water flow thru my pool pump system. I've got 2-1/2" SCH 40 pipes running water around, but a flow sensor for pipe that big is pretty spendy. It occurred to me I could break out a smaller parallel path that would fit a smaller, cheaper flow sensor.

My question is, what is the relationship of the flow thru the smaller pipe to the bigger pipe? Would there be water flow? Would the elbows reduce the flow thru the smaller pipe ?

[1990C4S]

What flow sensors are you proposing?

[Amail]

I was thinking something along the lines of this

[cockerpunk]

there will likely be some relationship in flows, but it will not linear, and without a larger sensor to first calibrate it, it will be pretty tough to predict with any reasonable accuracy.

[BK911]

Do you have pressure gauges before and after the pump?
Most pumps have a pump curve that you can determine flow by the dp across the pump.

[1990C4S]

Originally Posted by Amail

I was thinking something along the lines of this

That sensor will generate pulses from a paddlewheel? Then what?

[aschen]

I predict a highly nonlinear relationship between the two flows as well.

I think measuring differential pressure across a restriction (or from the pump as BK suggests) would be preferable.

Depending on your accuracy requirements you need some sort of calibration method for anything resembling quantitative results IMO, can you do a time to fill calculation or similar


Surely others have tried similar things so the all knowing google probably has a solution

[John Rogers]

I would suggest measuring how much power the motor driving the pump is using and that will equate to how much water is being pumped. Having the small line off the side of the main line will never give a reading relating to the actual flow. You would need some restriction in the main line to force the water through the small line and then it would not match actual flow.

I would suggest installing a tee and short section of the 2 1/2 pipe and a valve that opens with no restriction down stream of the filters, etc. A valve in the main line would also be needed so you could shut off main flow and open the test line to see how long it takes to fill a 5 gallon container. This would be pretty close to actual flow rate.

[sammyg2]

Here's how I would do it:
Forget the small pipe. Build as flange into the larger pipe and install an orifice plate in that flange, small enough to restrict the flow slightly and cause a pressure drop across it.

then drill and tap the pipe on either side of the orifice and plumb up a delta P gauge (differential pressure gauge). Use the gauge to measure the pressure drop across the orifice, which should be reasonably linear until it starts cavitating.

You may have to play with the size of the orifice and range of the gauge until you can get both to work together, and you'll have to do some calcumalations to determine calibration, but it will be reliable, cheap, accurate, and kind of fun to do.

[sammyg2]

Or you could go the easy route and get a copy of the centrifugal pump curve, and measure the amps the motor is pulling.
Convert the amps to HP, and along with the delta P across the pump (discharge minus suction pressure) converted to head feet and voila!
Plot it on the curve and you have your flow.
After several measurements you could use as a graph to indicate flow at various discharge pressures, assuming the suction pressure is constant.

[BK911]

I didn't recommend the orifice because it will reduce flow and a 2.5" one won't be cheap.
Couple of taps on either side of the pump, Couple of ball valves and a pressure gauge and BAM! Done.
Deadhead the pump and read pressure.
That will put you on the right pump curve.
Then use dP and that will give you flow.
Couple of bucks and less than a 6 pack in labor.

[BK911]

I didn't recommend the orifice because it will reduce flow and a 2.5" one won't be cheap.
Couple of taps on either side of the pump, Couple of ball valves and a pressure gauge and BAM! Done.
Deadhead the pump and read pressure.
That will put you on the right pump curve.
Then use dP and that will give you flow.
Couple of bucks and less than a 6 pack in labor.

[Hugh R]

get some more pipe and elbows and time how long it takes to fill a container of a known volume like your trash can you drag to the curb once a week.

If you have a DE filter with a backwash valve, get a slack hose from the pool supply place and put it on backwash and fill the large container as you time it. That way, you account for DE filter pressure and flow drop.

[Cajundaddy]

Flow varies a great deal over a seasonal filter cycle. If accuracy is important, plumb this in:

https://www.poolsupplyunlimited.com/h2flow-fv-c-flowvis%EF%BF%BD%EF%BF%BD-flow-meter-check-valve-with-valve-body/142242p1

[Evans, Marv]

I was going to suggest something along the lines of what Hugh said. If you are OK with plumbing in an alternate path to measure flow through a cheaper & smaller type of meter, you can put in a valve on your main line and put in a T before it with another valve for a shut off. Then just close the main flow, open the alternate flow, and do a timed measurement into a container. Cheaper & easier than more valves & plumbing, plus it might serve some purpose later on.

[Cajundaddy]

Timing a 50 gallon drum fill is certainly the cheapest way to go but accuracy will be about +/- 25%.
Remember that by diverting flow from the filter you have eliminated all the head loss between filter and pool return, including heater, return valves, check valves, and any 90 ells. Typically greater than 1/2 of your head loss is in the return side.

If a really rough estimate is all you need this will get you there.

Calculating head loss:
http://blog.poolproducts.com/swimming-pool-feet-of-head-calculations/

[Amail]

Thank you all for your replies. I’ve been out of the office for a few hours, so just getting back to this now.

Originally Posted by cockerpunk

there will likely be some relationship in flows, but it will not linear, and without a larger sensor to first calibrate it, it will be pretty tough to predict with any reasonable accuracy.

I wasn’t sure what the relationship would be, or indeed if it would flow at all. I figured if there were a straightforward way to calculate the flow in the larger pipe based on flow in the smaller pipe, then we’d be close, but that assumes a straightforward calculation.

Originally Posted by BK911

Do you have pressure gauges before and after the pump? Most pumps have a pump curve that you can determine flow by the dp across the pump.

I have a pressure gauge after the pump, but before the filter. I use that to give an indication when it’s time to backflush. In order for the pressure to be useful, I have to see what my flow is. For instance, I’ve noticed times when the pressure is quite low. This usually indicates a full secondary strainer or crud in the impeller. Having a flow reading along with the pressure reading would tell me if I have a restriction before (or in) the pump, or if the filter needs cleaning.

Originally Posted by 1990C4S

That sensor will generate pulses from a paddlewheel? Then what?

Well, the assumption is you can deduce the water velocity from the pulse rate. Given the pipe ID you should be able to convert that to flow in GPM.

Originally Posted by aschen

I predict a highly nonlinear relationship between the two flows as well.
I think measuring differential pressure across a restriction (or from the pump as BK suggests) would be preferable.
Depending on your accuracy requirements you need some sort of calibration method for anything resembling quantitative results IMO, can you do a time to fill calculation or similar
Surely others have tried similar things so the all knowing google probably has a solution

I’ve done a cursory Google search but figured the PPOT Brain Trust would like to sink their teeth into this one. I don’t see what putting a pressure gauge before the pump will accomplish. It’s a straight run from the pool to the secondary strainer, then to the pump. If anything, I’d measure suction on the pool side of the pump

Originally Posted by john rogers

I would suggest measuring how much power the motor driving the pump is using and that will equate to how much water is being pumped. Having the small line off the side of the main line will never give a reading relating to the actual flow. You would need some restriction in the main line to force the water through the small line and then it would not match actual flow.
I would suggest installing a tee and short section of the 2 1/2 pipe and a valve that opens with no restriction down stream of the filters, etc. A valve in the main line would also be needed so you could shut off main flow and open the test line to see how long it takes to fill a 5 gallon container. This would be pretty close to actual flow rate.

Power consumption is one option, but I think that would assume a clean secondary strainer and pump impeller (not always the case).

Originally Posted by sammyg2

Here's how I would do it:
Forget the small pipe. Build as flange into the larger pipe and install an orifice plate in that flange, small enough to restrict the flow slightly and cause a pressure drop across it.
then drill and tap the pipe on either side of the orifice and plumb up a delta P gauge (differential pressure gauge). Use the gauge to measure the pressure drop across the orifice, which should be reasonably linear until it starts cavitating. You may have to play with the size of the orifice and range of the gauge until you can get both to work together, and you'll have to do some calcumalations to determine calibration, but it will be reliable, cheap, accurate, and kind of fun to do.

That sounds like a good option.

Originally Posted by sammyg2

Or you could go the easy route and get a copy of the centrifugal pump curve, and measure the amps the motor is pulling.
Convert the amps to HP, and along with the delta P across the pump (discharge minus suction pressure) converted to head feet and voila!
Plot it on the curve and you have your flow.
After several measurements you could use as a graph to indicate flow at various discharge pressures, assuming the suction pressure is constant.

Wouldn’t this assume debris free impeller and strainer?

Originally Posted by BK911

I didn't recommend the orifice because it will reduce flow and a 2.5" one won't be cheap.
Couple of taps on either side of the pump, Couple of ball valves and a pressure gauge and BAM! Done.
Deadhead the pump and read pressure. That will put you on the right pump curve. Then use dP and that will give you flow. Couple of bucks and less than a 6 pack in labor.

Again – doesn’t this all mean a clean impeller and unrestricted inlet?

Originally Posted by Hugh R

get some more pipe and elbows and time how long it takes to fill a container of a known volume like your trash can you drag to the curb once a week.
If you have a DE filter with a backwash valve, get a slack hose from the pool supply place and put it on backwash and fill the large container as you time it. That way, you account for DE filter pressure and flow drop.

The hope is to build an alarm to indicate when the secondary strainer is full (or some other supply side obstruction), the impeller needs cleaning, or the filter needs cleaning. I think your suggestion would work if the system is always in the same condition. It’s also needed to tell a chlorine generator when it’s okay to apply current to the generating cell.

Originally Posted by Cajundaddy

Flow varies a great deal over a seasonal filter cycle. If accuracy is important, plumb this in:
https://www.poolsupplyunlimited.com/h2flow-fv-c-flowvis%EF%BF%BD%EF%BF%BD-flow-meter-check-valve-with-valve-body/142242p1

I looked at that, but I need to build this in to my automation system. I’d need to instrument that device, but looks like it would be a good candidate.

Originally Posted by Cajundaddy

Timing a 50 gallon drum fill is certainly the cheapest way to go but accuracy will be about +/- 25%.
Remember that by diverting flow from the filter you have eliminated all the head loss between filter and pool return, including heater, return valves, check valves, and any 90 ells. Typically greater than 1/2 of your head loss is in the return side.
If a really rough estimate is all you need this will get you there.
Calculating head loss:
Swimming Pool Feet of Head Calculations | SPP Inground Pool Kit Blog

Cajundaddy, thanks - that's a really informative link! At some point I want to switch over to a variable speed pump; looks like this will help pick out the right size.

Again, thanks for your suggestions. If I’m wrong about the impeller and secondary strainer condition affecting flow rate calculations, please tell me.

This is a fun exercise, and I’m looking forward to coming up with a good solution!

[Arizona_928]

Originally Posted by BK911

I didn't recommend the orifice because it will reduce flow and a 2.5" one won't be cheap.
Couple of taps on either side of the pump, Couple of ball valves and a pressure gauge and BAM! Done.
Deadhead the pump and read pressure.
That will put you on the right pump curve.
Then use dP and that will give you flow.
Couple of bucks and less than a 6 pack in labor.

visual

[Amail]

Originally Posted by AZ_porschekid

visual

Thanks for the video. In there he shows pressure sensors on both the suction and discharge side of the pump. If you close off the pipe on the discharge side, I see how you get pressure (thanks, pump!). How do you get any pressure on the inlet side though? For instance, my pool water level is about a foot above the inlet of the pump, so I could see measuring a foot of head on the inlet side. But my friend's pool has the pump higher than the water level, so wouldn't it measure suction?

[Arizona_928]

Originally Posted by Amail

Thanks for the video. In there he shows pressure sensors on both the suction and discharge side of the pump. If you close off the pipe on the discharge side, I see how you get pressure (thanks, pump!). How do you get any pressure on the inlet side though? For instance, my pool water level is about a foot above the inlet of the pump, so I could see measuring a foot of head on the inlet side. But my friend's pool has the pump higher than the water level, so wouldn't it measure suction?

It's the change pressure. It doesn't matter.