How to calculate flow with valve position and pressure differential

SCADA_Dude

Member
Join Date
Feb 2008
Location
California
Posts
115
Is it possible to accurately calculate flow rate through a valve using valve position and pressure differential across the valve?

It must be possible either using modeling or some advanced formula, I am wondering if anyone has got this to work reliably. In my application, I would be calculating water flow (obviously viscosity would be a factor in this calculation)

Thanks!

I found this on google for a blackbox solution (ClaVal X133):

Q = Cv * sqrt(DP)

Where
Q = flow (gpm)
Cv = returned Cv at the current valve position
DP = difference between upstream and downstream pressure (psi)

An example of using the function for a 6 inch 100-01
valve with
Upstream pressure = 100 psi
Downstream pressure = 50 psi
Valve position = 30 %

Q = cv_6(30) * sqrt(100-50) = 82.6 * sqrt (50) = 584 gpm
 
Last edited:
Is it possible to accurately calculate flow rate through a valve using valve position and pressure differential across the valve?

It must be possible either using modeling or some advanced formula, I am wondering if anyone has got this to work reliably. In my application, I would be calculating water flow (obviously viscosity would be a factor in this calculation)

Thanks!

I am sure it is possible, you would just consider the valve to be the orifice plate. You would just need to calculate the exact cross sectional area of the open path through the valve for a varying valve position.
Achievingany accuracy might be optimistic though.
 
Do I think this can be done? ... Yes
Will it be accurate? ... No

I think, if there was an accurate way of doing this, it would have already been done. Over the years so many ways of measuring flow using Differential Pressure have been developed: Orifice, Venturi, Pitot Tube, Annubar, etc. All of these have highly defined fixed features to create a DP related to flow. Introducing another variable into the mix, IMO, would would just add to the difficulty of the calculation and the inaccuracy.
 
It can be done, just like HLeap said, but expect accuracy to go out the window.

The problem is that the physical characteristics of the valve doesn't match that of an orifice plate. The valve will generate much more turbulence, which will affect the accuracy. As to how much, I can't immediately say.
 
I know your question was about how to "calculate" flow, but if what you really want to do is "determine" flow, you may have an option.

As stated by others "calculating" flow through an orifice is a science. The techniques used in this endeavor INVARIABLY begin with a known, fixed, orifice with extremely low turbulence-generating characteristics. Your VALVE, no matter WHAT kind it is, will not be suitable to create a pressure drop WITHOUT creating turbulence.

In addition, your ability to precisely and accurately calculate the exact cross section of the opening of the valve, based on the handle position or how-many-turns from closed or open, will be next-to-impossible with a butterfly valve. It MAY be possible with a gate valve, but its not going to be as easy as you may think.

Yet another factor is that while you're squeezing down the "flow" to generate a pressure drop to determine "flow" you are affecting "flow". I find that to be bad science.

If you are trying to use a valve to "determine" flow, you COULD rent or borrow an ultrasonic flow meter and MEASURE and record the flow at different valve settings. Then, all other variables being equal, returning to a particular valve setting should closely replicate the previously measured flow for that valve setting.

I would recommend a cheap, saddle-on flow meter instead (around 400 bucks). The time you'll save and the reliable info you'll get over it's lifetime, will more than offset the cost.

The cheapest one I know of can be found at setpointautomation.com

Stationmaster
 
Have you looked for manufacturer's data on the valve?

The Cv(position) term data should be provided by the manufacturer as a formula or look up table. If this is an expensive valve the manufacturer should provide specific data for each valve by serial number.

The manufacturer should supply this data. If they can't then it shows they have no qualify control and their design is a kludge.
 
This should give you an approximate flow rate if you are measuring water flow. However, since this is a 6 inch valve I am going to assume this is a butterfly valve. The Cv is very non-linear, and isn't neccessarily a precise value as various valve trim conditions will affect it. Additional error will be induced by the position feed back from the valve operator. In my experience these are very bad and subject to drift and calibration error. You will get additional error from from the DP transmitter. The result is going to be a very non-linear measurement and particularly innacurate as the valve approaches the closed position.

If you are measuing anything other than water you will have to correct for specific gravity, and if you are measuring a gas additional corrections are required.

If you are satisfied with a "more than before" or "less than before" kind of measurement it might be sufficient.

By the time all is done you would be better off with an insertion type averaging Pitot tube flow meter like an annubar. These are still subject to innacuracies if your pipe configuration isn't good (upstream elbows etc.). They will give you more accuate measurement than the valve. Another more expensive choice is a flow tube, similar to a venturi. Primary Flow Signal and BIF offer them, and they can correct for known pipe configurations and give quite accurate measurements.

You get what you pay for. If you go cheap with the valve and DP measurement you will probably be unhappy with the results.
 
I did exactly the same thing at a water district a few years back. We got the valve curves from CLA with Cv information. We then plugged these into a 2-D curve fit program and a formula representing flow came out on the other end.

Put these calculations in a PLC can be quite tricky but not difficult. I made a separate & long ladder subroutine to to do the flow calc. The results were very comparable to other CLA-VALVE blackboxes. I didn't think the accuracy was that bad.
 

Similar Topics

Dear Experts,:) I have E&H flow meter (promag L400 ,5L4C4F, DN450) which has pulse o/p with these specs: Pulse width Adjustable:0.05 to 2000 ms...
Replies
4
Views
2,076
Hello everyone, i know there are some posts about this topic but none convinced me. I have an input pulse that comes for every m3 of water that...
Replies
7
Views
5,319
Hi guys. First post ;) We are working on a project for my company where we get data from some analog and digital transmitters (pressure, on/off...
Replies
6
Views
2,083
I am doing some studying on Flow Measurements and stuck on this problem? A dp transmitter is calibrated 0 to 80 inH20 and transmits a 4 to 20 mA...
Replies
30
Views
9,168
Hey This might have been up several times, but i haven't found an answer that solves my problem. I have some flow meters that measures with 10...
Replies
18
Views
7,526
Back
Top Bottom