Calculating GPM Flow by using a differential pressure sensor

A venturi meter used differential pressure through an orifice in a pipe to determine flow rate. Usually it is used for incompressible fluids and the flow rate calculation assumes that you know several factors, that is pipe diameter, orifice coefficient (usually a function of reynolds number, but generally a constant of the meter) and fluid density. As for flow rate it is fairly easy to modify the manometer scale such that under normal operating conditions, flow rate can be read directly--in the old days this was done using a scale monograph--today if the differential pressure is measured digitally--you let the computer crunch the nomograph directly to flow rate.

Venturi meters can also be used for compressible fluids, however gas conditions are different and choked resulting from Mach effects at sonic throat velocities.

See this web site for a flow calculator, formulae, and a better discussion venturi flow meter

Rap

Are you taking the piss rap?

It all seems a bit ironic to me but perhaps that is due to the literary influences I have sadly been exposed to.

Either way it's pretty good.

The GPSA gives the following formula for orifice meters in liquid service:

Qh = C' SQRT(Hw)

where

Qh = liquid flow in gpm
C' = orifice constant (Fb x Fgt x Fr)
Hw = differential pressure in inches of water
Fb = orifice factor. These are tabularized values, depending on the pipe diameter and orifice diameter. They are given in the GPSA, and probably also in Perry's.
Fgt = specific gravity temperature factor for bellows meters. Also a tabularized value
Fr = Reynolds number viscosity factor for liquids. Also a tabularized value.

I am surprized that an engineer wouldn't already know this, or couldn't figure it out himself/herself from the reference materials.

Thanks for the response, however the setup is rudimentary. He just wants to install the connection to pipe without an orifice. I keep telling him it can't be accomplished this way but he insists it can but can't also give you the formula.
Thanks
Frank

An orifice type flow meter typically gives you accuracies of plus or minus 2%. What your engineer is suggesting will give you an answer, but the accuracy will be far worse, and you probably won't be getting a reading on your DCS or board.

A well designed liquid flow system typically has pressure losses of 1 psi per 100 feet of equivalent pipe. To have a 10 psi differential, you would have to have two pressure gages in your system separated by 1000 feet of equivalent piping, and then factor in differences in elevation. Once you have your differential pressure and you know your total length in equivalent feet, then by trial and error you can back calculate your flow, using the equation:

deltaP = (2 f L V2 density)/(D gc)

where:

Delta P = differential pressure in pound-fotce per foot squared
f = friction factor, from the Reynolds Number graph
L = length of piping in equivalent feet
V2 = velocity of fluid flow squared, in feet squared per second squared
Density of fluid in pound mass per cubic feet
D = diameter of the pipe (internal) in feet
gc = conversion factor in (feet pound mass)/(pound force second squared)

This method will be cheap, but it will only give you a ballpark, "quick and dirty" answer. If the flow rate is something you really need an accurate value for, reconsider the orifice meter.

Midwest Instruments makes aneat digital Q sensor for almost any sze pipe or viscosity of fluid. THESE CAN BE RENTED , so their is no real excuse to not hold Q as a fixed value if need be.(Knowing an accurate gpm is usually a test function so it can be controlled no?)

UOP Johnson has a handbook in which the GPM can be calculated +/- 10% by measuring the arc of the discharege line for any given diameter pipe(assuming that the fluid is flowing pipe full).


How critiCAL IS YOUR GPM VALUE? USGS REQUIRES A +/- 1% ACCURAACY FOR PUMPING TESTS ON AQUIFERS.


Hey rap, long time no see, where ya been?

Flow sensors
There are also sonic flow sensors which are costly but accurate. If you are using the length of pipe method, it is better the longer the run of pipe is. But you also may need to consider inflow conditions as these produce losses. If the engineer is a good ME, he should know all this. But if not have him check it out with a good ChemE. I am a professional ME.

CBS

Is this just to say that ifFb = orifice factor. These are tabularized values, depending on the pipe diameter and orifice diameter I had a 1/2 inch pipe that the value would be .5?

All of the formulas above have one important relationship in common... pressure drop is proportional to flow rate squared (for incompressable liquids). If you have flow through a system with significant pressure drops, you can calculate flow based on pressure drop assuming you have a reference point and that the resistance of the system doesn't change over time (due to scaling, valve operations, etc.) This is a dumb way to calculate flow since it is almost certain that the system resistance to flow will change over time, but it's possible to do it this way. Is there any reason you wouldn't put a cheap rotameter on it?

Here's my situation. I am a controls electrician and I have a project for proving energy savings. One of the aspects of the project is to measure the BTU that are used by a Finn Tube radiator in a classroom. What I have is a differential pressure sensor on it with supply temp and return temp sensors. The piping is half inch copper piping with no change in diameter. Any help is really appreciated!

Is this a steam system then? What kind of temperatures are you looking at?

No, the coefficient is not related to the pipe diameter like that. You need to go to the tables which demonstrate head losses for various configurations of piping systems. If the flow through the pipe is well developed, no elbows (turns) either upstream or down stream, then the standard head loss formulas can be used. What happens is that the flow stream through a turn starts to spiral because the flow along the outside of the turn and the inside of the turn are along different paths and on exiting the turn there is a differential in speed which will affect the flow on either side of the turn. So pressure sensors will not indicate the correct condition until the flow becomes uniform. In laboratory bench tests you can use the pump to determine the flow and from that calibrate the pressure sensor data. Trying to measure flow rate depends upon how accurate you need to be. Knowing the pump head curve and power input of the motor will often be a better way to understand the flow rate.

Here're my thoughts, worth exactly what you paid and based on my non-existent knowledge of your system.

You don't need a robust system for measuring flow, you need a one shot deal that produces a reasonable answer for your experiment and is consistent. I would measure flow into your boiler/heater or at the suction or discharge of your pump. Assuming your are maintaining a constant level in your boiler (and this is a steam system), the mass in has to equal the mass out, so what is going into the boiler must match want is going through the radiator. (It is very important to monitor for accumulation in the boiler!). You can measure this with a simple in-line rotameter which you could leave in place after the experiment. You might already have one. I bet there is some flow monitoring system already in place for the pump. Measuring superheated steam is much harder. Measuring steam/water mixtures is a killer. If this is an oil system, the rotameter idea should work as well.

This is a hydronic system.

There is plenty of elbows in the system before the sensor. But between then up stream and down stream sensor wells it only has two elbows. Fortunately this dosent have to be super accurate. I am only looking to get a good approxamation.

You should be able to use a rotameter to measure flow directly anywhere in the system. Do you have any flow measurement points at all?

how would you figure flow rate for chilled water in an 8 inch pipe with 34 pounds of DP across the overall secondary chilled water system? the secondary pump speed is about 50% (30HZ).

I think the system DP is too high and the building operators think that it needs to be this high to fix 1 load that is not getting enough cooling. the flow will increase and decrease based on load and all my Onicon flow meters have stopped working

weswashere

I am very week in maths