Can anyone help me on determining the flow rate of argon? 1

[haydenle]

I am an intern at a steel mill. I very new to this, so bear with me. I currently have a project to size a flowmeter. I know the pressure upstream is around 200psig, this is coming from and Airliquide vessel outside the plant. Down stream we have a pressure regulator and it is regulated down to 80 psig. The headlosses for the system will take forever to determine. The temperature leaving the vessel is -20 degrees F and I am not sure what the temp is at the regulator. Please any help would be greatly appreciated.

 

[Zapster]

Constant enthalpy?
Heat gain/loss from piping to 80 psig regulator depends on flow rate, ambient temp, insulation, etc.

Look at

http://webbook.nist.gov/chemistry/fluid/

for argon thermo physical properties.

I assume you want the temp of gas downstream of the 80 psig regulator.

Determine the process, look up the thermo physical properties.

 

[haydenle]

Actually I need the flow rate. Sorry for the confusion. I am trying to determine the max and min flow rates for the application. I believe the temperature of the gas is 70 degree F, because the pipe goes underground.

 

[BigInch]

All you need to find out is what range of flowrates you want to measure and where you will have room to put in a flow meter, in the 200 psi line or in the 80 psi line.

You will also need to determine how accurate you need the flowrate measurement. A rough flowrate can be determined from a simple orifice plate, for +/- 0.5% you'll need a proper flow meter of some type of which there are several types to chose from, a meter tube with a machined orifice fitting or a turbine meter, or you could use a coriolis meter. You should have a straight place about equal to 20 x the pipe diameter to install a proper metering device.

You may be able to get an idea of the average flowrates by finding out how many times they are filling the tank and divide by the hours of operation, then apply some "factors".

If you still can't estimate the hi and lo flow rates, make a rough estimate of the lengths of pipe of various pipe diameters going from the vessel to the regulator, measure the temp at the tank and regulator and repost the answer here, or send me an e-mail if you like.

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[Latexman]

Do you have a level instrument or, better yet, a level recorder on the argon tank? If you have a recorder or data achive, use that to determine minimum, average, and maximum usage rates. If the tank vents off vapor to control it's pressure and temperature you'll have to figure that out too.

Good luck,
Latexman

 

[danw2]

Another suggestion for guesstimating flow rate:

How often do is your argon delivered and in what quantity?

Divide the volume of delivered product by the number of minutes/hours between deliveries for a guesstimate.
Modify that by whether the argon is used continuously, like a continuous bleed into an atmosphere furnace, or unevenly in batches.

I've had similar situations, where the flow rates were total unknowns.

What I did was install a set of orifice flange unions, and a valved bypass on the straightest run of pipe I could find. I got a multivariable DP transmitter which takes a temperature reading from an RTD in addition to knowing the DP and static AP upstream of the orifice (I use Honeywell SMV, Rosemount makes one, too.) Rather than install a
thermowell and RTD for just a flowrate test, I configure the meter for a fixed gas temperature and buy 3 orifice plates sized to beta 0.75, 0.5 & 0.3 for the pipe diameter at the working pressure of the line.

Then I'd install the 0.75 plate and check flow rate. If the flow was down in the noise region, I'd go to bypass, swap to the 0.5 plate, and see what it looked like. If low on the 0.5 plate, I go to 0.3.

My assumption (not always valid) is that someone sized the pipe to maintain some reasonable flow rate, so I have a fair chance of hitting a readable flow somewhere over a range of 0.3 - 0.75 beta for a given pipe size.

I've done this 4 times so far and it's worked OK 3 times. The one time it didn't the flow rate was pitifully small for a 2" line, so I rigged a section of pipe with 3/4" line and got a reasonable flow rate from the 3/4" line.

I don't leave the multivariable in the line because it's a test tool. It's been replaced with 3 different flow meters: a rotameter, a vortex meter and a turbine meter, depending on service, but each of those flowmeters needs a flow range to be sized correctly.

Clamp-on ultrasonics might now be useable on gases, but when I did my projects, I couldn't hire anyone to measure gas flows with clamp-ons. Technology has likely changed over 5 years.

Dan

 

[haydenle]

Well, I need the flow rate going to the purging system. This is the 80 psi end of the regulator. There is at least 100 yard(estimate) of piping and who know how many bends. There is also a elevation change of about one story, so lets say 20 ft. The temp leaving the vaporizer, right before going underground, is -20 degree F below ambient. I believe the temp at the regulator is ambient temperature. The flow rate at the outside regulator is 480000 scfh. This flow rate is teed off and slit into two seperate piping system. I am assuming the flow rate is 24000 scfh for each line.

 

[StoneCold]

That is a very big tank. Are you sure about the 480,000 scfh? Since regulators do not indicate flow I assume that you are basing the flow rate statement on changes in tank level? Correct?

StoneCold

 

[Cockroach]

As stated in a previous thread on orifice metering, the volumetric flow is simply the stream velocity times the area of the bore of the pipeline. You can use the Bernoulli Equation to determine velocity of the fluid in the pipeline as a function of pressure drop.

Without getting heavy handed, you can make the assumption that the flow is isentalphic and not compressed. It has been correctly pointed out from Zapster's comment on isenthalpic flow that in the absence of throttling, heat loss is neglegable. Other assumptions are that the pipeline is perfectly level (no potential gradient), frictional losses are low (ideal fluid), efficient exit condition (unity for velocity coefficient) and isentropic flow (reversable). Then:

P1/rho1 + 1/2 v1^2 = P2/rho2 + 1/2 v2^2

set v1 = 0 (vessel very much larger than pipeline; this means the surface drop is small compared to fluid exiting the pipeline) and we would get the velocity of the fluid through the pipeline as:

v2 = sqrt(2[P1-P2]/rho) rho=fluid density

This is the Bernoulli Equation under all the stated assumptions, thus giving us stream velocity. Since the area of the pipeline bore is A=1/4(ID^2) then the volumetric flow of your stream is:

V' = v2 X A Volumetric Flow Rate

You have enough information to get stream velocity, but failed to state pipeline size and schedule. I cannot compute a final, approximate answer.

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada

 

[Cockroach]

The area of the pipeline bore is "1/4 pi (ID^2)", I forgot the coefficient of pi=3.14156 in the original post. Sorry.

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada

 

[iainuts]

Hi WayneLee. As StoneCold mentions, the 480,000 SCFH is rather high. That's the equivalent of roughly 4200 gallons of liquid argon per hour. You'd need a decent sized ASU on site to furnish that, not a tank.

I'd agree with the folks saying you should check your delivary records, but there's an even better way to get more accurate and time dependant data. You could chart the tank liquid level directly. Most, if not all of your industrial gas companies use something called "telemetry". What happens is there's an electronic box attached to the dP guage that's used to determine tank quantity. That box takes readings every 15 minuts or every few hours, depending on how it's hooked up. It then stores a large number of data points, on the order of thousands, and then when it is programmed to, it phone calls the industrial gas supplier and downloads that information to a database.

I'd suggest you ask your industrial gas provider for that information in EXCEL format and then you can plot the data over any period of time. It should be as easy as them pulling it down from their database and emailing it to you. They should also be able to correlate for you the liquid contents versus dP if they don't already have that in the database. Also, if the data they have is not taken in small enough time increments for what you need, you could ask them to set the telemetry to almost any time period you want and have them send you that data over any time interval. You could for example, ask them to have the telemetry record a data point as little as every 15 seconds. Just tell them what you want and they should be able to pull it down from a database or create a special one for you.

Anyway, talk to your industrial gas salesman for this type of information. He should be glad to send it to you, and if he doesn't, tell us what company it is you get the Argon from and I suspect someone here could "help you out" so to speak.

 

[haydenle]

Im sorry, I meant to put 48,000 scfh. There is a meter on the tank outside of what level of liquid argon is left in the tank. Maybe I could see if there is some type of delivery record. The pipe size is 1 inch and schedule is 40. Also, could I just put a anemometer at the end of the hose that connects to the vessel. This would give me a velocity, which I can then use bernoulli's equation to get flow rate in the pipe.

 

[haydenle]

Is it safe to use bernoulli's equation and neglect any any elevation change.

p1+rho*v1^2/2=p2+rho*v2^2/2

let v2 be negligible
p1=80psi
p2=0

Solving for v1 gives me p2-p1=-80psi
Can I just take the absolute value? To solve for v1 I would need a positive number under the square root. However, I end up with a negative number under the square. Please can anyone tell this is feasible?

 

[BigInch]

Yes you can neglect elevations,because gas density doesn't change much with small elevation changes, but you must adjust rho at different pressures. P2 does not equal 0. You must use absolute pressures (add 14.696 psi to each gauge pressure) and absolute temperatures must also be used if temperatures are to be considered. You are also assuming that it is an ideal gas, which probably doesn't matter too much, but you might want to check the compressibility factor for argon to be sure.

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[haydenle]

Well, the only way know how to solve the problem is: assume ideal gas and use equation of state.

P1*V1/T1=P2*V2/T2

(180+14.7)psia*24000scfh=(80+14.7)psia*V2 (neglect temperature change)

I end up with 900 scfm

Tell me, is this is a sufficient enough answer to size my flow meter?

Also, I am using scfm, do I need to convert to acfm to size my flow meter?

When converting to acfm, do you have to take into account that the fluid is argon?


Please any help will be greatly appreciated.

 

[insult2injury]

1. How can you assume ideal gas with a liquid supply?

2. The temperature change is 90 degrees, I wouldn't neglect that, not to mention phase change at some point.

3. Assuming the 24000 scfh is only accurate if both circuits after the "T" have equivalent losses.

4. With your bournoulli equation above, you need to solve for v2, v1 is the tank velocity which is assumed 0. That equation only applies if the mass flow rate at both points 1 and 2 are equal which, if I understand correctly, is not the case.

 

[BigInch]

Insult,


Bernoulli across the orifice,
(note i left out the density variable on purpose)

1. If its gas before the meter and gas after the meter and the meter only loses 10-20 psi with no temperature change across the meter, yes I would assume an ideal gas to size the meter.

2. The temperature changed is 90 degrees from where? If it doesn't change 90 degrees across the meter its not relavent.

3. Why not assume flow can be controlled to be equal?

4. V1 is the flow upstream, V2 and downstream of the meter, V1 = V2 (assuming upstream and downstream piping is the same diameter). What is missing is the head loss of the flow across the orifice.

H1 + V1^2 + P1- Hlo = H2 + V2^2 + P2

assume pressure loss is 10-20 psi, T1 = T2, ideal gas
Same pipe diameter upstream and downstream, V1 = V2
Estimate orifice diameter first using ideal gas
for small pressure drop rho1 = rho2
Assume elevation upstream = elevation downstream
H1 = H2

V1^2 + P1- Hlo= V2^2 + P2

V1 = V2
P1 - Hlo = P2
=====================================================
Lee,
To estimate size the meter (an orifice plate with dP)
Use Bernoulli upstream of orifice and through (inside) the orifice

V1 = velocity upstream
V2 = velocity through orifice
H1 = H2
rho1 = rho2, small pressure drop
T1 = T2

H1+V1^2 + P1= H2 + V2^2 + P2
V1^2 + P1= V2^2 + P2
P1-P2 = V2^2-V1^2

then downstream
P2 - P3 = V3^2 - V2 + orifice flow recovery factor * (P2-P3)

iterate until you only have a small pressure drop

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[insult2injury]

BigInch, I understood what you were saying, but he wasn't applying it correctly. I think he was trying to follow cockroach's instructions.

 

[haydenle]

Insult

I believe it is a gas supply. It is supplied from a 6050 gallon vessel, which then flow through a vaporizer.

Big Inch

I dont believe I have access to an oriface plate. Unless I am not understanding something, are you referring to bernoulli's equation or is that the actual equation?

 

[insult2injury]

If you can measure the pressure and temperature at two locations along that 80 psig line (preferably straight in between), you can then use Bernoulli's equation to back into the velocity because you know that the mass flow at each point must be equal. Furthermore, if the pipe is uniform and the temperature and pressure changes are small, then the stream velocity at each point must be equal as well so you could simply the equation.