Since you're doing this by hand:

Assume that the entire load is applied at the end of the run vs. at discrete points along the line. Measure the total length of piping and then count up fittings/valves/etc and determine equivalent pipe length. Use your equation then to solve for D given the load (at the end), desired pressure drop, and total length. With it being parallel, you'll have to balance the load across the two runs (i.e. iterative solution).

Take a look at the book Piping Calculations Manual by Shashi Menon. It may have some equations on argon. Otherwise, you'll have to find a source for a flow equation for it. Don't use Bernoulli, that's for incompressible flow.

Hi,
A good resource is " flow of fluids through valves ,fittings and pipe" Technical paper N 410M by crane .
The hypothesis of incompressible flow is valid if Mach number is below 0.3 .

Note : you should prepare a sketch ( simple diagram) and attached it to your post for us to better understand the query.

Pierre

Thank you, alchemon & pierreick. I will upload a sketch when I get home in about an hour.

Presumably the 150 psig pressure at your tie-in point can be maintained with all existing machines + the new machines running. If so, you can just size your new loop using that 150 psig source. Otherwise you may need to analyse the whole system.

You hint that it is a loop, which would normally mean that the argon is not consumed at the machines, its just used and recaptured for transport back to the tank. Please confirm. If so, you will presumably need a compressor to reinject it back into the tank. You will somehow need to establish the loop outlet pressure.

Bernoulli works, but you have to introduce additional terms for pressure loss, which is not included in the classic Bernoilli relationship. And that pressure drop will need an equation of state to determine the density of the gas at the various pressures and temperatures you will have in the loop, so you can determine gas viscosity, RReynolds number, the friction factor and corresponding pressure losses. All of those are generally included in compressible gas formulations to calculate pressure drop. We just need to select one of several available, which might be D'Arcy, Colebrook White, Churchill, etc.

I will await answers to my questions and your sketch, before proceeding further.

It sounds like you have a network or at least a complex piping model.

So yes, changes to that system to add another supply point will impact the rest of the network / system.

1) Forget Bernoulli - think friction flow calculations
2) If there is no good reliable diagram - draw one - great thing for an intern to do is send him or her out and say measure everything and tell me what our system ACTUALLY looks like compared to my drawing which was last updated 10 years ago. So every pipe length, size, branch, valve, elbow, tee, pressure regulator or flow control, each user etc.
3) Then remember that mass cannot be altered, so mass in = mass out and you can't get more mass out than mass in other than very short term flows
4) Remember that at a node (tee, junction or whatever, the pressure must be equal in all branches at that point.
5) You might need to plug all this into a network model program as otherwise it gets too complex and time consuming to calculate all the pressure drops and flows
6) In pressurised gas flow, ALWAYS ASK when someone gives you a flow rate whether this is at a certain pressure or whether it is in standard or normal volumes. So your 27GPM at 73 psi, is that really 27GPM at a pressure of 73 psi or 27 GPM at standard conditions, but the machine needs 73 psi. It makes a massive difference in gas flow design and lots of people get it wrong or don't know.
7) Be aware that pressure drop might not be your limiting factor but actual gas velocity. Some systems don't like going at high velocity, so make sure you check and know what your limits are to company standards.
8) Be aware and check what the actual ID of the pipe is. different pipes have different wt and ways of using sizes. So a 2" pipe might be 2" actual OD, it might be 2 1/2" actual OD or indeed somewhere in between. Check and find out as it makes a massive impact on frictional losses.

In terms of your point above, to be able to answer than you need to know what the pressure is going to be at your take off point once you add the extra flow. That's your start point to know if your pipe is going to be big enough or not.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

I agree with LittleInch .... especially his point about the need for a network model program and his point #7

MJCronin
Sr. Process Engineer

OK so here is my drawing and a little explanation of what I've done so far

• https://files.engineering.com/getfile.aspx?folder=635010bc-8744-4196-a5e2-9

Sketch

• https://files.engineering.com/getfile.aspx?folder=f2413645-ce49-4c41-9132-f

pipe velocity for comparion, I'm using the air range of 10-30 m/s as my target because air is mostly nitrogen

• https://files.engineering.com/getfile.aspx?folder=86118a8e-504f-40f9-a432-3

Is it nitrogen or argon? Your sketch shows liquid nitrogen. Also the loop doesn't show it returning to the source. Please confirm.

Assuming this is correct, I stand by my original statement. Assume that all of the load is applied to the end of the loop. Then use the procedure for a parallel pipe run. If it is nitrogen, compressed air equations will get you close enough.

The main issue I see is you'll need to estimate the inlet pressure to your new loop, from the existing loop. You are being provided 150 psi at the nitrogen tank. You'll have to determine the drop on the existing line with everything running. Perhaps you're in luck and you can look at a gauge at the end of the pipe with everything running. Otherwise, to do this by hand, we will have to make some simplying assumptions.

Per this, this is a traditional parallel pipe run, ie the flow will split between both branches. In theory if the flow resistance were the same in both branches, then the each branch would supply half the flow.

If the existing loop is 1", the techs recommendation of a 1-1/2" passes the smell test.

Its nitrogen, if I said argon that was a mistake, sorry. also the fluid is being consumed and vented to atmosphere not returned or reclaimed

Building these as loops is helpful because you have to always flow the 'long way' to get to service.
Is there pressure regulation at each use point?
As it has been said before know what they are actually talking about when they say 1 1/2". Is it tube (actual 1.5" OD) or pipe (1.900" OD), and then you need to know the wall (tubing gage or pipe schedule). We supplied a lot of SS tube for gas systems, but some customers bought Sch 5 pipe for them. Big difference.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

Agree with the others.

You either need to tap into the system at your chosen point to measure pressure or calculate it based on the flows in the system.

Then work out what the pressure will be at your point A when you add all that new supply.

The 150 psi will only apply when at the start of the system.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

This isn't related to your question, but it's worth throwing out there - ask about how the risk of having nitrogen being vented within the room(s) is being handled. If there is a credible mechanism for lots of nitrogen gas to get released if something goes wrong and makes the room oxygen concentration fall, then it is an asphyxiation hazard. I expect your company is aware of this risk and knows how to handle it, so it'll be an extra thing you learn.

I can hook a pressure gauge to the valve we are going to build off of. Once I have that pressure at point A, can I assume the rest of my calculations are carried out properly?

You should know the pressure at point A as a function of how much flow is being drawn by the users upstream of point A as alchemon pointed out. A single reading will be misleading as the starting point for your calcs. Take readings at different times of the day for a couple days to get a minimum pressure and use that as the pressure at point A.

This may not be possible at your site, but you could coordinate an event where all the machines are on and force the minimum pressure reading.

Yes - put a gauge on the valve at point A and record the actual pressure in the line while everything else is running, and at various points in time as mentioned. That should give you the inlet pressure to your new loop. Then do the loop calculation as we have discussed previously. When you are done, check the velocity at the OUTLET pressure, i.e. the minimum pressure, end point of your new loop. Remember, mass, not volume is conserved, and as the pressure decreases, the nitrogen will "speed up" to maintain mass balance.

If the velocity is too high, go one pipe diameter up.

The forced event is not possible at the site, there is a lot of machines that procees spans 2-10 days and the schedules are staggered all on top of each other. but I feel like I'm at a good place to take some readings and finish the calc. I will post update post gauge readings. Thank you so much!

Hi,
It might be safer (stability) to add a buffer tank at the discharge of the evaporator and from there to supply the 2 networks separately with their own pipeline .
how far are the equipment from the evaporator ?
https://www.atlascopco.com/en-gr/compressors/wiki/...

my thought.
Pierre