Line Sizing Calculation 3

[ProcessRookie]

I have been tasked with selecting the line size for a new piece of pipework carrying process water.

I can get flow rates, typical viscosities and densities and an idea of the pipe length/route.

How should I proceed in calculating the minimum line size?

At first glance I see that I may be able to use Darcy's law but getting information on the pressure drop or inlet pressure may be difficult unless I just use the pressure given on the line list.

 

[Latexman]

"How should I proceed in calculating the minimum line size?" You should not. You should pick the optimum line size.

Have you used Search? It's between Forum and FAQs near the top, left of page. That's a good start, because this has been discussed many times.

Good luck,
Latexman

Need help writing a question or understanding a reply? forum1529

 

[bimr]

Reasonable pipe velocities depend on the application. There is no correct velocity for all applications. Here is a general guideline.

Reasonable Velocities for the Flow of Water through Pipe:

Boiler Feed.............8 to 15 ft/sec
Pump Suction ............4 to 7 ft/sec
General Service.........4 to 10 ft/sec
City.......................to 7 ft/sec
Transmission Pipelines...3 to 5 ft/sec

 

[SNORGY]

In the absence of better guidelines, this might be a way to start:
(1) Use the API-14E erosional velocity correlation to estimate an upper limit on flow velocity. For process water, u = [100/(sqrt(rho))] = 3.16 m/s.
(2) for your flow rate Q, use u = 4Q/(3.142*d^2) and pick the closest standard line size.
(3) dH = f*L/d*u^2/(2*g)+Z, where Z is the elevation rise (metres) from origin to termination.
(4) dP = rho*g*dH
You can either accept the dP, or not. If you can, you're done.

 

[ProcessRookie]

Thank you very much for the help.

The line is gravity fed into an atmospheric sump so I think I can work out my driving force (dP) using the difference in elevation.

SNORGY, where can I find the reference for the equations/method you have given?

 

[bimr]

In the absence of better guidelines? I posted reasonable velocities above. They are taken from Cranes Technical Paper 401.

3.16 m/s is an extreme velocity as evidenced by the fact that it will generate a pressure drop of 9 psi/100 feet of pipeline.

Water lines in chemical plants are typically sized with velocities of 4-6 ft/sec.

If you want to size the line using gravity flow, use the Hazen-Williams formula for a full pipe:

http://www.calctool.org/CALC/eng/civil/hazen-williams_g

 

[SNORGY]

Sorry, bimr.

I was obviously still typing while you were hitting [Submit Post]. When I subsequently hit [Submit Post], it might have given the appearance that I was taking issue with your information, and that was certainly not my intent. Your guidelines are perfectly appropriate and indeed more in line with what typical Operators / Producers would use, although I have been used to seeing Vmax ~ / = 10 ft/s in upstream / midstream plants for produced water. My statement snippet "...in the absence of better guidelines..." was made under the assumption that the OP might not have had a Client's (or Employer's) guideline document at hand that specified upper velocity limits.

So, STAR for your post as well.

ProcessRookie:

API 14E is a reference document published by American Petroleum Institute. The equation I posted coincides with the perhaps more familiar "rho-V-squared" criteria that you might have heard from various sources. Typical rho-V-squared values are on the order of 10,000 with rho in lb/ft^3 and V in ft/s. Algebraically rearranging, you get V = K/(sqrt(rho)), with K = 100 for the units specified. For Metric units, K = 122. The other equations are just a combination of Darcy's formula for dynamic head loss and an adjustment for elevation difference. As bimr states, Crane TP-410 is a very good resource.

 

[bimr]

no problem

 

[SNORGY]

That said, I still fouled up in my units or my arithmetic somewhere. For water with k = 100 and rho = 62.4 lbm/ft^3, k/(sqrt(rho)) = V = 12.66 ft/s = 3.86 ft/s. In Metric, 122/sqrt(1000) = 3.86 ft/s.

Stupid non-Metric units. Stupid iPad. Stupid fat fingers.

 

[ProcessRookie]

Thanks a lot people. I now have Crane, API 14 E (my clients guidance document on it anyway) and a clear path forward. My transition from Graduate to reasonably good engineer will hopefully begin.

 

[SNORGY]

ProcessRookie:

Let's hope you end up better than me.

"For water with k = 100 and rho = 62.4 lbm/ft^3, k/(sqrt(rho)) = V = 12.66 ft/s = 3.86 ft/s. In Metric, 122/sqrt(1000) = 3.86 ft/s."

REALLY?

Try:

(Rev. 3) "For water with k = 100 and rho = 62.4 lbm/ft^3, k/(sqrt(rho)) = V = 12.66 ft/s = 3.86 m/s. In Metric, 122/sqrt(1000) = 3.86 m/s."

By the time this thread gets to 100 posts and Rev. 39, I might get it right.

Stupid units.

 

[MortenA]

another tip:

If this is an "open line" (like a drain pipe) you should NOT use Bernoilli (or similar) since its the gurgle point that determines the max flow. Base you capacity on the Fround no. instead.

Best regards, Morten

 

[ProcessRookie]

MortenA,

This is one of the problems I have, at normal operation I have fully flooded flow as the level in the tank is above the outlet pipe however, the low low levl alarm is set below the top of the outlet pipe so non flooded flow can occur.

Thanks,

PR

 

[SNORGY]

ProcessRookie,

If you Google one or all of "flow through partially filled pipe", "open channel flow", "Manning equation" you will come across a few good sites that have literature and examples on how to solve that kind of problem. One site even sells the spreadsheet itself for about $28.00 U.S.

 

[25362]

Some useful reading:

thread378-332035
thread378-46739
thread378-87742
thread378-86817
thread162-195742

 

[ProcessRookie]

25362,

Thank you for the links, some of the problems are very similar to my own.

It seems like the best solution is to design for fully flooded pipe using Darcy (as this will be how the system operates 99% of the time) and install a simple air vent at the local high point in the line to avoid air locking.

Due to the fact that there is an alarm and a trip stopping the tank level reach low low, I don't think it is justifiable to size lines based on Froude number; the air vent seems the best solution.

Does this seem logically sound to your experienced selves?

Thanks for all the advice, I think this forum has been a great find for me. A good way to get access to different experience than collegues.

Cheers,

PR