Line Sizing 3

[Chem2020]

Per API 14E there are some criteria for selecting and sizing pipe lines. For centrifugal pipes it is recommended the velocity at the suction should be 2-3 ft/s and discharge 6-9 ft/s.
I need to know what is the logic behind this? What would happen if velocity is less or more than the recommended range. Thanks alot

 

[TD2K]

The 2 to 3 ft/sec for centrifugal pumps (I'm assuming the pipes is a typo) is to avoid excessive suction pressure losses for bubble point liquids. If your fluid isn't bubble point, you can go higher, I've seen over 10 ft/sec in suction lines in diesel tank farm pumps for example. The best way to check is to estimate the pressure losses through your piping and see if it's acceptable for your system.

On the discharge side, the 6 to 9 ft/sec for plant piping based on reasonable pressure drops. Lower velocities result in larger and more expensive piping, higher velocities can result (especially in smaller lines) in higher pressure drops which have to be supplied by your pump. Calculate the pressure drop for water in a 2" pipe at 15 ft/sec to see what I mean. Pipelines are another story, the velocities in them are typically quite a bit lower because otherwise the pressure drop and pumping costs would become prohibitive. Inplant piping might be sized for 2 to 3 psi per 100'. Try to do that over a 20 mile pipeline and you have a good sized pump.

Try picking a few flow rates and see how the velocity and pressure drop is affected by going up and down a line size.

 

[moltenmetal]

TD2K has correctly explained this. These rules of thumb are "economic velocities" based on an optimization between the cost of piping (pipe, fittings, fabrication, installation, testing etc.) and the cost of energy to run pumps. Obviously that optimum velocity shifts depending on what you're doing with that piping, and DP per 100 ft of pipe is a better measure- again as TD2K mentions. Some erroneously conclude that the reason you avoid velocities above 15 ft/s for liquids in plant piping is to avoid erosion, or that using 2-3 ft/s will avoid NPSH problems at the pump suction without further analysis etc. There's more to properly sizing pipe than applying a few rules of thumb.

 

[silebi66]

Generally, velocity in the pipe greater than 15 fps in CS pipe can cause erosion.And, low velocities (< 3 fps) in the pipe may cause entrained solids to settle and plug lines.

Pump suction loss should consider NPSHA, for several pumps operate in parallel, on sudden failure of one pump to reduce the NPSHA due to flow rate increasing may require a higher elevation to prevent
cavitation. As general good practice, the suction line and leads to the pump should be sized for a unit pressure drop of 1~2 psi/100 ft.

 

[TD2K]

I'm not so sure I agree that 15 ft/sec is a problem with clean fluids. Add some sand to that or the potential for cavitation and that's a very different situation I realize. Too low a velocity as Silebi66 points out can be a problem in liquids with solids.

I wouldn't be surprised if the inlets and outlets of control valves, the velocity through an orifice plate, pump discharge nozzles didn't run more than 15 ft/sec more times than you might think. I bet some smaller bore cooling water lines in plants run in this range where you have a 20 to 30 psi differential between your supply and return header and something like an off the shelf lube oil cooler that is designed to take no where near this pressure drop so the flow increases until the pressure drop is being taken by the piping and the cooler.

We had one letdown valve in a pretty clean liquid service when I was in Saudi Arabia that ran about 25 ft/sec and while the corrosion group was monitoring it, they didn't see any signs of erosion. When I got cases where I was pushing the API 14E velocity limit I would talk to them but they weren't really concerned as long as it was a clean liquid. Flowlines downstream of wells going to a GOSP, we didn't push those velocities.

 

[moltenmetal]

Again I'm with TD2K here. We've used pipe and tubing in clean liquid services at velocities far greater than 15 ft/s and never seen evidence of erosion. River water with sediment in it? Slurries? Situations where there's risk of erosion-corrosion? Sure- those have their own velocity rules which vary from service to service.

For clean liquids, 15 ft/s is at the upper limit of the economic velocity in piping due to pump energy cost, NOT due to erosion risk. When you're wasting the energy anyway, i.e. in a pressure let-down train, as long as the control valve is still the controlling resistance in the circuit, we have no problem letting pipe take more of the pressure drop by allowing velocities to climb well beyond 15 ft/s.

We also make the switch from 2" to 3" pipe only if velocity goes beyond 15 ft/s, where available pressure drop permits. The capital cost jump in linesize from 2" threaded to 3" welded is huge- vastly greater than from 3" to 4" etc.

The velocity in pretty much every orifice plate, control valve trim etc. we specify is well above 15 ft/s, as are the exit nozzles of most centrifugal pumps. We don't see these wearing out quickly in clean liquid services.