Centrifugal Pump Suction-Inlet Design & Best Practices

You should check velocities at all points.
10 fps isn't all that bad as long as it leaves you with enough pressure (NPSHR) at the pump and the pipes don't shake on start or stop, or a quick valve closing.
Avoid velocities over 15 fps everywhere.
7 fps in long pipelines gives better pumping costs. If its not a long pipeline, generally no problems.
If its a suction line, then (again) as long as NPSHR is OK its fine. Low velocities are better; minimum probably 3 fps.
If you do not know the NPSHR of the pump, then you should always have at least 6 psig at a water pump suction inlet.
Your Sizes are typically OK.
Avoid ells near suction inlets.
Header diameters 1 to 2 Sizes larger than the largest branch diameter. Use 2 for larger diameter piping. 1 for smaller diameter piping.
3 pumps are usually 2x more difficult to use than 2. Generally the less, the better.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

What did this look like in terms of layout?

If you avoid getting 1800gpm at any point in your 6" pipe would be good. That's a very high velocity (22ft/ sec).

Is this manifold being fed by me than one tank? How many?

10 fps is a bit high but at those lengths you won't be hurting for NPSH so it should be OK. Go for two pumps if you don't mind losing one for a while if it stops.

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

Thank you very much.

If this was a permanent facility/plant install and pumps running steady state, then it would be engineered 100% and focused on pump efficiency; however we run remediation systems intermittently depending on type of project, whether we're treating with coagulants/flocculants for suspended solids on construction sites, which requires a larger number of tanks to keep tank velocities down for proper settling prior to sand filtration...

I attached a photo for reference... yes we set up tanks side by side parallel to each other; inject coagulant/flocculant into "Pretreat Train" filling into rear of tanks; solids settle out before conveyance out of tank front into manifold and onto sand filters.

How many tanks depends on overall flow rate, and what the contaminants of concern are. Suspended solids (construction, clay soils, sediment, etc...) it's good to target about 150 gpm per tank.... so 1,800 gpm/150 gpm = 12 tanks... Other projects like ground water remediation, where water typically runs clean in terms of solids but treating with GAC for dissolved metals, VOC's, etc... we can run a higher flow rate/velocity through tanks and have fewer, but still may be other limiting factors.

Yes three pumps is ridiculous and I'm trying to educate my team. Two is perfect as this provides what I call internal redundancy in case one goes down.

Thank you thank you!

jcth said:

Best practices: 5-10 pipe i.d. straight inlet run to pump suctions; 1-2 suction Dia sizes larger than pump suction Dia; suction velocity no higher than 8 ft/sec.

Click to expand...

Where have you taken those from?

for info

If flow rate is nearly steady then 2 pumps will be better.

If flow rate can be lower for some time then investigate whether 3 pumps will be better for handling lower flows such that one or 2 pumps can be switched off.

The velocity limitations are mostly to prevent excessive erosion for continuously running systems as well as to limit the required pump power and NPSHA. For non-continuous systems there can be certain relaxations of these criteria if this can be justified from cost point of view.

Now! Or it may be too late.

Hi,
You may want to take a look at the document attached

Pierre