Straight Runs for Pumps

For a non-self-priming pump, provided it has a net positive suction head [NPSH] then it doesn't matter.

The downstream pipework shape is immaterial. The pump will operate simply where its own characteristic curve and the system characteristic curve intersect.

It is good practice to install strainer facilities upstream of a pump, and place that equipment on planned preventative maintenance at suitable intervals.

Agreed, this is one area I insist upon (despite the opinion of others) - not to put strainers in suction line, unless you can monitor the pressure there.

Usually our pumps have NPSH (the pump level is kept below the tank oil level) and hence no pressure monitor is kept at suction side.

Several persons have suggested the strainer at inlet. but we have put it at return oil line where we can monitor rather than at pump suction.

Shanii- As far as bendes etc are concerned, since the suction lines are much larger than the dely line, and usually the pumps are kept flooded, and usually the flow velocity (rather the Re) will not be so high so as to have any significant effect.

The 5D min rule is applied to have a laminar stream lined flow.

In our systems say 150LPM pump, the line just comes out of the tank, one 3" elbow (LR) + Tee (for splitting to two pumps- one stand by ) - butterfly valve - pump inlet. there is no possibility (no space) for the 5D.

At outlet also, the additional losses in head due to the short pipe is taken care by the pump. Here too we have space limitation. So practically we do not use the 5D rules except in some areas (espacially where the instrumentation pick-ups, relief valves etc are connected)

sb

I think you are referring to a filter if you are putting it into the "return oil" line. For a process type centrifugal pump a strainer typically is there to catch nuts and bolts. Strainer holes could be 6 to 8 mm diameter with a free area of roughly double the suction pipe area. It is installed to provide protection of the pump from large foreign object damage. I think that this is what was meant.

I agree whole heartedly that unless there are very good reasons, and then only if Dp gauges can be fitted, filters should not be put into the suction line.

The suction pipe is usually bigger than the pump inlet (to minimize losses) the eccentric reducer (flat side on top) is absolutely necessary.

Except for turbulence the pump would not know what fittings are on the suction side.

If proper NPSH calculations is done bend etc would be in order.

Please describe your situation in more detail and better advice may be offered.

Actually i just want to know that straight run may or may not contain fittings, valves or strainers if they are arranged inline. In other words if 5D run is coverd in valve strainer and flanges,shouldn't i use pipe as straight run then?

See #1, #2 and #4.

The shape of the pipe continues to be irrelevant provided there is a NPSH!

This question is too broad.("Behind the detail is the detail")

What type of pump are you asking about? (Centrifugal, Positive Displacement, Sump, Lift, etc.)

What is the pump nozzle configuration? (end suction/top discharge, top suct./top disch, side suct./side disch., inline, etc)

What is the commodity?

What is the suction line and pump size?

Where is the suction source (Tank) relative to the pump?

What is the configuration and length of the suction line?

That is only to name a few of the details that effect a pump suction line.

You really need to supply more detail. If you are working in mining etc you may have some code to conform with.

I have seen 5 x diam used with meters in mining code. A meter is very susceptible to turbulence and 5 x diam is suggested between the pump (centrifugal) and the meter on the suction side and 10 x diam between strainer and meter.

For most centrifugal pump applications the type and placing of fittings is not that critical.

Bends (even 90deg) are often used next to the eccentric without problems.

An isolating valve on the suction side should not be used for controlling the flow and if no restrictions are caused it should be ok..

A strainer on the suction side may also block up and should preferably only block objects that may block the impeller.

2D to 5D straight run in suction is a good practice. Elbows excluded. It creates turbulance. Valve length can be included. Exclude reducer also.

Five pipe diameters straight run length in the suction line to almost any pump is good engineering practice. Even with a good NPSHa margin it is possible to get flow turbulence caused by poor piping design that can be very destructive to pumps.

Hence the API 686 requirement:

The pump suction line shall have a straight run (typically five pipe diameters) between the suction flange and first elbow, tee, valve, reducer, permanent strainer, or other obstruction sufficient to ensure stable and uniform flow at the pump suction nozzle.
Note: A piping straight run length of five pipe diameters, based on the pump
suction nozzle size, is usually sufficient to ensure stable and uniform flow
at the pump suction nozzle.etc etc.

It is basically a good practice to have a straight run between the pump suction nozzle and the first fitting. The first fitting means the first welding point be it a elbow or a strainer. But in actual practice for positive displacement pumps 5D is not mandatory. it is feasible to have this straight run requirment only when the pump is of double suction type. you can refer the below site

http://www.directindustry.com/prod/goulds-pumps/double-suction-pump-for-high-temperature-process-20610-220177.html

Thanks,

Kumaran

The main objective of straight run in pump suction is to get stable flow.

May be this simple rule could answer your question (taken from Troubleshooting Centrifugal Pumps and Their Systems by Ron Palgrave).

Straight run is just be a pipe, not included the valve or strainer.
At least 5D should be provided after bend/valve/strainer. If it not possible, consider for using cascade bend.

Cascade bend