Fitting distance from Pump 5

[mmmumuw]

I read from a pump manual that we should be avoid any fittings, including bends, compensators, valves in the area of 5 X DN suction and 5 X DN discharge.
I guess if before suction, the fitting will make the flow less smooth and can make cavitation. But what is the purpose of this 5 x DN?
Thanks alot,
MW

 

[Artisi]

You should avoid having any fittings that can cause unsteady flow into the impeller, that means that fittings should be placed well down stream from the inlet, the distance varies somewhat depending on the authority making the statement, although it seems to vary from 4 - 8 pipe diameters. The idea is to allow time for the flow to straighten and streamline before entering the impeller. Turbulent flow resulting from fittings does not cause cavitation - cavitation is a result of insufficient NSPHa, but fitting too close to the pump inlet can result in noisy operation because of poor flow approach to the impeller eye.
Any good hydraulic book will give you information on this subject.

Naresuan University
Phitsanulok
Thailand

 

[PUMPDESIGNER]

Check out this link for explanations and an awesome video showing the problems of incorrect pump connections.

PUMPDESIGNER

 

[PUMPDESIGNER]

Sorry, forgot the link:

http://www.irrigationcraft.com/flow_run.htm

PUMPDESIGNER

 

[mmmumuw]

What is the purpose of 5 X DN discharge?

 

[friartuck]

I used to work in fans and the principles are the same.

If you have bad inlet conditions, this alters the actual fan/pump curve. So the whole thing can be de-rated by possibly up to 30%. Thats why you should have a straight inlet. We tend top say about 1.5D upwards but 5D obviously better.

The flow leaving the fan/pump is swirling and if a fitting is introduced too close, the swirl will create a greater fittings loss hence the reason for a straight outlet.

The efectes of distance between pump/fan and fittings is handled in articles covered by ASHRAE (I think its called system effect or something very similar) There is a calculation procedure to take into account location of fitting near equipment.

The effect is more pronounced on fans as they only generate small pressures to start with and any loss causes problems. The effect is also worse with axial or tube axial fans and much less with backward curved impellers.

On this basis, pumps having backward curves would probably be fairly tolerant to having a bend close to the inlet or outlet. (I'm talking about small/medium heating and chilled water pumps and not monster high pressure pumps which might not like being 'strangled' in this way)

Friar Tuck of Sherwood

 

[scalleke]

Every time an impeller blade passes by the outlet opening of a pump there is a high pressure and alo pressure point.

High pressure occurs right prior to the blade passing the opening.

Low pressure occurs about when the mpeller blade has passed the discharge opening and the discharge opening is in the middle between two impeller blades.

These low and high pressure points will cause a discharge pressure wave pattern in the discharge piping.

The closer you have valves and other items are near the discharge of the pump the better these waves will be reflected and disturb the flow pattern of the liquid through the pump. You need distance between the pump and the first accessories such as valves etc so you get some mass of liquid that can somewhat attenuate this pressure wave.

5 x DN is a rule of thumb where things start to be OK. I have seen in the past that 10 x DN was recommended.

As you start discharge throttling down the flow throught a pump. The more you close your valve, the more your liquid gets trapped between pump discharge and discharge valve and the more the trapped liquid gets pressurized the more energy it contains, the better the pressure waves will start to resonate in this liquid, the more the pressure waves will start to influence the flow going through the pump. Finally flow will be disrupted and the pump will go in minimum flow.

The longer the distance between pump and valves, the better the waves will be attenuated and the longer it will take before the pump goes in min flow. (min flow will be lower). However the longer the distance, the bigger the mass of liquid trapped between pump and valve and the more energy the liquid in the discharge piping contains at min flow.....more energy....more damage at min flow.

Place valves etc as close to the pump as possible but respect the manufacturer's recommendations. Do a min flow test on the valve to determine where the min flow point is and make sure you do not get there. The pump manufacturer can help yoy with this.

If you do a min. flow test in the presence of the manufacturers rep you can even palace valves closer than the 5 x DN provided you ensure that the unit cannot operate at lower than min. flow.

Best Regards.

Scalleke

 

[scalleke]

Next time I will proof read prior to submitting. Sorry for the errors.

Scalleke

 

[PUMPDESIGNER]

Discharge has a slightly different reason for straight run, which we call a "flow run", but some pumps do not need that flow run, it really depends on the pump.

Here are some of the things to know:
If pump has a diffuser, then flow run may not be necessary.

Speaking specifically of pumps with Ns<6000, no diffuser, and discharge connection is close to the impeller, such as scroll type single and double volute pumps, they usually require some flow run.

Straight volute pumps (with no diffuser) have very high velocities at the discharge. Energy transfer from the pump shaft into the fluid is almost entirely by an increase in velocity, and then this high velocity energy must be converted into pressure energy by slowing the fluid down. Now the little known fact, this "slow down" or conversion of energy, from velocity to pressure, in many cases is accomplished after the fluid leaves the pump and enters the pipe system which is most often larger size than discharge nozzle, i.e. system resistance in the pipe system causes reduction in velocity. If you allow a fast turn just after pump discharge the high velocity fluid hits that ell or tee and can not only eat out that fitting, but also fluid may re-enter the discharge nozzle of the pump, called discharge recirculation and also discharge recirculation cavitation. Heck, even with a flow run you can have discharge recirculation at low flow rates.

A video for discharge recirculation at this link:

http://www.irrigationcraft.com/cavitation.htm

PUMPDESIGNER

 

[vanstoja]

The length of straight discharge piping needed downstream of a centrifugal pump depends on the quality of the flow (re axial velocity profile distortion and fluid rotation or swirl) exiting the casing discharge nozzle or flange. This is generally caused by secondary flows in the pump casing downstream of the diffusing section. Swirling discharge flow can result from such effects as oversized vaned diffusers blocking a portion of the usual tangential discharge nozzle resulting in a transverse secondary flow into the main discharge flow which produces two counterswirling vortices just like occurs in a 90 degree pipe bend. Unrecovered static head due to discharge velocity profile distortion can lead to incorrect pump head measurements if discharge pressure taps are located in swirling flowfields with high velocities(low pressures) at the pipewall. Downstream swing check valves in swirling flowfields have oscillating disks that can damage valve pivot hardware by repetitive backseat tapping and also cause surging in the discharge flow. Pipebends downstream of swirling pump discharge flows can sometimes intensify the swirl. It's hard to say how much run of straight discharge pipe is needed to dissipate axial velocity distortion or swirl without a measure of their degree. 5D of straight piping may be fine in most cases of minimual distortion/swirl but inadequate for other conditions.

 

[vanstoja]

Decay rates for swirling pipe flow and accompanying axial velocity profile offset have been determined experimentally and are reported in the the open literature. Most of the testing was done in Japan and reported in the Bulletin of the JSME by Murakami & Shimuzu(MS)(1973), Senoo & Nagata (SN)and others. Baker & Sayre (BS)in USA or Britain also did classic studies. Swirl momentum decay is an exponential fuction with a power exponent of -Beta x Number of pipe diameters where Beta depends on the friction of the inner pipewall and the viscosity of the fluid in the pipe. For water, BS found Beta to be in the range 0.02-0.07. MS found 0.037 and 0.155 for smooth and rough (c_f=0.02)wall pipe, respectively. SN found Beta to be between 0.037 and 0.04 for water with a pipewall c_f=0.12. Plotting swirl momemtum ratio vs pipe diameters, I got swirl ratios of 0.62 and 0.10 for MS smooth and rough pipe at 15 pipe diameters downstream of the swirl source. MS Beta values for decay of axial velocity profile offset(or skew) downstream of a single 90 degree bend was somewhat greater than 0.31 to reduce an initial offset of 0.2726 to zero at 8.5 pipe diameters downstream. For two 90 degree bends back to back, Beta values depended on spatial angle between bends with zero offset occuring at 9.2 diameters for a 90 degree angle and more than 11 diameters for a 45 degree angle. Such studies of flow disturbance (swirl and velocity distortion) decay rates have led to some very large numbers of pipe diameter separations for flow measuring instrumentation as well as for pumps in complex piping systems. Unfortunately, most of the pump gurus promoting low pipe diameter numbers have not read of these results and fail to warn of the consequences of pipe geometries and obstructions beyond the first upstream or downstream bend. This "system" problem should be addressed more critically by pump designers and particularly by pump users who are often victimized by defective system piping design practices or by unresolvable spacing constraints in the plant employing the pumps and their connecting piping.

 

[PUMPDESIGNER]

Hey vanstoja, nice posts.
Tired of architects assigning a closet for your stuff to go into? Have some fun, check out this link to a photograph our production manager made up:

http://www.irrigationcraft.com/download3.htm

PUMPDESIGNER

 

[Artisi]

PUMPDESIGNER
Seems like there is plenty of room left for expansion at some later date.


Naresuan University
Phitsanulok
Thailand

 

[PUMPDESIGNER]

Hello Artisi,
Complaints, complaints, we all got 'em.
So you are one of them witty guys too huh?
You would fit right in here. About five years ago that photograph appeared on my desk top while we were doing that project.

vanstoja, do you think your references supply enough information that a spreadsheet could be built to enable one to figure out intake and discharge piping minimums for scroll, center discharge, and turbine volutes? If so, is there a single source I could purchase those references from? I am in the middle of building a worksheet to calculate radial thrust (looking for more references on that too), but your references for decay rates look interesting for my next project possibly.

PUMPDESIGNER

 

[Artisi]

Hi PUMPENGINEER
Like you, I've been there and done that, it made life a little frustrating at times but interesting.
Nothing like an impossible problem to solve to keep you on your toes and prove the experts wrong.

Naresuan University
Phitsanulok
Thailand

 

[vanstoja]

Pump Designer
All the references I have seen on swirlflow are for piping mostly with 90 degree bends and do not address pump casing types or design features. I think the people now using pump CFD analysis will have to develop the data for pump casing discharge distortion control. We got into it experimentally in the 1970's to save downstream swing check valves hammering their disk bushings and pivot pins to destruction due to swirling/distorted pump discharge flows. We used 5-hole pitot tubes to measure discharge pipe velocity profiles in manufacturers water test loops and in an airtest rig using sheet metal ducting where we went inside the wooden pump casing to find secondary backflows from diffuser vane channels nearest the exit nozzle. What is a "turbine volute"?
On centrifugal pump impeller radial hydraulic thrust, I have many of the classic papers since the 1960's and some very limited test data but I think you should open up a separate thread on that subject.
Loved your picture of a typical pump piping shoehorn job.It seems like the pump piping is always the last gasp of the space cadets and the consequences can be monumentally bad. I suffered through several years of grief helping to resolve one such operational crisis.

 

[PUMPDESIGNER]

vanstoja, Sorry about "turbine volute", meant turbine bowl, as in multi-stage turbine bowls with diffusion vanes.
Will start a thread on radial thrust when I get the chance.

PUMPDESIGNER