Does direct recycling to suction improve NPSHa? 2

[hollerg]

I am short of NPSHa. I have heard that recycling some of the discharge flow to the suction of the pump may help.

I can't halt production to address root causes. I cannot cool the suction line, because the product is molten and only ~ 3-5 C above the melt point. Pump cannot be adapted to an inducer.

Does anyone have a reference or specifics? I think it was in Hydrocarbon Processing or Chemical Process or Chemical Engineering, but I can't find it. The recommendations were only a part of the article on improving pump operation. I cannot recall how to introduce the fluid back to the suction line, the proportion of the discharge flow (10%?) or whether there is a way to predict the performance gain.

FYI -- Googling found a patented (US 6,517,309)
suction side device that makes a centrifugal pump, operate more like a multiphase pump, but this was not the article I was thinking of.

The recycle to the invention reduces NPSHr.

http://www.pumpcavitation.co.nz/Zaher-abstract/Paper8.htm

The invention seems to use the recycle to entrain the flow from the source tank, boosting the pressure into the pump. EG a transfer of mechanical energy from the recycle flow to the source tank flow, I guess like an eductor.

 

[UmeshMathur]

hollerg:

Pumping a fluid raises its enthalpy in proportion to the pump power, so simply recycling the discharge back to suction pressure could risk flashing some of the incoming liquid, making an undesirable situation even worse. If you could cool the recycle stream sufficiently, that might lower mixed inlet temperature to a safe value, lowering vapor pressure and reducing or eliminating the extent of flashing.

Finally, it seems to me that using a venturi on the recycle line to help suck fluid into the pump might work, if (a) the venturi is designed properly, and (b) the discharge pressure before entering the pump is sufficiently high to collapse any vapor bubbles. However, the required recycle flow could be very large and might chew up a lot of the pumping capacity, which could defeat the whole idea.

Performing this analysis reliably would require accurate physical property data (heat capacities, latent heats, vapor pressures, etc.) for your molten material, not always easily available.

What do you think?

 

[25362]

From what can be inferred from publications, the subject of fluid prerotation (as caused by internal recirculation, by the turning of the impeller, or by external means) and its effects on NPSH are still being studied.

The Pump Handbook by Karassik et al. mentions an article on jet inducers, that appeared in Cavitation in Fluid Machinery, Symposium Publication, ASME, 1965, p. 109.

The article was written by R.A. Lewis and deals with an inducer consisting of one or more liquid jets directed axially toward the eye of the impeller. The liquid for the jets is supplied through a by-pass line from the pump discharge.

Another included reference is H.H. Anderson's Centrifugal Pumps, Trade and Technical Press, Surrey, England, 1980, pp. 69-70.

 

[Artisi]

I cannot see any logic to your proposal unless of course you are already by-passing some of the pumped product, however, if this is the case you would be better off reducing the thru-put and lowering NPSHr.

Recycling pumped product back to the suction means increasing the flow thru' the pump which requires a higher NPSHr - so a self defeating exercise.

following on from the above, if you can afford to recycle some of the pumped liquor back to the inlet this seems to indicate that you can tolerate a lower flow rate, if you can afford a lower flow rate - reduce the pump thru-put and lower NPSHr.

Naresuan University
Phitsanulok
Thailand

 

[25362]

Not necessarily if the patented small recycle fixture indeed moves further off the 1% (or 3%) head drop of the pump for incipient cavitation to a flow rate greater than the combined flow rates.

 

[EdStainless]

You are only a few degrees above melting point and you are cavitating? How far off of BEP are you?

Is it possible that your inlet temp is lower than it used to be? Is increased viscosity and higher pressure drop in the inlet part of your problem?

= = = = = = = = = = = = = = = = = = = =
Corrosion, every where, all the time.
Manage it or it will manage you.

http://www.trent-tube.com/contact/Tech_Assist.cfm

 

[checman]

hollerg: A similar set up is used all over the world to increase the lift capability of shallow well pumps. I have done it myself on process pumps and did get the results I was looking for. You must have a good start-up procedure and a well thought injector. You must be careful because you can and probably will to some extent cavitate at the injector.

Regards Checman

 

[vanstoja]

I don't understand how pumping cavitation performance is an issue for a molten material that is only slightly above its melting point where two-phase flow is more likely to be liquid-solid rather than liquid-gas which is the normal province of cavitation phenomena involving impeller surface damage and possibly liquid-gas two phase flow instability. How far is your molten material from its vapor point temperature? I don't think you have a liquid-vapor cavitation problem so concerns about NPSH margins seem to be totally irrelevant. Where did liquid cavitation NPSH data for your material (which I presume is not liquid water)
come from?

 

[hollerg]

This is an evaporator system has run this way for a number of years, before I arrived. I am removing moisture from the molten material. VLE shows that as I get to the 0.5 % water level, about 5% of the vapor is the organic. This is not a polymer. I runs adequately with excessive NPSHa, but unless there are "temporary fixes" between now and the next outage, I cannot improve product quality to meet customer demands. To do that I need to reduce residence time which means that the vessel inventory must be lowered.

The installation: 3x4x13 @ 1150 rpm pump, NPSHr = 3ft, 4" inlet line 8 ft long, Normally liquid level is 6 ft above suction line. I am operating just to the left of the BEP, in an evaporator bottoms pump service. Being Hastelloy and glass I am trying to avoid costly or lengthy changes in a tight capital environment. I don't believe that the inlet line has any significant cooling. It is insulated and the flow is 250 gpm, most of which recycles through the evaporator and back to the pump.

There is a strong vortex (actually observed) which I have just added a vortex breaker for. I am just now putting it in service and have no results heat

I have a sever restriction in the pump suction line. My calcs show that the 3" nozzle on the glass lined vessel is a killer on the suction side pressure drop. I have read that at higher temperatures there is a favorable shift in the cavitation index, meaning lower NPSHr, so I might get lucky,

http://caltechbook.library.caltech.edu/22/01/chap7.htm

but if recycling would help too, I would feel much more comfortable.

 

[PUMPDESIGNER]

I have the same questions as vanstoja.
More succinctly, how did you determine that NPSHa is low?
Low NPSHa is a diagnosis, not a report of symptoms.
I think most of us would prefer to know symptoms and how you conclude low NPSHa is the cause of those symptoms.

PUMPDESIGNER

 

[25362]

If I may add: is the fluid Newtonian, is the flow in the suction piping laminar?

In laminar flow [Δ]P[sub]f[/sub] = 32[μ]LV/D[sup]2[/sup], stressing the importance of the viscosity [μ] on the friction drop.
For a Newtonian fluid the viscosity is constant, it isn't so for non-Newtonian fluids.

 

[hollerg]

Summary of all details leading to lack of NPSHa conclusion follows. However I am not asking for others to confirm that I have an issue. I know I have a mismatch between calculated NPSHa and the pump NPSHr at my new conditions. I am trying to make a regular pump work where an inducer or elevating the vessel is normally required. Can't afford to change these things at this time, as previously stated.

Process Intent of change:
Drop the fluid level from 6 ft down to 3 ft of static head above centerline of pump, reducing product residence time in suction side vessel.

System details
Fluid: chlorinated organic acid & residual water
Flow 250 gpm
Pump is inline 3x4x13 (12.5" imp)@1150 rpm, Hast C
NPSHr is 3-4 ft, w/ no safety factor
Fluid is @ the bubble point (80 mmHga @ 240 F),
At bubble point because boiling off water and some organic.
Fluid density of 84 lb/cuft,
visc of 18 cP.
Lines are Hast C, elect traced and insulated

1 ft 3" inlet nozzle, 1 ft below surface of fluid

4 ft of 6" (6x3 & 6x4 reds, two elb, (1.5' in vertical)one tee in branch)

3.5 ft of 4" with regular port ball valve

Re & vel data
3" noz on vessel Re# 18.4K @ 9.7 ft/s
6" line, Re# 9.4K @ 2.5 ft/s
4" line, Re# 13.9K @ 5.6 ft/s

Do not confuse the term molten to mean no vapor pressure. Being only 5-9 F above melt point does not mean it does not have a vapor pressure (i.e. water at 36 degrees or sulfur). It also does not mean I must have solids. There are no solids in the samples. I am using the term molten to mean solid at room temperature.

Again, your help in finding the literature references about the impact of directly recycling a portion of fluid from the discharge of the pump to the suction would be greatly appreciated. I have seen it someplace.

 

[25362]

Is it possible to pressurize rhe suction vessel with an inert gas such as nitrogen ?

 

[rmw]

Regarding your initial question, in order to maintain your process flow out of the evaporator vessel, you would have to carefully balance the pressure drop of the recycle (to the pump suction-not the evaporator) so that the contribution of the recycle flow to the pump suction did not not 'back out' any of the process flow from the vessel and hence increase your residence time, and in the process all you would accomplish would be to move your pump farther out to the right on its curve, where the NPSHr is most probably increasing anyway. (Comment re NPSHr curve made without knowing what your NPSHr curve looks like, but you have already stated that you are close or worse.)

Sounds like a dog chasing its tail to me, to put it into technical terms.

Since your goal is to lower the level inside the vessel, it appears obvious that some piping modifications are going to have to be made. As I read it, it is unclear to me whether the 3" piping is in a stand pipe inside the vessel so as to determine the level or is at a nozzle on the side of the vessel?

Whichever is correct, then when you lower the level by 3 ft, it appears that you are gong to have to modify some piping and getting rid of the 3" piping (which I assume is there as a flow restriction of some type) is essential, and upsizing the balance of the piping to the extent possible from the vessel to the pump is about all you can do. You can control the flow with downstream devices.

If you can remove 3 ft of piping losses from the suction piping and valving at the flow rate at which you wish to operate, then you can lower the level by that amount. Other than that, raise the vessel, or get another pump with a lower NPSHr.

rmw

 

[rzrbk]

Can't help with the article, but per the attachment to your original post, it seems like the intent was to spin the fluid in the suction line to more closely match rotation of the impeller. I can see this causing less required acceleration at the impeller eye as the fluid hits the vane inlet and tries to match speed; resulting in less NPSHr. It seems like it might work in an end suction pump, but be less effective in other types where the suction path twists and turns before it reaches the impeller.

Other than lowering the level till cavitation approaches unnacceptable and living with it, the only other suggestion I can think of not already discussed would still require a short shutdown. This would be to install a baffle (or inert media with a suction standpipe/sleeve)in the vessel to reduce the liquid volume but still allow the 6' level above the pump.

 

[Artisi]

" I can see this causing less required acceleration at the impeller eye as the fluid hits the vane inlet and tries to match speed; resulting in less NPSHr. "


Pre-rotation is not a component of NPSHa or NPSHr. It is a commom but erroneous belief that pre-rotation causes a change in NPSHr when in fact the effect of pre-rotation is a miss-match of the inlet flow and the impeller blade causing a drop off of performance independant of NSPHr/a.

You can have a situation where you have plenty of margin between NPSHa /r but the inlet flow is pre-rotation resulting in inlet side problems which has nothing to do with NPSH.

For the problem as reported, certainly straighten the inlet flow by what-ever means suits the application and run the pump at the point of minimum NPSHr.

Naresuan University
Phitsanulok
Thailand

 

[25362]

This is what I could learn from specialized books.

Changing prerotation, as with guides, imparts a circumferential component to the inflowing fluid, thus altering the energy conversion process in the pump and with it its characteristic.

However, the so-called prerotation NPSH-reduction and control range is unstable and applies to a narrower band of flowrates than with impeller blade adjustment, and therefore, it is even more susceptible to cavitation !

 

[checman]

I have found that the impeller design imparts the correct rotation by design and fluid going to the pump should be free of rotation for the most part when approching the pump. The recirculation line is not what normally comes to mind. It should extend into the pipe and discharge towards the pumps suction, taking full advantage of the velocity head of the bypass. I have used a 90 degree elbow on a nipple that was welded in the suction pipe perpendicular to the pipe. The sizing and placement of all the components requires some thought and is installation and process specific.

Always outside the box!
Checman

 

[moltenmetal]

Are you running the pump "flat out" on the discharge, or is there a flow control valve there? If there is, consider opening this valve wide and installing a variable speed drive on the pump, spinning it only as fast as required to obtain the necessary discharge head/flow. Reducing pump speed should reduce NPSHr to obtain the required discharge head. Plus you'll save some electricity.

 

[ChickenSoop]

Some options to check out:
- does the pump supplier offer a flow inducer to reduce NPSHr?
- investigate a larger size suction pipe to reduce friction losses, remove any screens on the suction line.
- install max sized impeller, and use VFd to slow down pump
- insure you are operating on the pump curve and not beyond run-out. Maybe valve off the discharge a bit to raise NPSHr if you can live with a bit less flow.
- if above doesn't work out, consider a new pump - a larger pump and run it even slower

forget about fance gizmos or recirculating anything.