Pump hystersis! 1

[tronical]

Hello to All Pump specialists,

We are on a mining project, and the site is filled with water. We have to dewater the site before mining by drilling wells and installing deepwell pumps. We started a one-pump dewatering test, and we have experienced a very funny behavior that we cannot explain:

As we ramp up the VFD that is driving the pump, we increase pump capacity. Everything was working fine until we reached a frequency of about 40Hz. At 40 Hz, our capacity was up to 3000 m3/hr. after we increase beyond 40 Hz, pump capacity starts to drop. This 3000 peak limit is 500 m3/hr less then the pump vendor designed the peak for (peak should be 3500).

When the decrease in pump performance unexpectedly occurred as we kept increasing the frequency, we decided to repeat the test. At the repeat, same thing happened at 40 Hz, so we decided to decrease the frequency. HOWEVER, as we decreased the frequency, the pump capacity never recovered to the same flows as before. If you envision a plot of flow (y-axis) vs. frequency (x axis), you get a hysteresis curve, where flow goes up until it reaches (40Hz,3000m3/hr), and then loops back to the origin with a different curve path (lower flows).

When we switch off the pump and repeat the test, we get the same hysteresis again as we ramp the frequency up and back down.

The well is 240 meters deep, so the pump is well beyond the 40m minimum submergence needed. We ensured there is no blockage. We checked all these factors to see if cavitation is the problem, and found nothing. Even if it was NPSH, usually when cavitation happens the pump is expected to operate normally when you ramp back down. This hysteresis behavior is unexplainable at the moment. We tested the discharge during the test to see fines/particle effects. All solids were fine particles that are within acceptable limits.

Can anyone exlpain the hysteresis? and why we experience premature peak flow?

 

[electricpete]

Perhaps as you increase flow you draw a vacuum on the suction which drags air in through some leak path. Once the air is drawn into the pump it continues to interfere with the pumping. When the pump is secured and restarted somehow the air is removed.

Just a thought

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[BigInch]

If you have a high specific speed high efficiency centrifugal, Consider the results reported by Kaupert & Staubli (1999) where changes occur in the useful output power the flow receives in the hysteresis flow regime, as the impeller flow recirculation and prerotation begins and ends with throttle adjustments.

Hysteresis in a pump characteristic results from instability phenomena involving complex three dimensional flow with recirculation. The unsteady flow field on the top and bottom branches of a hysteresis loop in a high specific speed (?S 5 1.7) centrifugal pump characteristic was experimentally evaluated. A hypothesis for recirculation zones and prerotation as power dissipaters is proposed for explaining the discrepancy in the pressure and shaft power hysteresis.

Theoretically the rate of change of useful hydraulic power in the hysteresis regime during transient pump operation was found to be a function of throttling rate. Quasi-steady behavior existed for slow throttling, |df/dt| , 0.005 s21. A second-order nonlinear dependence on the throttle rate was determined for the change of useful flow power during the commencement/cessation of the impeller recirculation.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[electricpete]

So you set up some unhealthy flow pattern at a given speed and once you establish it it doesn't go away with decreasing speed?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Artisi]

It is assumed you are talking about a pump installed in a deep well.

just something to think about, is the well capable of more than 3000 m3/hr - possibly you are trying to exceed the inflow capacity to the well, this could account for the decreasing flow as you increase speed. You could be pumping to the intake level of the pump within the well and entraining air, this could explain the pumps behaviour.

Naresuan University
Phitsanulok
Thailand

 

[BigInch]

EPete, It seems to fit with the OPs experience, with the 40 m submergence criteria he mentions, I gave a low chance it that it was related to air or NPSH, but perhaps that probability should have been even higher. But, yes exactly. Its claimed that a prerotation condition can be set up in some pumps which will remain in effect, even though throttle setting is reduced, until a recovery point is eventually reached. It is also related to high efficiency designs. The study further warns that this tendency towards prerotation should be evaluated against the increased cost of purchasing a high specific speed - high efficiency unit. As prerot set-up appears to be related to rate of speed change towards the operating point, maybe the OP could do some experimentation with various speed advance rates to see if there is any change in the effects he writes about now.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[katmar]

BigInch, the pre-rotation phenomenon would explain the hysteresis, but does it explain the drop off in capacity as the VSD frequency increases above 40 Hz? The pre-rotation idea seems the most plausible explanation of the hysteresis, but could it be occurring in combination with some other problem? I would have guessed that the pump was sucking air, but this is unlikely given the submergence.

Tronical, do you observe air in the discharge when the capacity starts to fall off?

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[BigInch]

Possibly, as the prerotation energy must be drawn from somewhere.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[EdStainless]

What happens if you start the pump at 50Hz? Will the flow start out at the proper design level and then drop off? It really sounds like the pump is intake starved. Do you have sensors to measure downhole pressure?
What head is the pump designed for?
What pattern do you see in motor current as you increase the freq?

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

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

 

[tronical]

Hello guys,

Sorry for building up your questions..had to be at site.


The pump is completely submerged in a deepwell. As soon as pump performance drops the level of the well starts to rise. The lack of intake is not an issue at all, as the pump is submerged in a deep deep well. The water table is 240m above the pump. As we pump up and reach peak (premature peak that is) flow, the level in the well dips down to about 150m (will verify not too sure). This is still way more than we need. As for the net inflow of water, water is very very abundant. We can add another 15 pumps and pump away for 15 years before we dewater the pit.

We tested for offgassing. nothing. we had valves to test for offgassing on the top of the well, and different high points along the pipeline, and we observe no offgassing nor bubbling. there is no air in the discharge.

Once the unhealthy flow pattern starts, it does not go away until the pump is completely shut off. Again, there is no air in the discharge.

BigInch your pre-rotation suggestion is yet to be looked at in detail. I obtained the paper you refered me to...just didnt get enough time to read it, but it seems like it is worthwhile.

Edstainless, we tried to ramp up quickly to 60 Hz, we still see the same problem as we exceed 40.

How do I post pictures/pdf here? I will try to post a curve on monday/tuesday when I get back to office.

 

[BigInch]

you can post pictures using the " tag.

1.) Upload a picture to some site like [URL unfurl="true"]http://imagevenue.com[/URL]
2.) copy the URL link for the picture that they return,

[URL unfurl="true"]http://img14.imagevenue.com/aAfkjfp01fo1i-11989/loc567/97449_Al_Hambra_122_567lo.JPG[/URL]

3.) paste the URL link into this thread box, as above
4.) add "[img " to the front of the pasted URL (no "")
5.) add "" to the end of the pasted URL (no "")

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[Artisi]

A bit more detail would be interesting for this discussion.
1. when the pump reaches its premature peak, will it maintain that flow if you maintain the 40hz speed ie, for how long have you run the pump at 40hz.
2. what is the relation between the screen diameter/well liner diameter and the pump/rising main diameters.
3. where is the well screen positioned in relation to the pump inlet.
4. what is the recovery time after stopping the pump for the water level to recover from 150 to 240 metres.

If, as you say there is enough water to keep 15 pumps running for 15 years it would be expected that with only one pump operating in a well, capable of the flow rate you are expecting, not to see any drawn-down in the well casing. Dropping the water level in the well casing by 90 meters would indicate that the inflow to the "well" is less than the pump capability.

Naresuan University
Phitsanulok
Thailand

 

[tronical]

Hi guys,

We pulled out the pump and found rocks clogging the inlet. The rocks were all beyond 0.5 inches, which is why wee didnt see any solids in the discharge. The rocks were getting sucked in only when we had enough pressure (at 40hz), at which point they would get sucked and stuck in the inlet screen. Them getting stuck explained the premature peak flow. When they got stuck, a reduction in frequency still didnt allow the rocks to get unstuck, which explained the hysteresis. Only a complete shutdown would allow the rocks to fall off the screen, which is why we were getting essentially a carbon copy repeat-test results, and which is why when we flushed the pump and repeated, it didnt solve the problem.

The whole well is supposed to be guarded with an inlet screen to prohibit rocks greater than 0.5 inches from entering the well. Clearly somewhere a screen broke. We have not yet located the source of these rocks, but it is no longer a pump question.

Thank you all. Hopefully its an incident that you wont come across yourself, and if you do, remember this discussion!

 

[Artisi]

Just reinforces my belief, that the majority of "pump problems" are not really pump problems but installation / application problems.
Anyway, a good result for you and it's appreciated that you took the time to get back to us - something that doesn't always happen.

Naresuan University
Phitsanulok
Thailand

 

[tronical]

Getting back to you was the least I could do. I really appreciate everyone's help honestly; after I discovered it was rocks the first thing i thought of was "what am i gonna tell my buddies at eng-tips? they sure are expecting a more interesting failure (maybe I shold cause one lol)"

Honestly I appreciated your inputs greatly, and eitherway your inputs made me as well as others learn, even though my case was something else.

tronical

 

[thewellguy]

Depending on you location, there are companies that have video cameras taht you can lower down the well and see what is going on. Look for a large weel driller in your area or I could try and reccomend one if your in the states. Dependong on what type of pump (submersible/turbine) you may be able to put a flow sleeve over the pump to distribute the suction area and reduce the likelyhood of sucking up rocks.

 

[BigInch]

Rocks. They'll do it every time.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com