Flow surging with a submersible pump 5

[Marke]

Hi

I have a bit of a problem that I can not explain.

We are test pumping a new well.
The well depth is 230M and the static water level is 113M below the surface.
We have an 8 stage submersible pump at 150M below the surface and are trying to run at different levels of constant flow for well analysis.
We are using a VFD to control the pump speed and this is operating in a closed loop setup with an electromagnetic flow meter being the feedback source.
We specify the desired or target flow and the VFD regulates the speed to keep that flow. The flow and water depth are logged over a period or hours at five different flow levels and the water engineers then use these figures to determine the performance of the well and either approve or dissallow the use of the well.
We have used this test set up for severla years without any issues.
The water level is monitored by the use of a submersible pressure transducer mounted on the riser above the pump.

In this pparticular well, we expect to be able to run well in excess of 40 litres pers second, but in operation, if we operate at 20 litres or below, the well performs as expected.
If we increase the flow to around 23 litres per second or more, we have a cyclic problem. Approximately every 20 minutes, the water level drops and the flow drops.
At a target flow setting of 25 litres per second, we run at a constant speed of 43.6 Hz, the water above the pump is 15.4Meters. Then every 20 minutes, the level drops to around 9 meters above the pump and appears turbulent and the flow drops to around 20 litres pers second. The VFD slowly accelerates to 48 Hz and the flow does not change with the speed increase. (I have limited the maximum speed to 48Hz). All of a sudden, the water level jumps up to around 13 M above the pump and the flow jumps up to 26 - 27 litres. The VFD pulls back to 43.6 Hz and the flow and levels stablize as before. Shortly after, we get very airated water appearing at the top of the well which is obviously associated with the problem.
I have set the VFD for constant speed and monitored the flow and level and the same issue occurs. It always corrects itself, the problem is probably there fo around 2 minutes every 20.
At greater than 10 Meters for water above the pump, I would not expect a vortexed drawdown of air into the pump inlet, or am I wrong here? It seems that every twenty odd minutes, the well is producing pockets of air from the bottom and limiting the water flow below 22 litres per second.
The client is looking for 40 litres per second, at this stage, I am not convinced that a pump at a lower level will be able to produce this.
Any thoughts?
Best regards,
Mark

Mark Empson
Advanced Motor Control Ltd

 

[bimr]

You have several things going on.

You are drawing down the well from the static level of 113 m to 221 m.

Your testing is showing that your desired pumping rate is greater than the yield of the well.

Note that as you draw the well down, your well pump will have to pump at a higher head. The pump discharge head is equal to the inlet pressure plus the pump head. The inlet pressure is decreasing as the water level drops. This will cause the centrifugal pump performance to slide up on the pump performance curve and give less capacity at higher head.

It is probably not vortexing in the well column, it is just drawing down the water level and allowing air to enter the pump.

Check with the pump manufacturer and determine how much water column you should have on the pump. I would expect that you need more than 9 meters.

Basically what is happening:

1. You are overpumping the well.

2. As the water level drops, the VFD will have to increase the pumping rate to maintain the setpoint.

3. Since you have limited the VFD setpoint, you are allowing the well to recover (with the VFD setpoint you are pumping at lower flow rate than the desired flow rate).

4. The cycle then repeats.

 

[MikeHalloran]

I'd say the well is telling you it has a capacity of 20 to 22 l/s, and no more.

When you attempt to extract more than that, the reservoir can't fill the well fast enough to keep the water level up, so the water level above the pump slowly falls until the pump starts pulling in air, then the pump operating point takes a big jump because of the reduced density.



Mike Halloran
Pembroke Pines, FL, USA

 

[cvg]

assume:

static water level= -113
pump suction inlet = -150
bottom of well = -230
minimum pumping water level = -141

you might consider lowering the pump closer to the bottom of the well in order to better test the drawdown. Is the bottom 80 meters of the well screened?

 

[Marke]

Hello bimr

Thank you for your comments.
We are not drawing the level down to 221M as the pump is at 150M. we are drawing the level down to 150 + 14 = 136M. - an approximate draw down of 23M, so 1 meter per litre under normal operation.

We just seem to have this very rapid drop in level and flow that occurs for about 1 minute every 20 and it is very regular.

From my perspective, it appears that there is a burst of air comming in through the bottom of the well, effectively the water inflow is replaced by air inflow, dropping the level of the water and causing the pump to pump less water. It appears that air in the pump affectss it's performance as expected.
The alternative scenario is that the inflow is restricted for this short time and the water level drop (to around 7 - 9 meters above the pump inlet) allows air to be drawn down from above.
I would not have thought that likely at 7 - 9 M of water above the pump?

I do not think that it is a case of a constant limited flow with recharge, as the timeing is exactly the same whether we are pumping 21 litres or 25 litres. I recovers very quickly, the flow jumps from 20 litres or so to 26 - 27 in about two seconds and the water level bounces back up again just as quickly. The chages are very dramatic and quick.

MikeHalloran. Thank you for your comments, yes we may have to limit the draw to around 21 litres, but it is a very strange situation that I have not seen before. Over a twelve hour period, the pattern is extremely consistent.
My concern is whether the air is comming from above or below. If from above, then dropping the pump lower will overcome this, but if from below, it will not help. This is a test pump to get an indication of the inflow and drawdown of the well and the behaviour has the local experts perplexed.

Hello cvg.
Yes, dropping the pump lower will alter things and allow for a greater drawdown, but this is a test rig so we are limited to the capacity of the pump and motor.
The wet end has just been refurbished, just in case it was an issue, and the pump is capable of around 45 litres at this depth so it should not be a capacity issue.
To drop further we would need to purchase an new motor and pump which is off the budget as a test unit.

Best regards,
Mark

Mark Empson
Advanced Motor Control Ltd

 

[DubMac]

This may sound crazy but I remember hearing about the practice of dropping ping-pong balls into the backside (annulus) to stop air entrainment problems in certain water wells. The thinking was that as water inflow fell from perforations above the pump and splashed into the water level, it was entraining a lot of air into the pump.

The ping-pong balls would cushion the fall and keep air from entraining.

I just remember hearing about this many years ago at Worthington Pump Co.

Not sure if it relates to your problem, but I would like to know if anyone has heard of this practice??? I know I've seen crazier things crammed down wells in the oilpatch.

 

[MikeHalloran]

If the gas were coming from below, it likely wouldn't be air, so if you can separate it, you might be able to light it, or at least have it do something funky to a flame.

I don't think it's that. I think that you are indeed pulling a vortex down 9+ meters, which is how long it takes to get the entire water column above the pump spinning fast enough for it to happen.
A vortex breaker should therefore at least change the period of the instability. I don't think it will get you to 40 l/s.



Mike Halloran
Pembroke Pines, FL, USA

 

[cvg]

there is no air coming in from the bottom of the well. even if this was theoretically possible, it would not come in regular 20 minute increments. I believe BIMR is on the right track. Not sure what your scope of work is, but testing and documenting the well flow and drawdown might not be very convincing to your client if you cannot establish equilibrium at your design flow and test from approximately the proposed depth of the pump.

 

[cvg]

whats the casing diameter and pump / discharge piping diameter?

 

[bimr]

As I posted, you need to check with the pump manufacturer and determine how much water column you should have on the pump. I would expect that you need more than 9 meters.

Not sure what pump you have, but I took the trouble to look up a typical Grundfos multistage pump performance curve.

Note that the NPSHr on the attached curve may be more than 9 meters.

http://www.leesupply.com/userfiles/pdfs/grndfs_6_8_10_subpumps_perfcrvetechdataLSPTL030_sec4.pdf

It is obvious that your pump setting depth is too shallow. This is causing your pump to cavitate and it will eventually destroy the pump.

You may be able to set the pump deeper. The pump setting depth must allow for adequate NPSH below the drawdown level..

 

[Marke]

Thank you all for your comments, most helpfull.

bmir
We made contact with the local office of the pump manufacturer very early on and they confirmed that there would not be a problem provided that the top of the pump was covered, suggestion being that a minimum water level of 1M above the top of the pump was very acceptable. They have been to site and noot expressed any concern about what is happening from the point ov view of the pump usage. They have commented that the pump performance is dropping off when it gets air in it, but no suggestion of air coming from above.

I have found reference in the manufacturers engineering manual, to 10M of water required above the pump. There are many pumps operating in this area with the protection set to only 2 or 3 meters above the pump.
Today, we were advised of three other wells in the neighbourhood that are producing airated water, so I suspect they are having the same issues.

From you comments and observations, it would appear that the inflow into the well is capable of much greater than 25 liters/second, but there is a periodic limitation of the inflow to around 20 minutes which is very cyclic and repeatible but for a short period of time.
After 24 hours of continuous operation, the drawdown at the problem time has reduced suggesting that the temporary inflow limitation is actually increasing with time. - this is a new well and it is not uncommon for the maximum inflow to increase with usage.

I think that the mechanism is when the inflow is restricted, the shortfall is taken from above the pump and because the water velocity down the side of the pump is high (the reduction in level is almost instantaneous), and the water level above the pump is reducing, a vortex is developed causing air to be drawn down into the pump. The inflow increases and the air is flushed through the pump and everything returns to normal untill the next event.
The vortexing is only happening as a result of the high velocity down the side of the pump and the reduced water level above the pump.

An interesting problem!!

Thank you all for your help.
Best regards,
Mark

Mark Empson
Advanced Motor Control Ltd

 

[cvg]

high velocity down the side of the pump...
that is why I asked about the casing and pump diameter. The well casing has to be large enough diameter to allow unrestricted flow into the pump suction inlet from both above and below. This is especially important if only part of the well casing is screened.

 

[bimr]

Not sure what training the local office has, but 1 m of water above a medium size pump like you have is inadequate. Of course, they will pleased to sell you replacement components when the pump fails prematurely.

The paper in the link is actually proposing a safety margin of 1.5 meters above the NSPHr or 1.35 times the NPSHr which is greater.

http://turbolab.tamu.edu/uploads/files/papers/p25/P25-Special1.pdf

One would expect that the pump will cavitate prior to a vortex occuring. In any case, there is a potential for damage to the pump.

One final point is that you may have dissolved gases coming out of solution. In particular, dissolved carbon dioxide if present will come of solution at low pressures. If this is occuring, there is a potential for scaling to occur on the pump.

 

[Marke]

Hello cvg

Thank you for your comments.
The well casing is 12 inch and the pump diameter is 8 inch.

bimr, that paper looks interesting, it is going to need a quiet time to properly digest, but the bottom line appears to be that we need more water above the pump, and we have a variable flow into the well.

Thanks everybody, this has been a great help, we are now wiser for the next "different" installation!!

Mark Empson
Advanced Motor Control Ltd

 

[SteveWag]

I have had this problem when the water bearing aquifer is above the water level. Water cascading down the dry upper portion of the well carries air with it. This air enters the pump, pump flow drops, well level rises, covers the exposed aquifer and the cycle repeats. This is a real problem in a siphon system.
Steve