Could distance from VFD to pump cause problems? 3

[powersoff]

Hello,

I have 5 well pumps ranging from 25HP-60HP. The 5 pumps are being fed from 5 VFD's. The closest VFD to a pump is 200 feet the farthest is 400 feet. The control of the pumps is from a pressure transducer the pumps try maintain a setpoint. Only 1 pump is on a PID, when called the other 4 run at 100% speed. The second pump will turn on if the PID pump runs at 100% speed for 1 minute and then 1 minute later pump #3 and so on until all 5 pumps are running. To shut off the reverse is true if PID pump falls below 95% speed for 1 minute it turns off pump 2 a minute later until all pumps off except PID pump.
There is also a low pressure cutout switch if PSI falls below 30 no pumps can run unless done manually.
Does anyone have any suggestions as to why 2 pumps have failed in the last month?

 

[gepman]

What do you mean by "pumps failed"? Do you mean damage to the pumps or damage to the motors or something else? Long distances between VFD and motor can cause problems with the "reflected wave phenomena" which has been discussed in many threads in this forum. Search the forum for those and other similar words.

 

[powersoff]

VFD would fault off.I had continuity between pump motor leads and ground(earth)

Thanks

 

[itsmoked]

Any chance the two that have failed were on the longest two cables?

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[SteveWag]

OK, when the pumps “failed” you had conductivity from the motor leads to ground. This means the motor insulation has been pierced. Find the Emerson submersible motor site on the web and look at their requirements for load side inductors for use with VFD’s. The fast rise times of common VFD (dv/dt) cause extremely high momentary voltages every 120th of a second on each phase wire. The load side inductors round the corners of the pulses and remove, or lessen the high voltage transients. The problem is at it’s worst at 480 volts and load conductors over 75 feet, getting worse as voltage increases and distance grows. With submersible well pumps, it is VERY easy to lose motors when load side inductors are omitted.
Steve

 

[Valvecrazy]

OK, I know I am going to catch it from the VFD guys here but, this is the kind of problem I deal with everyday. Yes the VFD is causing motor failures. The longer the wire, the higher the voltage spikes to the motor. Load side inductors only lessen the problem, and is just one more expense added to an already expensive and complicated installation. Motor manufacturers love Drives, because that means they get to sell you motors more often.

Submersible motors can last 20 to 30 years when installed correctly, and running on smooth sinusoidal AC voltage, as with an across the line starter. Of course motor manufacturers don't like this so, they are always going to suggest a Drive.

This is a case where you don't need Drives. Four of the five pumps run at full speed all the time anyway. Which means the drives are costing you extra energy, and reducing the life of the motors. The one pump that does use a PID loop, could easily be controlled by other means like, dare I say it, a simple control valve.

I would solve the problem by removing the Drives. As someone on this forum keeps reminding us, "you can't solve the problem by using the same kind of thinking that caused the problem in the first place".

Of course in some states they give you cash incentives to install Drives, which creates the kinds of problems you are having now, and usually doesn't save any energy. Even in a different situation where a Drive might be able to save some energy, there is no way to do the math which shows the effect of short life of the motors and other equipment. A Drive would have to save a LOT of energy to be able to justify replacing the motors so frequently, not to mention frequent replacing of the Drives themselves, when otherwise all the equipment should last decades without the least bit of maintenance.

 

[Skogsgurra]

Amen.

Yes, you probably need to install motor reactors.



Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[gepman]

Valvecrazy is correct that if the pumps are running at full speed that you shouldn't be using VFDs, it is a waste of energy and money. I will reserve judgement on the last pump which varies its speed to maintain pressure until you give the system hydraulics. There is a very long thread regarding this if you care to read it,

http://www.eng-tips.com/viewthread.cfm?qid=205555&page=3

I was so interested in answering his question that I didn't pay much attention to his initial statements. That being said it is most likely an incorrect application of the VFD that caused the problem. I recommend that you read the following application guide for VFDs.

http://www.nema.org/stds/acadjustable.cfm#download

This guide will NOT apply exactly if you have a submersible pump motor since this is not a standard NEMA motor configuration however the general principles apply. If you have a submersible pump motor then follow the recommendations of the motor manufacturer (not the drive manufacturer, which is probably written for NEMA motors).

You can get some information on VFDs and submersible pump applications at

http://www.goulds.com/product.asp?ID=174&MASTERID=2

and

http://www.goulds.com/product.asp?ID=281&MASTERID=2.

Constant pressure control may not be the best control algorithm to accomplish your objectives. If you look at the Gould's Aquavar VFD control they have an algorithm which varies the pressure with the flow so that there is always sufficient pressure to meet demand without providing excess pressure which wastes energy.

 

[Valvecrazy]

I know that there is a voltage loss when using a line reactor or filter, which increases the amperage and heat. The last line filter I observed reduced the voltage coming into the Drive from 480 volts to 462 volts, and increased the amperage proportionally. I also observed considerable heat coming from the line filter itself. How would you figure heat or power losses from the line or load reactors or filters? Also when you have a voltage drop because of a line or load filter, how does the increase in amperage and heat affect overall power consumption, compared to a Drive without any filters?

 

[gepman]

Valvecrazy A "line filter" can mean almost anything. If it was a line reactor that you observed on the output of a drive it was probably due to your measurement device which will not work well on the output of a VFD. Or there could have been poor connections.

You can see what the loss is for typical line reactors at

http://www.mtecorp.com/instr-011_line-loadreactor_060227.pdf,

see pages 25 and 26. The losses are relatively low.

The drop in voltage and increase in amperage does not affect the power consumption by the motor unless the voltage goes outside the typical rated limits for motors. The increase in power can be measured by the loss in the filter/line reactor which can be measured by I2Z or I*DELTAV*SQRT3.

 

[Valvecrazy]

I was reading voltage and amperage on the input to the drive, before and after connecting the reactor. So the supply voltage to the Drive was reduced by the reactor, which increased the amperage. This increase in amperage must increase the overall heat losses but, I don't know how to add that back in to the equation.

Thanks for the link Gepman. If I am reading it right, it shows about 40 watts loss for each reactor on about a 20 HP load, and the losses increase as the FLA increases. Just something else that should be added back into energy use calculations. There is also a little less than 1% loss on an active harmonic filter as discussed in a previous thread. Every time you add a filter, you are losing something. These numbers may seem insignificant but, when there is only a few percent difference in power consumption between a drive and no drive, they become very significant.

However, a percent or two extra power consumption by a filter or reactor, is much less significant than the motor and or Drive, not lasting as long as an Across The Line Starter, and a motor running on smooth sinusoidal AC power.

 

[eerickson]

powersoff, we have similar problems at our facility. Check your VFD’s operational manual for the recommended carrier frequency, and verify this setting on your drive. I found several good references on IEEE Xplore, but really haven’t determined the best design approach for today’s IGBT VFD(s). The two best reads I've found are:

1. Reflected wave modeling techniques for PWM AC motor drives, By Skibinski, G.; Kerkman, R.; Leggate, D.; Pankau, J.; Schlegel, D. This paper appears in: Applied Power Electronics Conference and Exposition, 1998. APEC '98. Conference Proceedings 1998., Thirteenth Annual,Page(s): 1021 - 1029 vol.2, 15-19 Feb 1998
Volume: 2 Issue

2. Cable characteristics and their influence on motor over-voltages
By Skibinski, G.; Leggate, D.; Kerkman, R.
This paper appears in: Applied Power Electronics Conference and Exposition, 1997. APEC '97 Conference Proceedings 1997., Twelfth Annual
Page(s): 114 - 121 vol.1, 23-27 Feb 1997
Volume: 1 Issue

Currently, I’m waiting on more information from Southwire (

http://www.southwire.com).

They design a triplex conductor interlaced with grounding conductor system. From what I gather the peak voltage destroying the motor systems is due to twice the dc bus voltage times the impedance mismatch ratio of the motor and cable. There is nothing you can do about the motor impedance (short of inverter duty design motors), but the cable impedance can be altered (i.e. line reactors). Filtering does costs and will reduce the system performance. My preference is not to filter.

Good luck.

 

[Skogsgurra]

I found a comparison between motor voltage and the Rema curve showing typical peak voltages. This is an 800 kW dynamometer machine for truck motor testing. Grid is 690 V and standard cables have been used. No screen, no symmetric PE.

The peak voltage is very close to the theoretical limit and the motor had problems with insulation faults after only a few month's operation.

I think that reactors have been installed now. But have not done any measurements after that.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[thewellguy]

What brands of equipment are involved in this installation. Some brands of vdf already have output filters build in. Also you need to look very closely at the nameplate info for the subs and conpare it to the vfd's. sub motors as a rule of thumb pull more amperage then an above ground motor. For example the Ugly book give the fla of a 50 hp 460 vt motor at 65 while the Franklin motors o& m manual states that a 50 hp 460 volt 6" motor can pull as high as 77 amps.

Also per the manual Filters or Reactors: Required if (1) voltage is 380 or greater and (2) Drives uses igbt or bjt switches (rise-times , 2 msec) and (3) cable from the drive to the motor is more than 50 ft. A low pass filter is preferable. Filters or reactors should be selected in conjunction with the drive manf. and must be specifically designed for vfd operation.

 

[ozmosis]

I'm still not clear on what the actual "fault" is.
The last post by powersoff mentioned "VFD would fault off.I had continuity between pump motor leads and ground(earth)"
The description of the VFD "fault off" tells me that there is a problem that the VFD is picking up. Nowhere have I seen that the motor has failed. All points mentioned so far are very valid but until we know more details from powersoff on what is the actual problem, a lot of what we say could be meaningless. It is quite possible the user has used screened cable and over such a long cable distance the screen creates a high capacitance and the VFD sees this as an over-current fault. Or, the stray capacitance due to a motor that is 400feet from it's ground point could cause the VFD to see this as a ground fault.
It is quite common that "faults" indicated by the VFD are seen by certain users as "failures" of products whereas a fault indicated by a VFD is an indication that whatever it is measuring (current, voltage, temperature...etc) has simply been measured outside the tolerance of the VFD. It doesn't always mean something had "failed".

So, powersoff, perhaps you'd like to expand further on the fault then these fine brains in VFD's and valves can provide more accurate tips before we end up saying get a bucket on a piece of rope and get your water up that way.

 

[powersoff]

The pump motor leads had continuity to ground causing the VFD to fault.Apparently the VFD does not like a .019ohms measurement to ground from motor lead. The wires between well head and vfd are meggered and test o.k.
The wire is direct burial cable in pvc conduit to well head.
The VFD brand is Altivar.

Thanks

 

[gepman]

If you checked continuity with the motor connected you could have a short (or fault) in either the motor or the motor leads. Hopefully it is not the motor. As I and others have mentioned the reflected wave phenomena plus the dv/dt of PWM drives can cause burn through of the insulation either in the motor or the motor leads. Allen-Bradley does not recommend THHN/THWN conductors for their PWM drives in wet conditions which is anything underground or obviously in a well.

The brand name Altivar is I believe a Square D brand and although not my favorite, I have used hundreds without any problems when my clients want them. I, like others, prefer to use what we are most familiar with.

To summarize the above replies, your best solution is to determine if the fault is in the motor or the leads, remove the VFD's where you don't need them, follow the manufacturer's recommendations for the motor if it is a submerisble pump motor (you never have really said), and install some type of attenuation device for any VFD's that remain including load reactors. The motor leads should be rated probably at least 1000V, I usually use something like a Belden or Olflex VFD cable or at least XHHW-2. Shielded cable as ozmosis says can cause high capacitance problems.

Let us know what your resolution is.

 

[thewellguy]

A question for Gepman. I'm going to assume (I know) that powersoff is referring to a submersible well pump. You suggested that 1000v cable be used in this application. I'm would guess that this is so the cable does not fail due to spikes. I would think that it would be cheaper to have the cable fail due to spike than the motor. Therefore, my guestion is why up the cable rating?

 

[gepman]

thewellguy
Yours is an interesting concept but if you really want to design the system so that one component fails before the other (similar to a shear pin) then you need exact information on the dielectric capabilities, capacitance, etc. of both the motor and the conductors. This will probably be hard to get. Also powersoff may replace the submersible pump motor (now or later) with a better insulated unit later and then he wouldn't have to replace the cable or conductors.

If powersoff installs a load reactor it is probably a moot point since the cable insulation would not be as much of a factor.

My philosophy is to design the system so that no part fails. I have used the "shear pin" method a few times but it takes good engineering information to be certain that your shear pin will fail with an overload but not with a normal load. I don't believe that information is available in this case.

 

[powersoff]

Just to be clear the submersible pump is smoked/fried/junk/needs to go to a motor shop to be fixed.The VFD is good and is currently controlling a brand new submersible pump.
I am grateful to all who replied and am researching options.