VFD issues

[KAG]

I have a vfd that is giving me some ramping problems in PID control. It was running smoothly and as it was reaching it's setpoint went into high current overload. After some wire switching and moving to local control it ran smoothly at a 40hz and above. At 60hz current readings using a fluke ammeter show dead on manufacturer specs but the VFD shows at least 15% higher readings hence the overcurrent fault. ABB believes it is a faulty VFD. Just looking for some input from those of you out there who have experience with these things.
Also Just wanted to say this is a great forum and I am a new member!

Regards,

 

[Marvo]

Welcome to the forum KAG.

The problem with measuring the run current of a VFD is the carrier frequency. Many ammeters can't give an accurate reading at those kinds of frequency ranges. Can you state exactly which VFD and exactly which ammeter you have please.

As a rule I'd always take the run current calculated and displayed by the VFD as being the real life run current at least up to the point where it's proven without doubt to be inaccurate. The only times where I've encountered a VFD that was giving inaccurate readings it was down to the inaccurate motor information that was inputted into the VFD. I've never come across one that was inaccurate because it was actually faulty.

 

[KAG]

With the VFD running at 60hz, ABB ACS310, the display reads 8.4amps, the output from the vfd reads 7.4, and the current from the disconnect ahead of the VFD is 7.5 approximately. Measured with a Fluke True RMS clamp on. Can't remember the exact model. This thing was running fine at 60hz in PID mode then began to show current excess on the display and went to high current warning. Runs fine in local control. Strange?

 

[KAG]

By the way thanks for the help. I would expect the current readings to be stable at 60hz? Also the voltage was fluttering up and down 10-30volts on the motor side of the VFD as well.

 

[KAG]

AT 60 hz I would expect the readings to basically be motor name plate. At one point during the running of this motor it ramped up to max drive current and stalled to 22.5hz. after a bit of looking and wire switching it took off and ran fine in hand mode.

 

[Marvo]

You're going to have the same problem measuring the output voltage of the VFD, again it's very possible the carrier frequency is beyond the range measuring capability of your Fluke.

Have a quick read of this if you're not sure how the modulated output of a motor drive is made up of carrier frequency (can be way over 10kHz for some drives) as well as the output frequency (0-60hz for USA systems).

Just out of interest what is the motor doing?

 

[KAG]

The motor is powering a submersible water pump. Even at SFA the current draw should be well under the vfd 10.8a limit. Itis under its greatest load when the water level in the dewatering well is high and motor current should drop as the water level falls. I spoke with the ABB tech support rep and he feels there is too much difference in the clamp meter reading and the vfd amp readings. It is possible the motor could be at fault (it is brand new) but it runs fine in hand operation!

 

[KAG]

Excellent read. Makes sense.

 

[Marvo]

Do the pump manufacturers state specifically that the motor is suitable for VFD running? Some motors don't have good enough insulation on the windings and the VFD output can cause tracking internally, symptoms of which could present as high and unstable run current.

 

[KAG]

Meter was a Fluke 30/32 True RMS - What would you recommend for an accurate (as accurate as can be expected) short of a scope.

 

[KAG]

Yes they do. They are wound specifically for that purpose with the caviat that they never be run below 30hz. Below 30hz the bearings (water lubricated) do not receive sufficient flow to properly lubricate. It is certainly possible that the pump intake picked up a small amount of debris however at 3450rpm it would simply chew it up and spit it out without hesitation. A millisecond spike would be all I would expect.

 

[Marvo]

If you're a Fluke fanboy you could go for a Fluke289. Here's a US link where Fluke explain the problems of taking readings from VFD outputs with standard instruments. I'd normally suggest you try our forum sponsors who offer members a decent discount but they're UK based so unfortunately you're out of luck.

 

[KAG]

Much appeciated

 

[Marvo]

Obviously without seeing the installation in the first person it's difficult to go through the normal proceedure of fault finding. Rereading the thread in its entirety a few thoughts spring to mind.

For now at least I'd ignore what your Fluke is telling you. There's a good chance it can't give you accurate readings so rather take the run currents from the VFD itself as gospel. By all means beg or borrow another instrument and confirm them but I don't see why the drive would be incorrect if you've inputted correct motor info during setup. I think the current reading discrepancy is introducing unnecessary confusion and leading you astray from finding the root cause of the tripping.

Pumps don't always follow that lower RPM = lower shaft power requirement = lower running current. There's many things that can produce an unexpected run current, 1. lock rotor as you mentioned a blockage may have occurred 2. The pressure across the pump, if it's too low you can get unexpected cavitation or impeller lift which can increase the run current. Most pumps have a minimum head pressure requirement to prevent this. It's worth noting incorrectly sized pipework or an elbow too close to the inlet or discharge can have a similar effect. Also excess static pressure ie a blockage on the discharge side can either increase or possible decrease the run current outside of expected limits depending on the type of impeller.
These are all pump problems that could give the kind of symptoms you're seeing.

I think the PID mode might be a large part of the problem. With an unpredictable load it can be hit and miss at best. PID can also be very complicated to set up correctly on a VFD, if we ever have an application that requires PID control we tend to stick to known combinations of VFD and machine where someone has done the painstaking work of setting up the hundreds of parameters for that particular combination. At least to start with I'd go through the entire commissioning procedure without PID mode. Once you're sure everything else is behaving to expectations then last of all recommission in PID mode and at least you'll have baseline figures to work from when testing.

 

[KAG]

Agreed. I have another meter, Extech EX830, which I'll use to confirm readings. When I left the installation is was running at a smooth 40hz with no issues in local mode. I tend to agree with the PID part of your response. The pump manufacturers like a gain of 4 on these things with an integral setting for @ 20 seconds. Other issue is submersible pumps typically like quick starts so the builders say 1 sec max start time. The problem here is varying volumes so the PID will be necessary unless there is another way to control the speed as the water rises. These 4" submersibles typically draw max current during the lowest head conditions and this particular installation is just outside the curve for the model. Even in the service factor however the amp draw is within the confines of the vfd's maximum. I may allow it to run a few days at a reduced manual speed and see how it all shakes out. If all goes well then I will try the PID macro again and see what happens. ABB says the drive is faulty so we'll replace it and see if that makes any difference. A few seconds of dry run periodically is better than losing a motor. I may also induce some artificial head on the discharge in the form of a valve! Last resort is to go to a smaller pump and overspeed. Franklin electric allows up to 80hz on their submersible motors, with reduced overall life of course. I will however respond when I get to the bottom of it but I believe you are correct about the PID.

 

[Marvo]

I'd be interested to know what transpires if you can find the time to post again in a few days.

If ABB are prepared to swap out the drive at no cost then it would be wise to do so just to rule out that as a possibility.

If your pump is a multistage centrif then low head is not a place where you want to be, even if the head is slightly below specified minimum it can have a marked detrimental effect on mean time between failures so artificially introducing extra head pressure is a close second best to re-specifying a pump with a lower head requirement.

<1 second to >3450rpm is a steep start ramp, it's steeper than the default ramp settings of most drives so I'd double check these settings which may be critical to ensure bearing hydraulic action and possibly seating of the impellers. Also with bearings that rely on water hydraulics for lubrication I'd avoid dry run at all costs, it could lead to the demise of the pump faster than a marginally high run current.

Maybe consider a 0-10 volt level sensor (or something similar that's drive compatible) that can signal the drive to modulate the motor speed between 30-60HZ, it might be simpler and more reliable than PID operation.

Good luck and I'll look forward to reading what happens.

 

[KAG]

FYI I do have a 4-20ma transducer in the well. 0-15psig. Set at 31ft with a setpoint of 10ft (21 from surface). These are submersible open design impellers. Motor is always submerged and the stack can take a ton of run dry without issue. We once ran a pump and motor out of the water for 75 minutes without damage. I think we got lucky! Anyway it may be a programming issue and although I was given parameters directly by the ABB tech people that doesn't mean it's correct. I'll keep you posted once we iron it all out. This has been one of THOSE projects anyway! Again your help is greatly appreciated.

 

[Engineer54]

Now this has been an interesting thread, and i'm also looking forward to what happens on the OP's end....

 

[i=p/u]

If you don't understand I can talk you through what they on About.....

 

[KAG]

I'm certainly open to any suggestions for alternatives to pid control or perhaps programming suggestions. Feel free to email me and I'll give you the blow by blow. I'm thinking I may have to limit the motor to a level where it stays in the curve. Perhaps 50-55hz.

 

[Marvo]

Can you give more info about the end game you're trying to achieve? You have a pump in a well with a level transducer, is that transducer actually there for level control or is it just dry run safety? Where is the pumped water going and what processes on the discharge side of the pump are you trying to control?

 

[KAG]

Ok. Here we go. We have a basement in a building that floods during rain events. There is a layer of porous sand at a foot below the basement floor that is perched on top of the solid rock. When it rains it fills rapidly and sump pumps can barely keep up to the demand. We drilled a hole just outside the foundation wall to drop the water to a level where we could intercept a large portion of it before it enters the building. We installed a submersible pump inside of a sand packed pvc liner to a depth well below that sand level to force the water out and down. Once we begin pumping we discharge the water to a storm drain prior to entering the facility. Long term the constant pumping will drain the zone and make the rain events less of a catastrophe so ideally this pump will run continuously unless it somehow is completely dewatered. All I really need to do is keep the water level just below that sand layer at all times. here are some details.
Hole depth - 46ft
Diameter 4.25" ID inside sand packed 6" hole
Pump intake at 33ft
Transducer - 4-20 hanging at 31ft
Current draw 7.5 SFA 8.5
2hp 200v 3 phase
ABB ACS310 drive

Basically the discharge water just runs to city storm sewer. The motor went into overload once either as a result of lack of head or some other mechanical factor. Screen slots on the pump are less than 1/8 inch and sand grade outside the liner is 1/8 x 1/4 with .125 mchine slots. No doubt some rogue particles can get through but even if they pass the well screen and enter the pump somehow they should be pumped out without difficulty. Certainly not enough to send a pump running at full speed into overload. I am suspecting lack of head on the submersible may be the culprit.

 

[Tomma23]

Hi Kag & Marvo . Is it so that using PID control for what appears to be just a matter of controlling the Water Head level a bit over complicating the function . Could you not use say a two stage Bilge Level switch giving a High and a High High input to the VFD for 2 fixed speed outputs with suitable ramp up times . Use the lower level of the High switch to stop the Pump and protect from dry running . Just a thought . Regards

 

[Engineer54]

If you don't understand I can talk you through what they on About.....

Now you ARE joking!! no-one can understand your one line posts, let alone you trying to explain issues with a VDF controlled pump motor!! lol!!

 

[KAG]

Looked like plain english to me!

 

[i=p/u]

So that's what is you on about. .... Pittsburgh rocks.

 

[Darkwood]

I had a similar issue although its clear its not your issue after reading through; .... a submersible pump had been repaired by others but next heavy rain fall it tripped O/L ... through investigation it had been connected up with phases swapped and was running backwards now the obvious thought would be it wouldn't pump but it did pump successfully at about 50% flow rate reduction at full speed but with much higher current demand as the natural water flow was fighting rotation.

I doubt this is anything to do with your set up as you say you can run it fine without PID but just putting it out there as a future ref' it tends to go against the grain and is counter-intuitive.

Another thought though is the Inverter in damp conditions or got any dust contaminants?
Is it set up for reverse lockout so a erroneous input, tracking signal wont give a reverse direction command...?

 

[Engineer54]

Looked like plain english to me!

You need to look at his other posts to see what i mean, or even his last post here #25!! lol!!

 

[Silly Sausage]

Hahaha...I hope he (i=p/u) doesn't talk like this to his clients/bosses!!! :smilielol5::smilielol5:

 

[Knobhead]

I’m trying to get my head around why you need a VSD on a sump pump? Basic level control of a DOL starter makes more sense to me.
Can you post a link to the pump manufacturer and the type?

 

[Silly Sausage]

Phew, it's not just me then Tony!
A pump with some level sensors for control and alarm conditions...surely?!?!

 

[Knobhead]

Makes sense to me.

Of the 100’s of pumps I’ve worked on only two had VSD’s, they nearly always ran at about 90/95%. Look at the flow and speed graphs they aren’t linear. Under speed they are inefficient, over speed and they suffer with turbulence.

Crazy setup!

 

[Marvo]

All I really need to do is keep the water level just below that sand layer at all times.

Tony, I assume the PID control is either because he's trying to maintain a level rather than have the pump cycle on/off as it would with a traditional 3-probe liquid level controller. Possible it's also because there a massive difference in pump performance required depending whether it's raining or not. If the pump is sized to deal with the maximum likely rainfall then it will almost certainly 'short-cycle' when it's not raining and probably exceed the maximum number of starts-per-hour specified by the manufacturer.

You can often get away with having an increased number of starts per hour just by using a drive with a ramped start rather than a DOL starter but with a multi-stage centrif pump it needs to start on a steep ramp so you won't get away with more than maybe 10-20% over the manufacturers limit.

I'm also a big fan of simple and PID control can often be anything but simple. It might be possible to use the existing level transducer and have the drive handle it as a proportional only analogue input but without seeing the volume requirement figures it's difficult to say.

 

[KAG]

Marvo you are exactly correct. This is a dewatering application where the level must never rise above a preset depth. In this case the contractor poorly constructed the elevator pit in a commercial building so when a rain event happens it floods. By keeping the water below the elevator pit the problem goes away! We have done many of these but in most case the water production rates remain constant. In this case the volume of production at the soil bedrock interface varies from 10 - 70gpm depending on weather and time of year. To all reading this it is not a sump pump but a high volume submersible turbine. Very similar to you typical water well except larger diameter and greater volume. The idea is to create a cone of depression and essentially dewater the zone so the rain events have as little impact as possible.
Now for the solution. It appears that the folks at ABB called out some settings that were not so compatible with the type of motors used in submersible pumps. After a call to the motor and pump manufacturers we were able to set the gain, integral, ramp up, and carrier frequency to make the pump and motor happy. The current draw is now under service factor and the unit is operating as designed. It appears you were correct (Marvo) that the readings shown by the drive were in fact correct and once the parameters we correctly set we're off and running.
Again I appreciate the help!

 

[Marvo]

Thanks for the explanation, I am familiar with the dewatering technique you've employed and I'm glad to hear you got your settings issue cleared up. Thanks also for coming back and tying the thread up neatly.

 

[Darkwood]

Cheers for the feedback...
Now to implement a second pump and they can run alternately and if and when a trip or pump failure occurs the controls put the remaining pump in charge and alarm up for maintenance well this was my last set-up but it was for chemical waste and keeping it from flooding into the local river so may be overkill in your set up...?

 

[KAG]

Certainly in all cases when you can convince the facility owner to implement a spare anything it's good practice. Economics being what they are today they opted for a high level alarm!

 

[Darkwood]

Certainly in all cases when you can convince the facility owner to implement a spare anything it's good practice. Economics being what they are today they opted for a high level alarm!

and a mop and bucket

 

[Knobhead]

Best laugh I’ve had on here for a long time!

Thanks DW!