VFD Failure on Generator Power

[drktmplr12]

This is a shot in the dark, but quick enough to ask.

I know this is a loaded topic!

I am tasked with preparing a technical document explaining why water production wells were failing and restarting while running on generator power during and after Irma. This occurred in Southeast Florida.

I am not asking the forum to provide me a dissertation, that is my job. I am more looking for simple (and probable) theories to consider from the informed opinions on the forum. I will provide any additional information that is requested, granted I can easily get my hands on it.

Thanks in advance for anyone who takes the plunge here.

My current theories:

• Process conditions caused wells to experience overload conditions, causing VFD protections to trigger. The motor cooled down and automatic restart triggered. The cycle repeated until operations intervened.
• Generator became unstable due to harmonics resulting in overvoltage conditions. The VFD would fault, generator would stabilize and automatic restart triggered. The cycle repeated until operations intervened.

I know the following:

• Failures were observed while wells were operated with portable generator power.
• Wells may or may not be on same utility feed.
• Same behavior at all wells with VFD (Altivar 61)
• Well controls have been equipped with VFD’s for approximately 5 years.
• Wells controlled by VFD have been operated on generator power previously and without incident.
• It has proven difficult to repeat the behavior with the same generators.
• Data gathered includes generator nameplate data, electrical data (voltage, power, harmonics), VFD configuration parameters and flow measurements.
• The configuration parameters one VFD was downloaded to be inspected for possible causes for VFD failures.
• A fault log for one drive was obtained, which revealed a series of overvoltage faults without time stamp. These could have happened before, during or after the storm.
• Automatic restart is enabled.
• A Dranetz Power Platform 4300 Spectrum Analyzer was used to measure voltage and harmonics present at each site.

Well 24, 75 HP
FPL AB 486V, AC 486, BC 486
Gen. AB 514, AC 514, BC 514
Wacker G125 Prime Rated, 98/122kVA @ 0.8 pf

Well 29, 175 HP
FPL AB 483V, AC 483, BC 483
Gen. AB 480, AC 480, BC 480
CAT 350/437kVA @ 0.8 pf

 

[Besoeker]

This is a shot in the dark, but quick enough to ask.

I know this is a loaded topic!

I am tasked with preparing a technical document explaining why water production wells were failing and restarting while running on generator power during and after Irma. This occurred in Southeast Florida.

I am not asking the forum to provide me a dissertation, that is my job. I am more looking for simple (and probable) theories to consider from the informed opinions on the forum. I will provide any additional information that is requested, granted I can easily get my hands on it.

Thanks in advance for anyone who takes the plunge here.

My current theories:

• Process conditions caused wells to experience overload conditions, causing VFD protections to trigger. The motor cooled down and automatic restart triggered. The cycle repeated until operations intervened.
• Generator became unstable due to harmonics resulting in overvoltage conditions. The VFD would fault, generator would stabilize and automatic restart triggered. The cycle repeated until operations intervened.

I know the following:

• Failures were observed while wells were operated with portable generator power.
• Wells may or may not be on same utility feed.
• Same behavior at all wells with VFD (Altivar 61)
• Well controls have been equipped with VFD’s for approximately 5 years.
• Wells controlled by VFD have been operated on generator power previously and without incident.
• It has proven difficult to repeat the behavior with the same generators.
• Data gathered includes generator nameplate data, electrical data (voltage, power, harmonics), VFD configuration parameters and flow measurements.
• The configuration parameters one VFD was downloaded to be inspected for possible causes for VFD failures.
• A fault log for one drive was obtained, which revealed a series of overvoltage faults without time stamp. These could have happened before, during or after the storm.
• Automatic restart is enabled.
• A Dranetz Power Platform 4300 Spectrum Analyzer was used to measure voltage and harmonics present at each site.

Well 24, 75 HP
FPL AB 486V, AC 486, BC 486
Gen. AB 514, AC 514, BC 514
Wacker G125 Prime Rated, 98/122kVA @ 0.8 pf

Well 29, 175 HP
FPL AB 483V, AC 483, BC 483
Gen. AB 480, AC 480, BC 480
CAT 350/437kVA @ 0.8 pf

The large difference between utility and generator distortion figures suggest that the generator impedance is attenuating them. The inference from that is the generator voltage is being significantly distorted and that may be causing the drive supply monitoring to see it as a fault.

Have you seen what fault code is displayed when the drives trip?

 

[drktmplr12]

Have you seen what fault code is displayed when the drives trip?

We were having difficulties repeat the fault after the storm. The only fault codes available were Mains overvoltage, but with no time stamps.

Also, the drives have no input filter or reactor, we are recommending they change that!

 

[Besoeker]

We were having difficulties repeat the fault after the storm. The only fault codes available were Mains overvoltage, but with no time stamps.

Also, the drives have no input filter or reactor, we are recommending they change that!

Not a bad idea. Can you put a scope on the input voltage when running on generator power?

 

[drktmplr12]

Not a bad idea. Can you put a scope on the input voltage when running on generator power?

That can be arranged, although not preferable since we cannot currently replicate the issue. The Analyzer might have stored the voltage traces. I am going to see what I can locate with the scope's software.

 

[ptonsparky]

I've had rapidly chattering motor starters cause an overvoltage fault on a drive that were not related directly to the Drive operation. Other than that it is usually due to deceleration ramp time being to short. I would imagine the generator response to loading will be a major factor but I would think that would have been apparent since install.

 

[ptonsparky]

We were having difficulties repeat the fault after the storm. The only fault codes available were Mains overvoltage, but with no time stamps.

Also, the drives have no input filter or reactor, we are recommending they change that!

A must, though that did not help with my chattering starters.

 

[Jraef]

Was there flooding and/or wind damage? A lost neutral connection or a high resistance ground fault, too low to be picked up by your GFE system. That could have caused an imbalance and /or the generator AVR was increasing the voltage on the other legs or in general as a result. Now whatever caused that has been fixed or gone away (flooding in conduit subsided) and you can no longer see the issue.

 

[drktmplr12]

Was there flooding and/or wind damage? A lost neutral connection or a high resistance ground fault, too low to be picked up by your GFE system. That could have caused an imbalance and /or the generator AVR was increasing the voltage on the other legs or in general as a result. Now whatever caused that has been fixed or gone away (flooding in conduit subsided) and you can no longer see the issue.

Above grade flooding was not observed as far as I know. This doesn't mean the water table was not affected, changing the suction head and causing motor to operate on different location of the curve.

Wind damage was limited to down tree limbs. The wells are fairly protected and did not experience physical damage.

Are you suggesting a ground fault on one phase of the drive output, saturating that phase and causing the generator controls to overshoot the other phases / phase imbalance fault @ the drive?

 

[drktmplr12]

I've had rapidly chattering motor starters cause an overvoltage fault on a drive that were not related directly to the Drive operation. Other than that it is usually due to deceleration ramp time being to short. I would imagine the generator response to loading will be a major factor but I would think that would have been apparent since install.

The wells are generally on dedicated utility transformers, no other motor loads to speak of.

 

[Besoeker]

That can be arranged, although not preferable since we cannot currently replicate the issue. The Analyzer might have stored the voltage traces. I am going to see what I can locate with the scope's software.

Even if you cannot replicate the fault, voltage waveforms might give a clue about how to prevent a recurrence. Jut a thought.

 

[drktmplr12]

Even if you cannot replicate the fault, voltage waveforms might give a clue about how to prevent a recurrence. Jut a thought.

I'm looking at the data on Dran View, I've not used this software before so I'm stumbling through it.. I looks like the guys that pulled the field data only took snapshots of harmonics. I can't find any waveforms that show transient events.

 

[petersonra]

It might be possible for some weird set of harmonics to cause a vfd to trip on over-voltage. Doesn't seem real likely but I suppose it's possible. I don't think that line reactors are going to make much difference when you're on generator power.

Realistically if you can't repeat a phenomenon it is going to be hard to debug it. I can't think of anything off-hand that might cause a generator to trip a vfd during a storm but not after the storm. I suppose there is some possibility that lightning was involved but I hate to even suggest that because it is blame for so many things that it almost never is responsible for

 

[Ingenieur]

Above grade flooding was not observed as far as I know. This doesn't mean the water table was not affected, changing the suction head and causing motor to operate on different location of the curve.

Wind damage was limited to down tree limbs. The wells are fairly protected and did not experience physical damage.

Are you suggesting a ground fault on one phase of the drive output, saturating that phase and causing the generator controls to overshoot the other phases / phase imbalance fault @ the drive?

without error codes hard to determine

your observation about water table increasing, therefore decreasing discharge head and overloading the pump sounds plausible

I would look at how they are configured and see which shutdown faults autoreset/auto start and which ones lockout and need reset

sre you saying that the pump would fault on util and not restart until util was lost and it transfered to the gen? That might reset latching lockout codes

edit: looks like it only failed then restarted on gen?
no issue on util power?

 

[drktmplr12]

I would look at how they are configured and see which shutdown faults autoreset/auto start and which ones lockout and need reset

This is my exact approach. Then narrow it down from there. I am thinking this is a case of the VFD self protecting and automatic restart beating up the motor a bit.

sre you saying that the pump would fault on util and not restart until util was lost and it transfered to the gen? That might reset latching lockout codes

They run fine on utility. during the storm, utility was knocked out so they had portable generators ready to go. They fired them on and they ran for a period of time. Then 7 wells began with the erratic starting and stopping. This is why we were thinking they were process related. Funny part is, there are two wells that feed the lime plant, and the other 5 wells feed the RO plant (there are 30 wells in total). Curiously enough, both sets were acting up so that makes that argument more difficult to support.

no issue on util power?

No issue on utility.

 

[drktmplr12]

It might be possible for some weird set of harmonics to cause a vfd to trip on over-voltage. Doesn't seem real likely but I suppose it's possible. I don't think that line reactors are going to make much difference when you're on generator power.

The more I think about this, the more i think it's not probable. I know harmonics do weird things, but I thought it would more likely be undervoltage. I don't have a large amount of experience troubleshooting harmonics so I'll just put that out there. We also are not of the opinion that harmonics caused the faults, it was more of an "oh by the way you need input reactors or filters."

Realistically if you can't repeat a phenomenon it is going to be hard to debug it. I can't think of anything off-hand that might cause a generator to trip a vfd during a storm but not after the storm. I suppose there is some possibility that lightning was involved but I hate to even suggest that because it is blame for so many things that it almost never is responsible for

I was in my house during the storm and I didn't hear any thunder at all. Hurricanes don't generally produce a lot of lighting, although i'm sure there are exceptions.

 

[Ingenieur]

Once running do the pumps
run continuously or cycle (ie floats, levels, etc)?
is the speed modulated by a pid or constant speed when running?

how close are the pumps operating amps vs rated fla?
does the vfd limit starting current?

the small gen looks tight
the larger is more than adequate

 

[ghostbuster]

Our group has been taking harmonic measurements for over 30 years and are puzzled at these recorded measurements.


At a quick glance ,it would appear there maybe something wrong with the harmonic measurement process and computations. For example the harmonics (assuming these are current harmonics being recorded and displayed in your chart) are significantly lower on generator power (which is also suprising) but the overall voltage THD is equal to or larger compared to when being supplied with utility power.Something does not add up !:?

 

[drktmplr12]

Once running do the pumps
run continuously or cycle (ie floats, levels, etc)?
is the speed modulated by a pid or constant speed when running?

how close are the pumps operating amps vs rated fla?
does the vfd limit starting current?

the small gen looks tight
the larger is more than adequate

the modulate on flow/pressure depending on which process they are feeding (RO or Lime plant).

Nothing seemed out of the ordinary with respect to running amps. They do not have historical data for VFD output amps @ SCADA.

Acceleration ramp time is set to 1 s, down from default setting of 3 s. Rated Frequency is 60 Hz, so 0 - 60 Hz in 1 second. It is limiting starting current to a degree, but 1 s seems relatively steep.

Our group has been taking harmonic measurements for over 30 years and are puzzled at these recorded measurements.


At a quick glance ,it would appear there maybe something wrong with the harmonic measurement process and computations. For example the harmonics (assuming these are current harmonics being recorded and displayed in your chart) are significantly lower on generator power (which is also suprising) but the overall voltage THD is equal to or larger compared to when being supplied with utility power.Something does not add up !:?

I did not personally take the measurements, but I have a high degree of confidence in the guys who did. That being said, I understand no one is perfect.
The 3rd 5th and 7th listed are current harmonics.

I presented the results to Mirus International (specializes in filters, reactors) and they weren't terribly surprised. They did not have comments on the erratic behavior. They just recommended a tuned passive filter at the PCC, which we all agree should be there.

I recall studying harmonics in school, but in a academic context.. and I understand how they distort the waveform, but the immediate practical implications are somewhat lost on me currently. I am working on improving that, read all types of white papers, etc. That being said, I don't know what values to generally expect when looking at harmonics. I am going to take another look at our readings and see if I can understand a bit better.

 

[Sahib]

Generator AVR's checked okay? Check with replacing the AVR's.