All my VFD's are tripping on Over-voltage

Status
Not open for further replies.

csc_wyo

Member
Location
Denver
Occupation
EI&C Engineer - Oil & Gas/Heavy Industry
I'm on a facility where every single 480V VFD is simultaneously faulting on over-voltage. :thumbsdown: I have roughly 20 pumps and fans on drives, varying HP size. This fault isn't consistent, but this is the second time it's happened. The first time we wrote it off to a potential lightning hit, but there hasn't been any storms in the area lately and it happened again last night.

VFD Info:
The drives are all AB PowerFlex 753 AC Drive, Open Type, 480V, 3 PH, Frame 2, Filtered, CM Jumper Removed, DB Transistor. I've verified VFD frame sizes for their respective load, cable sizes, VD, grounding, etc. and I cannot see anything out of the norm.

Site distribution:
The site power distribution consists of two, 20MVA, 13.2/4160 Virginia transformers from an on-site substation. These dual feeds go into one, main-tie-main, 4160V, 3P, 3W switchgear which then feeds the plant load from the two main busses. There are three main 480V MCC buildings, each with their own 1MVA, 4160/480 step down transformer for plant load, including the feeds for the VFD's that are tripping. One low voltage MCC building contains a 4160V, RVSS for a 6100HP compressor that has not been commissioned yet. Additionally, there is one 4160V MCC Building that houses: two RVSS's for two 800HP compressors; One Toshiba sync VFD for three, 3750HP compressors; and an additional Toshiba sync VFD for three, 1250HP compressors.

Additional Comments:
  1. I'm having my field techs pull the time logs, but I'm wondering if I'm getting the spike if/when we have a process shutdown of either our 3750 or 1250 compressors, and that sudden drop in load is causing the energy to flow into the smaller loads. We did have a process shutdown around the same time, hence pulling the fault times to see if they're somehow related. FYI, a process shutdown for this facility is rapid shutdown, sort of like an ESD but without killing everything on site. What happens to our large sync VFD's is that they will open the bypass contactor on the across the line motor and ramp down immediately (-5sec) on the motor that is still under VFD control. The third motor in each compressor set is future.
  2. No other devices are faulting that are fed from circuit breakers, including motor starters that are on E1 Overload's.
  3. The upstream substation's two main 751 protection relays are not recording any over voltage events, so I would have a hard time believing that I'm getting any spike from the Utility. The substation is running just a little high @ 4300V, but that's well within IEEE tolerance.

So am I looking at an inrush issue? Harmonics Issue? All of the above? :?

All comments and information are welcome and I thank you in advance. I'll provide any information necessary to help me out.

Regards,
RT
 
Last edited:

csc_wyo

Member
Location
Denver
Occupation
EI&C Engineer - Oil & Gas/Heavy Industry
check the overvoltage setpoints for the VFD's.

Ron,
Thanks for the quick response. All set points are at factory standard, 1.25% of nominal voltage. So in this case it's at 810VDC w/in the VFD.

**An additional note to add is that the upstream substation is not recording any over voltage events.**
 

kwired

Electron manager
Location
NE Nebraska
Overvoltage on DC bus is usually related to deceleration issues or "on the fly" starting which may be happening if you have a "blink" in the supply line.

Whatever the reason for the "blink" it may be that you need to check out settings for how a power failure/power resume is handled.

If it is trying to restart on the fly and parameters associated with that task are not set correctly, the still spinning load will generate power and dump it into the drive the drive needs to correctly determine current motor speed and adjust it's output accordingly when it does come back on line - this is what is causing the overvoltage on the bus. You get similar problem if you set deceleration time too short.
 

Jraef

Moderator, OTD
Staff member
Location
San Francisco Bay Area, CA, USA
Occupation
Electrical Engineer
What types of loads are the VFDs feeding?

What kind of harmonic mitigation are you using on both the LV VFDs and the MV VFDs?

Do you have PFC capacitors for the compressors that are on the RVSS starters? If so, how is that being accomplished?

My suspicion, without yet knowing the whole picture, leans toward PFC caps on the compressors that are not properly coordinated and thereby over correcting once the compressors shut down, and/or the synchronous compressors going into leading PF when you shut down, because the drives are not immediately turning off the field excitation to the synch motors. Because this is taking place down stream of the substation, your 751s don't see it as an OV event, but the added VARS are interacting with the VFD bus caps in the LV drives that are still on line to charge up their DC bus voltages beyond the limits.

SWAG #2 is that something about your process changes drastically when you shed the compressors and your loads go into regen, but the drives have no DB resistors to absorb it, so all they can do is trip off line.

Are you aware that you can look at what the DC Bus Voltage was when the drive tripped? It's stored in parameter #958. You can then compare that to whatever else happened at that moment in time.
 

csc_wyo

Member
Location
Denver
Occupation
EI&C Engineer - Oil & Gas/Heavy Industry
Whatever the reason for the "blink" it may be that you need to check out settings for how a power failure/power resume is handled.

kwired,
If we see a loss in power, the VFD's are not programmed to auto-start, even though they are in AUTO. They will require a manual restart which puts them back into the process control system.

RT
 

petersonra

Senior Member
Location
Northern illinois
Occupation
engineer
If you are seeing an over voltage trip (presumably this is DC bus over voltage trip), the over voltage can only really come from 2 places.

It either comes from stored energy on the load side being fed back to the drive, or from the line side.

It seems unlikely that all 20 drives would simultaneously see some kind of process disturbance downstream feeding energy back to the drive that would result in a DC bus over voltage trip, which suggests it is coming from the line side.

I would be looking at what the line voltage is when the trip occurs. Unfortunately, these things can happen pretty quick so a lot of times are missed by the DCS if it is even being recorded.
 

petersonra

Senior Member
Location
Northern illinois
Occupation
engineer
Do you have PFC capacitors for the compressors that are on the RVSS starters? If so, how is that being accomplished?

My suspicion, without yet knowing the whole picture, leans toward PFC caps on the compressors that are not properly coordinated and thereby over correcting once the compressors shut down, and/or the synchronous compressors going into leading PF when you shut down, because the drives are not immediately turning off the field excitation to the synch motors. Because this is taking place down stream of the substation, your 751s don't see it as an OV event, but the added VARS are interacting with the VFD bus caps in the LV drives that are still on line to charge up their DC bus voltages beyond the limits.

how would that work? The caps on the compressors are separated from the 480V feeding the 480V VFDs by a transformer? I would think it would be pretty hard for the 4160V caps to interact with the VFD caps through a transformer.
 

csc_wyo

Member
Location
Denver
Occupation
EI&C Engineer - Oil & Gas/Heavy Industry
What types of loads are the VFDs feeding?
This is a gas processing facility, so even though it's not a constant load/torque like you see in manufacturing or some conveyance lines, we're fairly constant, but considered a variable torque load.

What kind of harmonic mitigation are you using on both the LV VFDs and the MV VFDs?
LV's have factory in-line filters. I believe they are the FN3258-30/55/75-nn filters respective to 2, 3 and 4 frame size? MV's are T300MVi/MTX Medium Voltage VFD's, 36 pulse, 5-level output, .95pf fully complying with 519, and 3300V IGBT's.

Do you have PFC capacitors for the compressors that are on the RVSS starters? If so, how is that being accomplished?
I don't have the job book for the RVSS on my laptop, but I believe there is no power factor correction capacitors on this unit.

...Because this is taking place down stream of the substation, your 751s don't see it as an OV event, but the added VARS are interacting with the VFD bus caps in the LV drives that are still on line to charge up their DC bus voltages beyond the limits.
I agree this is what must be happening.

SWAG #2 is that something about your process changes drastically when you shed the compressors and your loads go into regen, but the drives have no DB resistors to absorb it, so all they can do is trip off line.
I'm all about SWAG's...it's how we learn. We don't have any dynamic braking applied in this scenario. As soon as we dump the compressors, our suction and discharge valves "slam" shut w/in 5 seconds. We might have an issue there and we might have to allow it to come to a more gradual stop...as gradual as the line pressure will allow us to coast to this stop.

Are you aware that you can look at what the DC Bus Voltage was when the drive tripped? It's stored in parameter #958. You can then compare that to whatever else happened at that moment in time.
My programmer is heading there this afternoon and pulling that exact data.
 

csc_wyo

Member
Location
Denver
Occupation
EI&C Engineer - Oil & Gas/Heavy Industry
It seems unlikely that all 20 drives would simultaneously see some kind of process disturbance downstream feeding energy back to the drive that would result in a DC bus over voltage trip, which suggests it is coming from the line side.
I completely agree. Some of these drives are for exhaust fans and small process pumps which weren't running at the time.

I would be looking at what the line voltage is when the trip occurs. Unfortunately, these things can happen pretty quick so a lot of times are missed by the DCS if it is even being recorded.
I'll be pulling the SEL 751 data during that period just to see if there is any sort of anomaly. Unfortunately for me, the engineering company who designed this site didn't put any sort of pqm on the 480 side of the site! :thumbsdown::thumbsdown::thumbsdown:
 

mike_kilroy

Senior Member
Location
United States
480 comes from a transformer from 4160v....

Might there be a loose connection upstream causing a huge voltage spike? Not so likely with your good fn rf filters, but?

I too don't see the drive dc bus caps adding with all that isolation between the two voltages.
 

csc_wyo

Member
Location
Denver
Occupation
EI&C Engineer - Oil & Gas/Heavy Industry
Substation information

Substation information

Finally got the trend data from the main substation incoming monitoring relays. Absolutely no incoming voltage spike.

I'm still at a loss. :happysad:
 

mike_kilroy

Senior Member
Location
United States
Still at a loss; any of these ideas worth considering?


  1. What is update time on the trend data? If not real time analog, then maybe it missed the spike(s)?
  2. Did an OV fault occur while the data was being recorded?
  3. If so, can you pinpoint that time on the report?
  4. If not, then the data is not of value yet and it still needs to be recorded?
  5. What if there is a loose connection in the downstream 460v common feed area? Maybe you need to record there or right at the input of a drive?
 

kwired

Electron manager
Location
NE Nebraska
Finally got the trend data from the main substation incoming monitoring relays. Absolutely no incoming voltage spike.

I'm still at a loss. :happysad:

Most overvoltage issues I have seen are because of deceleration problems.

Don't know enough about your application but what are chances of being in a deceleration mode when this is happening, or that a deceleration command is happening even if it is an error?

Do you have high inertia loads? If so could be the load is "overhauling" the drive output so to speak.
 

7EA

Member
Location
California
I've been able to correct similar issues in the past with line reactors. Utility company was switching in power correction capacitors at relatively the same time every day causing the issues.
 

kwired

Electron manager
Location
NE Nebraska
Finally got the trend data from the main substation incoming monitoring relays. Absolutely no incoming voltage spike.

I'm still at a loss. :happysad:
This be a local substation to your plant or a POCO substation some distance away?

Switching capacitors or some other plant going on/off line with a major load, changing taps at regulators all can cause voltage fluctuations or spikes - that may not show up at a distant substation.
 
Status
Not open for further replies.
Top