Jump to content


Photo

Multiple Vsd Failures


  • Please log in to reply
14 replies to this topic

#1 JonAW

JonAW

    Intermediate Member

  • Full Member
  • PipPipPip
  • 14 posts

Posted 25 November 2009 - 04:28 PM

Hi guys, I just registered today as I have run out of options and can only hope there are smarter people out there than we have had in on this site. Basically below is the story and if you have time i'd appreciate you reading it and giving me some ideas no matter how trivial they may seem.

The company I work for got a contract to supply and fit the electrical control panel and wiring to motors etc of a waste recycling plant. We have done about 50 of these plants, maybe more, with the stipulation that we bear no responsibility for the electrical supply to our control panel, that lies with the customer. We have had no faults in the past.

This contract required a slightly larger panel than anything we have done before but not by much. It consists of a large PLC controlling about 30 motor starters and 10 variable speed drives (inverters). Plus a few other bits and pieces. Each VSD has its own input reactor (choke). The supply is from a 550kva generator, under full load our plant only draws around 120kva. (there is a terrible power factor of 0.6 so the apparent power used is only 80kw.)

Basically what is happening is that every day or so a vsd fails, always at the input side ie. the dc link is completely wiped out. We have tried different generators with varying degrees of success. With one generator we had no failures for 2 weeks but then the generator itself failed. We put a power analyser on the panel and see very little voltage fluctuation, about 5volts as the plant comes on and off load. We have measured harmonics with a result that we have 2.5% deviation on the voltage and 3% deviation on the current. Frequency could not be more stable at 49.9hz. Nothing else has failed on the plant except a 230vac/24vdc psu which failed this morning.

We tried changing the input chokes to output chokes to see if that would make any difference but we are still having failures. There has also been a power quality expert on site and he can find no fault with the supply and no fault with our installation. I know the supply has nothing to do with us but we have to help in finding a solution as we are replacing the vsd's until a conclusion has been found. Plus I have now been onsite for nearly 4 weeks incase there is a failure so that I can fix it as quick as possible. The thing is I am fixing the symptoms not the problem.

Can anybody offer any help or advice, we have completely run out of ideas and the company that bought this plant is losing money hand over fist as the downtime spent in solving different situations is crazy.

thanks

Jonny

#2 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,602 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 25 November 2009 - 06:06 PM

Hello Jonny

Welcome to the forum.

From your description, there is no doubt that there is something happening at the input of the VFDs that is causing them to fail, or they are faulty VFDs.

I am unable to give you definite answers to your problem based on your description but there will be an answer.

1. What sized and brand are the VFDs?
2. Is there power factor correction on the site?
3. What else is being switched near the VFDs?
4. Are the inputs of the VFDs being switched? i.e a contactor on the input?

Best regards,
Mark.

#3 JonAW

JonAW

    Intermediate Member

  • Full Member
  • PipPipPip
  • 14 posts

Posted 25 November 2009 - 08:34 PM

Hi

The drives are Weg drives mostly 4kw with a couple of 3's and one 5.5kw so all small drives. There is no power factor correction as the mains is generator powered. The generator is producing a perfect sine wave and as stated before harmonics are minimal when all load is applied. The nature of the plant means that once up and running there is no switching of any component. The inputs of the vsd's are only broken when an emergency stop is activated. Although due to motors jamming etc the emergency stop is used incorrectly by the plant operators to stop the plant. This occurs on average once every 1 and 1/2 hours and as the plant is a 24hr system this is far too frequently. The thing is that this is true for many of our other plants and we have no problem, however this is the first time we have used weg vsd's, we normally use hitachi. In the words of our vsd suppliers ' there is a catistrophic failure of the dc link'. In my mind this can only be caused by a high voltage spike across the input.

A vsd failed today again whilst my power analyser was on. There is no other signs of failure in the vsd other than a rise of 18A in L1 and L3 where the short occured. We also had a transient meter on at the same time and no sign of anything ontoward. Each vsd is protected by a manual starter which are really being tested!! Hope this extra information helps. Thanks for helping.

Jonny



#4 ggalea

ggalea

    Junior Member

  • Members
  • PipPip
  • 4 posts

Posted 26 November 2009 - 05:44 AM

QUOTE (JonAW @ Nov 25 2009, 03:34 PM) <{POST_SNAPBACK}>
Hi

The drives are Weg drives mostly 4kw with a couple of 3's and one 5.5kw so all small drives. There is no power factor correction as the mains is generator powered. The generator is producing a perfect sine wave and as stated before harmonics are minimal when all load is applied. The nature of the plant means that once up and running there is no switching of any component. The inputs of the vsd's are only broken when an emergency stop is activated. Although due to motors jamming etc the emergency stop is used incorrectly by the plant operators to stop the plant. This occurs on average once every 1 and 1/2 hours and as the plant is a 24hr system this is far too frequently. The thing is that this is true for many of our other plants and we have no problem, however this is the first time we have used weg vsd's, we normally use hitachi. In the words of our vsd suppliers ' there is a catistrophic failure of the dc link'. In my mind this can only be caused by a high voltage spike across the input.

A vsd failed today again whilst my power analyser was on. There is no other signs of failure in the vsd other than a rise of 18A in L1 and L3 where the short occured. We also had a transient meter on at the same time and no sign of anything ontoward. Each vsd is protected by a manual starter which are really being tested!! Hope this extra information helps. Thanks for helping.

Jonny

Hi Jonny: I need to ask you some questions, just bear with me:
1. What fuses do you have on line side? Fast or slow blows. How are they sized? ie 1.5 times the input current, keep in mind 10amp minimum regardless of the size of drive.
2. How are the drives normally or usually "stopped"? Is a control, 2 wire, 3wire or key pad "stop" used? Or, is the power cut and then energized again? Watch out for operators who love to hit the E-Stop! This will blow up the drive's DC Bus (caps) in about 3 or more actuations of cycling the E-Stop rapidly! I believe you MUST NOT power up the drives again until you wait 6 minutes; that's WEG factory spec if my memory serves me right. Only higher end drives protect against this like Allen-Bradley 1336 as they have a timer on the bus when power is cut. It will not turn on again until the caps drain. This is good except a pain in the butt when you need to get going.
3. Is the line power at 50 or 60hz? Is the drive configured in the apprpriate parameter settings for the power in?
4. What is actually frying on the drives? Is the same thing happening to all the drives?
5. Are the line reactor 3 or 5%Z(impedance)? Are you using an MOV Surge Arrester (Metal Oxide Varistor) in conjunction with the line reactors?
6. Are the enclosures cooled? Have you done a thermal or j junction heat dissappation calculation for the Watt loss of the (each) drives ie 3kW is about 87Watts? Where are the drives located in the enclosure at the top or the bottom - the layout? Arwe they one underneather each other or are they staggered/off set so the heat from one doesn't get sucked up into the heat sink of the the other? Where are the line reactors ie beside the drives or underneath them.
7. Are the drives set for vector or variable torque mode? How slow do the motors run? If below 30hz in vector mode, do the motors have a blower kit on the bell ends?
Let's start with that...
Thanks, George

#5 JonAW

JonAW

    Intermediate Member

  • Full Member
  • PipPipPip
  • 14 posts

Posted 26 November 2009 - 09:31 AM

George you gave me a lot of q's to answer, I appreciate the time taken to ask them. I'll go through them one by one.

The drives are each protected by a telemechanique manual starter which has overload and short circuit protection built in. Each manual starter is matched to the rated input current of the vsd's. I have never had a manual starter trip when its protecting a drive unless the drive has been faulty. After the manual starter we have a contactor which as stated before only drops out when an e/stop is activated.

The drives use a 2 wire control to start and stop them using a relay. The operators are supposed to use the push buttons on the panel to start and stop the system, but as you correctly stated an e/stop is regularly pulled out of lazyness.

the line frequency is 50hz, to be honest I haven't checked the settings on the vsd's, i'll do that this morning.

The line reactors are basic line reactors sized to each vsd. They were supplied to us by our vsd specialist for this particular job.

Yes the enclosures are cooled, and due to the size of them I reckon they are not heating above 25oC. The drives are at the bottom of the enclosure fitted side by side using the manufacturers minimum spacing requirements. The reactors are underneath the drives but a little distance away.

All the drives are vector control. All motors have rear fans and we were advised by the company who supplied both the motors and the vsd's to run them no lower than 15hz. To be honest the motors are not overheating as they are oversized for the job and current levels are lower than expected for this same reason.

Hope this helps, thanks for the replies



Jonny




#6 JonAW

JonAW

    Intermediate Member

  • Full Member
  • PipPipPip
  • 14 posts

Posted 26 November 2009 - 10:20 AM

i just checked the manual and apparently the drives take either 50 or 60hz input, no setting up required.



#7 JonAW

JonAW

    Intermediate Member

  • Full Member
  • PipPipPip
  • 14 posts

Posted 26 November 2009 - 12:14 PM

I have just been looking through some posts similar to my problems.

As stated i have input reactors on my vsd's but no output reactors. i have come across the term capacitive coupling and I'm thinking that this may be affecting our installation. The longest cable run that we have is 80m on a vsd. The shortest is 18m. All our motors are wired using swa cable bunched on cable tray. The armouring of the cable is directly earthed to the motor frame and the control panel which in turn is connected to an earth spike, measured at just 1 ohm impedance. Our vsd's are bolted directly to the steel backplate of the panel as well. As i understand it capacitive coupling affects motors as well as vsd's but we are not seeing any motor failures or signs of pending failures such as overheating or overcurrent faults. Is it possible for this effect to damage our vsd's but not our motors? The two vsd's which are giving us the most problems are a 3kw which drives 2*1.1kw motors 80m away from the drive, and a 5.5kw running a 5.5kw motor 34m away. Although these are the worst, all drives have failed at some point.

regards

Jonny

#8 yuri

yuri

    Member

  • Full Member
  • PipPipPipPip
  • 93 posts

Posted 26 November 2009 - 03:28 PM

Hello.
Contactors as I know are not used before VFDs, motor-starter is no good at all protection of a VFD system - too late tripping point (need be quick-action fuses).

#9 AB2005

AB2005

    Senior Member

  • Full Member
  • PipPipPipPipPip
  • 406 posts

Posted 27 November 2009 - 04:18 AM

QUOTE (yuri @ Nov 26 2009, 08:28 PM) <{POST_SNAPBACK}>
Hello.
Contactors as I know are not used before VFDs, motor-starter is no good at all protection of a VFD system - too late tripping point (need be quick-action fuses).


Hello JonAW,

Very intrusting case is it.
I want to share with you my personal experience. I have faced such problem at our corrugator machine which have 10 ABB servo drives for web tension control. We bought the machine from Italy where those drives were worked correctly but here, in our country, 9 drives malfunctioned in only 1 year and they also did not work correctly. I worked the same like you. The enclosures were attached with machine body which subjected to create little vibration. I removed the enclosures from machine body and installed on separate frame and, believe me, 80% problem solved. The days were June July where temperature was so high. But now, when temperature is very good, no any drive malfunctioned in last 4 months and they have been working perfectly.

So, I want to ask some questions;

1) Are the enclosures attached with machine body or separated?
2) What is the inside temperature of enclosures?
3) Can you give us the wiring diagram of system?
4) What component usually damage in VFD?
5) Did you ask to manufacture about these failures, what are their comments?

One possibility is that there is some problem in control circuit which may be subjected to contactors chattering. Anyhow, if you send the wiring diagram, we may be able to understand the problem.


Best Regards
"Don't assume any thing, always check/ask and clear yourself".

#10 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,602 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 27 November 2009 - 08:14 AM

Hello Jonny

In your first post, you said that the input of the VFDs have all failed.
I was presented with a problem at a site where the "inputs failed" and the finger was being pointed at the generator supply. When I analyzed the failed VFDs, I found that the inputs had failed, but in addition, the outputs had failed also. - The inputs failed due to the failure of the output stages.
I found that the real problem was that the output was being severely overloaded and the VFD was not able to protect itself.
Have you checked that the outputs are all OK?

In a recent post, you talk about two motors on the output of one VFD with a distance of 80M. If each motor is fed by a separate cable from the VFD, then you must add the two cable lengths together so the cable on the output is actually 160M.
What is the maximum cable length that you can connect to the output of the VFD? Many VFDs can only drive 10M - 20M of screened cable. If the cable length is too long, you can cause a failure in the IGBTs and this will cause a failure in the input rectifiers unless there is a DC bus fuse.

Best regards,
Mark.

#11 JonAW

JonAW

    Intermediate Member

  • Full Member
  • PipPipPip
  • 14 posts

Posted 27 November 2009 - 09:23 AM

Hi guys thanks for the replies, i'll try to answer your questions 1 at a time.

Yuri, we must use contactors infront of the VSD as a means of a mechanical break in the supply in the event of an emergency situation. The machine falls into the highest category of danger therefore we must make it as safe as possible. The WEG manual for the vsd states that it cannot be used as an emergency stop device and mechanical means must be applied. I take this to mean that the control signals cannot be used as a reliable way of cutting power to the motor in the event of an emergency. I cannot comment on the use of fast blow fuses although our power logger records a short circuit of only 18A before the supply is cut when a failure occurs.

AB2005, the machine is separate from the panel, temperatures have not changed much in the panel due to good ventilation and a large panel space. Although I have not measured the temperature, when the panel door is opened there is nothing that would cause me concern. There is definately no contactor chatter as I have been watching the control system for 3 weeks and everything seems fine. I have seen chatter in the past so I can definately say this is not happening. On most of the vsd's there has been damage to the dc link only, on one of them there has been damage to both the input and output. Weg are currently examining the vsd's to compile a list of possibilities.

Marke, the 2 motors are being fed from a 7 core 2.5mm cable. Last week I changed the input reactors to the output on most of the vsd's and have had no failures since. The only failure I have had is on a vsd with no output reactor.

Thanks for the replies. I hope this is the problem as we really need to get to the bottom of this very soon.

Jonny



#12 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,602 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 27 November 2009 - 07:39 PM

Hello Jonny

QUOTE
Last week I changed the input reactors to the output on most of the vsd's and have had no failures since. The only failure I have had is on a vsd with no output reactor.

I think that is your answer, you are using too long a cable run on the outputs. This results in high capacitive currents and damages the IGBTs which in turn causes the input bridge to fail.

Have a look at the maximum screened cable length for the VFDs that you are using. With some VFDs, this can be as low as 5M.
Large VFDs may allow for 150M, but small ones are commonly much less. If there are multiple output cables, these must be added together.

Best regards,
Mark.

#13 ggalea

ggalea

    Junior Member

  • Members
  • PipPip
  • 4 posts

Posted 28 November 2009 - 07:04 AM

QUOTE (JonAW @ Nov 26 2009, 04:31 AM) <{POST_SNAPBACK}>
George you gave me a lot of q's to answer, I appreciate the time taken to ask them. I'll go through them one by one.

The drives are each protected by a telemechanique manual starter which has overload and short circuit protection built in. Each manual starter is matched to the rated input current of the vsd's. I have never had a manual starter trip when its protecting a drive unless the drive has been faulty. After the manual starter we have a contactor which as stated before only drops out when an e/stop is activated.

The drives use a 2 wire control to start and stop them using a relay. The operators are supposed to use the push buttons on the panel to start and stop the system, but as you correctly stated an e/stop is regularly pulled out of lazyness.

the line frequency is 50hz, to be honest I haven't checked the settings on the vsd's, i'll do that this morning.

The line reactors are basic line reactors sized to each vsd. They were supplied to us by our vsd specialist for this particular job.

Yes the enclosures are cooled, and due to the size of them I reckon they are not heating above 25oC. The drives are at the bottom of the enclosure fitted side by side using the manufacturers minimum spacing requirements. The reactors are underneath the drives but a little distance away.

All the drives are vector control. All motors have rear fans and we were advised by the company who supplied both the motors and the vsd's to run them no lower than 15hz. To be honest the motors are not overheating as they are oversized for the job and current levels are lower than expected for this same reason.

Hope this helps, thanks for the replies



Jonny


Thanks Jonny: We have all been in your shoes and it gets better as you gain experience on these jobs. Unfortunately, you always are learning on the spot.
1. I strongly recommend fast blow fuses after the manual starters on line side of the drives: see http://www.highspeed...m/abouthsj.html
2. It sounds like Reflective Wave Phenomena aka High Voltage Due To Long Leads. ie: if you have a 10HP drive with three 2hp motors on it and each motor lead is 50 meters then the equivalent effect (or circuit) is 150 meters as seen by the drive and the motor. the motor will experience Double the rated voltage and the drive will be stressed. see http://www.mtecorp.com/motor.html or http://www.transcoil...putSide/V1k.nws for dv/dt motor protection. You can typically run at a Carrier Frequency of
1.5 to 2.5Khz for about 30 meters with this problem of reflection. After that you need a RLC (dv/dt tank circuit) Load Filter. So, if you have 3 motors on one drive that means no longer than 10 meters each motor lead.
3. So, here's the KICKER! I have never know any drive manufacturer who allows multiple motors on a vector torque application. It is impossible for the drive to go into tune or calibration mode. This requires the drive to signal and pulse the motor for rotor resistance etc. which, beforehand, the motor nameplate data has been entered into the drives parameters. Constant torque which is volts/hz yes, but not vector torque control. This means your drives may not be in vector mode only volts/hetrz. If they are in vector then the wrong data sets are employed and we may be seeing the end results.
Hope this helps,
George

#14 JonAW

JonAW

    Intermediate Member

  • Full Member
  • PipPipPip
  • 14 posts

Posted 30 November 2009 - 01:31 PM

Thanks everyone for the replies, I have acted on this forum's advice and applied input and output reactors to each vsd. Hopefully this helps although I don't think we are out of the woods yet.

Another system designed by a german outfit, which seems to be very well made, has just blown up a telemecanique altivar 71 250kw drive. The drive seems to have destroyed the input bridge as I was getting zero resistance between L2 and L3 with my ohms meter. This system is on the same site but supplied with a separate generator. The only common link between the two generators is the earth in the steelwork of the building and of the recycling system itself. The control panel is sited only 10m or so from the motors so excess cable length is not to blame this time. Another point is that the vsd has no reactors or filters on either the input or output. Before it short circuited we were measuring voltage harmonics of 30%+.

I neglected to explain at the start of the thread that quite a few other vsd's have been destroyed on the same site, which were not supplied or installed by our company. All have had input bridge failure. There have been numerous generators changed on site to try to combat the problem. They even decided to rip out all vsd's that weren't needed to limit the problem but by the time the other vsds were taken out, it only left our panel giving problems, or so we thought until this morning. The only other common link throughout the entire system is a generator which has now been relegated to lighting only. Again this is only being linked with the rest of the plant through the structural steel earth. Occasionally the lighting dims as if the gen set cannot supply enough power to the lights, but it is a 550kva generator running 20kw of lighting!

The reason I neglected to mention this is because we had completely isolated our control panel from all other supplies, apart from the earth, and we were the only system which seemed to be affected for about 5 weeks now. Throughout the entire time there has been no motor failures on any system at any time.

I suppose I have now confused everybody as much as our power experts are but if there are any ideas out there i would like to hear them.

#15 anu_rags

anu_rags

    Intermediate Member

  • Full Member
  • PipPipPip
  • 35 posts
  • Gender:Male
  • Location:India, Vadodara

Posted 05 December 2009 - 01:33 PM

Hi Jon,

Ill like to add as I think its useful in your application:

Use a VFD that has Half controlled bridge rectifier with half thyristors and half diodes. It eliminates the need of Charging resistor and charging contactor in the VFD. Therefore you can switch ON/OFF at the input without any limitation. Ofcourse after powering ON, the operator should wait atleast for the capacitors to charge 5sec before giving start command to motors.

Emotron make VFD can be programmed in Direct Torque control mode (similar to Vector) in multi motor application and can be tuned for static tuning.

Regards,
Anurag




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users