AC Drive FAILURE 6

[Elec777]

AT OUR PLANT (SUGAR REFINERT)THE CENTRIFUGAL MACHINES MOTORES ARE FED&CONTROLED BY VFD ,THE DRIVE RATINGS:380V/3PH/490AMPS ,THE MACHINES AND DRIVES WORKED FOR 2.5YEARS VERY WILL ,AFTER THAT We recently started getting heavy hardware failures in THESE AC Drives, Here are the failure symptoms:
The breakdown starts by short circuit in two out of six IGBTs of the motor power Block
,consequently the IGBTs in the another power Block
(DC power block) explode and the associated IGBTs firing control card got cracked and damaged ,
Also we found one power line fuse had been blown up and the feeder breaker(630amps) in the switchgear
Was tripped.
We also noticed the following:
•? ? ? ? ?The failure happens while the machine is running without prior alarm
•? ? ? ? ?the same identical failure happened in two different drives so far with time span of one week.
•? ? ? ? ?the MCC where the drives are located is clean and with air conditioning
I HAVE REPLACED THE DEFECTIVE BLOCKS WITH NEW ONES
AND EVERY THING IS ok NOW,COULD ANY FOLKS GIVE EXPLANATION
FOR WHAT ARE THE REASONES BEHIND WHAT HAPPENED ,IM STILL INVISTIGATING THE PROBLEM.
THANKS IN ADVANCE.

 

[Skogsgurra]

Hello Elec - and welcome to ENG-Tips. The best place for professional problem solving you will ever find. And it is free!

Need to have a few questions answered:

Is the grid with grounded neutral or isolated neutral?

Are all drives that failed on the same transformer?

Do you have automatic power factor correction installed? (Capacitors being switched in and out of circuit).

Also, it is not possible to see what your failure codes are - could you please write them again, with letters and numbers?

And, finally, what brand are the drives? Siemens? ABB? or other?

Also, we try to reduce the USE OF ALL CAPITALS. It is not common usage and there is a reason why the common letters exist.

Gunnar Englund

http://www.gke.org

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

 

[Elec777]

Dear Skogsgurra,
most of the time the plant is getting power from 10MW
steam turbine with grounded nuetral,and short time from neutral grounded Grid ,the failed Drives are bieng fed from two different transformeres 1600KVA each, we do
have the KVAR auto. compensator but we rarely use it.
the drive is Sinamics S150 of siemens and during
fault the drive screen reads:
(3:vector) fault 0622:
30022 PU: monitoring U_ce
cause: the collector-emitter voltage (u_ce)
of the semiconductor responded.
i didnt change any thing in the system setting ,i just replaced the defective blocks with new ones and resume working trouble free.
gratefull to your help

 

[Skogsgurra]

Where are you in the world?

An U_ce failure can be either because of long cables (capacitive loading), overcurrent (which shouldn't happen) or weak gate drive (fiber optic system, usually).

The U_ce failure can also be a secondary effect of the transistor explosion. Really hard to say what is the hen and whet is the egg.

Are these drives still under guarantee? (Guess not since 2.5 year's operation). Are the transistors by any chance SiC transistors? Are there motor reactors?

Gunnar Englund

http://www.gke.org

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

 

[ScottyUK]

Sugar mills have a repuation for having high levels of hydrogen sulphide in the atmosphere. Silver whiskers growing in a drive could certainly cause problems if they were on either the control circuitry or on the power stage. They often have the annoying trait of leaving no evidence as they are typically vapourised by the fault, leaving seemingly inexplicable damage.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[Skogsgurra]

Have seen zink whiskers - silver whiskers are new to me (except those formed in ceramic capacitors). They were really tough to find!

Gunnar Englund

http://www.gke.org

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

 

[Elec777]

Dears,
I understand that we are so careful before we state a firm conclusion,the problem that we have in the VFDs is really mysterious,no clues no evidence so far for anything wrong or so,that's why i have taken the risk and replaced the defective blocks with new ones(each VFD has two blocks)and since one month all went fine,however here are some hints:
1-the plant is a sugar refinery ,we refine raw brown sugar and produce white sugar.
2- i found a thin solid layer of sugar covering the triggering cards and i guess that sugar tends to absorb moisture.
3- the drives are get 25%overloaded for 30seconds every 3 minutes.
4-drives are running 24houres /10months/300days
5-we newly have installed the air conditioning after the first drive failed,also we closed all openings in the mcc room.
6- our refinery is in the middle east north africa region
do you think that increasing the Acceleration and deceleration times will give positive outcome?
regards

 

[Skogsgurra]

Deceleration with limited regeneration may cause problems (DC link overvoltage). But, as far as I know, the Sinamics guards against such failures.

Have you talked to other users in refineries if they have a similar problem with these drives? It could be a 'family problem' and then it doesn't help what you do.

I take it that you have used other power electronics equippment in that same environment. Did that cause you any problems? The Sinamics should be protected against harsh environment. There used to be versions that were "Tropenfest und Termitensicher" (Hardened for tropical conditions and safe against termites). I guess that all products are now.

Gunnar Englund

http://www.gke.org

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

 

[electricuwe]

To Skogsgurra:
It will still take a few years from now until you will find SiC- transistors in such drives. Currently you most probably will find Silicon IGBTs as well as Silicon freewheeling diodes in the modules used for such drive application. Maybe for special applications already IGBT modules containing Si IGBT and SiC diodes are already in use. But nevertheless, it could be a dV/dt related problem. But according to Elec777 there is no defect of the motor mentioned.

To Elec777:
To my mind two root causes are quite likely (2 and 3 according your numbering). No. 2 is quite ovious, as this can lead to a lot of malfunctions if the material absorbs humidity. No conformal coating on the boards ?
No. 3: IGBT modules are subjected to aging when exposed to repetive changes of temperature at it's inner connections. No problem if considered when designing the application. This has already has been done in traction drives and other drives for dedicated applications. For GPDs this is difficult, as the drive manufacturer doesn't now what the final application will be when the drive is designed.
On the drive design level as well as on the system design level the method to cope with this is oversizing. Keep in mind that not temperature, but the change in temperature is the reason for this aging

 

[Skogsgurra]

Uwe,

I have information saying otherwise. I was once hired to do measurements on drives that more or less exploded when cable lengths grew, the reason was SiC related. Those were not high-powered. Something like 20 - 30 kW if I remember correctly.

Gunnar Englund

http://www.gke.org

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

 

[Elec777]

dear All,
electricuwe is wright the blocks are with Silicon IGBTs ,
the power section of the drive consists of two identical blocks,motor block and active line block for regenerative function ,i also tends to agree on his idea about early aging because of temperature changes,otherwise how can we explain that new blocks are working trouble free now,however do you think that air conditioning is dangerous and should be removed
or you have certain advice on that,i would kindly ask you if you could advise any recommendations to prevent more VFD failures.
regards

 

[Skogsgurra]

Good that you do not have SiC. They do not seem to be a mature product, yet.

Gunnar Englund

http://www.gke.org

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

 

[electricuwe]

If you want to understand the reason for the problem ask you supplier to provide a failure analyis (even if the unit is so badly destroyed that repair doesn't seem useful). If there is no other reason found inside the unit, the IGBT modules shoupld be forwarded allong the supply chain.

As long as not everthing has been vapourised the IGBT manufacturer will be able to determine if significant aging has taken place.

 

[ozmosis]

I would be looking towards the environment the S150 drives are installed:
1) Temperature inside the enclosure the S150 drives are mounted. Check all filters on the enclosure to ensure they are not blocked. A high temperature inside would not always cause the drive to trip on over-temp but would reduce the life considerably of the DC link capacitors that in turn would potential cause problems with the IGBT's.
2) Contamination that could cause malfunction on the gate circuit and mis-firing on the IGBT's.
3) As this is a sugar refinary, do you have the sugar cane fields nearby? I have seen a number of times the final process in the fields of burning off the stubble creating a high density smoke and this 'air' being fed back into the plant room of the refinary. The air is heavy with carbon strands and this is simply sucked into the drives, lands on the (uncoated) sections of the boards and bang go the IGBT's.

Gunnar-SiC IGBT's are not used on the S150 units (or the G150 come to that). This 'trial' is used only on the G120 PM260 (690Vac regenerative unit). However, if they were used, I agree it is the first place I'd look...

 

[jraef]

Gunnar,
Silver whiskers are real, I have posted on them here a couple of times. thread238-172950


"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[Elec777]

Dears,
Im so gratefull for your smart coverage of the issue,what Im thinking now is to install a powerfull fans oppisite to
the IGBTs card(the IGBTs are itegrated in the firing card 40X20CM area)like the fans used for the processor in the PCs
may be it will help as aditional ventilation and remove the excissive heat generated by overloading.

 

[LionelHutz]

It's likely a component failure or even a component aging (going out of tollerance that caused the failures. The hard part will be determining if it's simply because of age or if the enviroment had an effect.

I would expect that there was both a temperature and a contamination problem, based on your comments that the MCC room was open until the first failure. VFD's should have ambient temperatures below about 40C and they also should not have contamination on their circuit boards. There are a lot of electrolytic capacitors inside a VFD and they tend to degrade quicker when in a hot enviroment. There is some relatively high speed switching inside a VFD and the timing of this switching must be quite precise. Contamination can mess with this switching and cause failures.

There are 2 output modules in parallel so it's no surprize they both failed. You mention a circuit breaker and fuses for line power protection. What kind of fuses? They should be semiconductor fuses that act very fast. The failure of the DC module (the rectifier) was likely because the DC output was suddenly shorted by the failed IGBT and the fuses or circuit breaker were not able to protect it.

 

[Elec777]

seems to be we are all now had agreed on two root causes of the failure:
1- over temperature in the IGBTs cards as a result of overloading and high ambient temparature.
2- the contamination that accumulated on the cards has caused a disturbance in the firing timing.
both 1 and 2 had contributed to early IGBTs aging ,so i would be so grateful if you advice your opinion about my plan to stop the tragedy of these VFD failures:
A- to keep the MCC room temperature under 40C by a powerful air conditioning system.
B- to keep the MCC room tight against dust .
C- to install a powerfull fans oppisite to
the IGBTs card(the IGBTs are itegrated in the firing card 40X20CM area)like the fans used for the processor in the PCs
may be it will help as additional ventilation and remove the excissive heat generated by overloading.

 

[electricuwe]

The IGBT modules themselfs have to dissipate several kW via the heatsink at the bottom side. Ensure that the filters and airducts are clean there.

The driver board looks generously sized and I do not think that additional fans are useful here.

What's the problem with sending parts back for analysis ? If it really pramture aging you have approximately two years of time available to wait for a reasonable result.

On SiC: the drives mentioned by ozmosis do not contain SiC transistors, but just SiC Schottky diodes as freewheeling diodes for Si-IGBTs. Siemens has published several papers on that. These drives are employing an integrated filter so that the motor will not be subjected to excessive dV/dt. However, if there are problems with these drives or the modules inside and someone likes to discuss it within the eng-tips forum, I'd suggest to start a separate thread.

 

[Elec777]

defective block already sent back to Siemens for analyze,as soon as i get it i will let you know the results.
appreciate your comments .