FAILURE OF 10 MVA TRANSFORMERS

[achilufya]

From the time i joined my organization, i have witnessed failures of 5 x 10 MVA, 33/11kV, Delta/Star transformers at one particular substation. They all fail in the same manner with the LV winding being down to earth. The latest one only worked for 6 months from the time it was installed. We decided to install a more intelligent relay on the 11kV side,the siemens 7SJ6221 and enabled the overcurrent and e/f, Negative phase sequence, and thermal overload. The main load supplied by the transformer is a textile plant with one boiler and the pumps. Additionally, there is REF HV, REF LV, standby E/F and this time around, flags were found on both REF HV and REF LV. The siemens relay also saw the fault and issued a trip but to our surprise the interrupted currents did not even reach the setting. We suspect the electromechanical relay REF LV could have issued the trip first. Could anybody have any idea of which quantity is burning the transformers so that our protection could be directed towards that? A power quality meter was installed during the life of the last transformer and the highlights of what was captured can be availed to you learned colleagues. The Sigra results within the relay will be downloaded and availed to you especially the one on the day of failure.

 

[RRaghunath]

What is the type of earthing on 11kV, is it high impedance type.

What is the rating of largest motor, type of driven equipment and frequency of starting on the 11kV side.

Does the transformer trip coincide with starting of any motor?

Any VFDs or other harmonic producing equipment in the 11kV system and how was the operating temperature of the failed transformer.

These are some questions. But, such failures require detailed study and I am afraid may not be easy to find solution through such forums.

 

[slavag]

Yeh, 5 trafos.
Possible add to Raghun's list additional Q's ( lot of Q's).
As I saied to customer: this case request work on the site, not via phone or e-mails.
Old or new protections are secondary.
First: what is a reason for this number of faults.
5 trafos, it's crazy. Start from primary installation, you have some serious problems:
1. earthing/grounding of trafo
2. harmonics
3. cable installation
4. cable shield earthing/grounding
5. unbalance loading
etc,etc,etc...

And last, are all 5 trafos from same mnf.?
maybe problem was at trafos?
I recommend check this also, it seems fantastic and stupid, but possible.

For protection issue, please send us scheme of CT's connection, current circuit wiring size, size of resistor's and varistor's used in the HI protection.
Good Luck to You.
Best Regards.
Slava

 

[ScottyUK]

How old are these transformers? Are you doing DGA and furan analysis on these transformers? Have you got any supplementary instruments such as a Hydran? What are the results? Any obvious trends?

With the losses you are taking at the moment (five transformers!) I would be considering an online monitoring system such as those marketed by ABB and GE Syprotec. They're expensive and normally they're only found on large critical transformers for economic reasons, but with this casualty rate it may pay for itself.

Are these transformers from the same manufacturer? Are they from the same batch (i.e. you have a large substation with many transformers) or do transformers of different ages keep failing in one specific location? You may be looking a a type fault with the design, or it could be something peculiar about your installation. Roughly where on the winding does the fault occur - line and, neutral end, or randomly? Is it always the same phase affected? How is the system earthed?

What has the manufacturer said about the failures? Request a stripdown of the core & windings with an independent witness to oversee the process. If it is a type fault the manufacturer won't want to admit it - independent experts are invaluable.


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

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

 

[slavag]

Sorry, Q not on the issue.
Scotty.
Are you used Hydran system ( or something like to this)?
It's very intresting for me. Are you connect signals from oil on-line monitor to trip or to alarm only?
regards.
Slava

 

[ScottyUK]

We have presently only have one Hydran on a GSU transformer, but it's location is not in the main oil circulation path and so I don't trust the readings as much as I do the DGA results. I need to make some effort to get the instrument moved onto a better tapping point. One of our GSU's has a hot fault somewhere which we have not managed to identify in spite of two close internal inspections. This GSU gets closer attention than the others because the DGA shows that a small (low volume) but hot (gas signature) fault exists. The hydran should have assisted with the monitoring. We had a long close look at the ABB T-Monitor which is (was?) based on the Syprotec design, but decided not to install it because the fault appears to be stable and relatively benign, and has remained that way for at least 10 years.


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

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

 

[slavag]

Scotty thank you for the information.
In my area it start as new trend, few companies installed Hydran's within last year. One customer ask us connect them to SCADA via DNP or Modbus. Few power plants intrested at T-monitor. I heard on this moment only marketing words,see at the future, what is a benefit of oil on-line monitors.
Regards.
Slava

 

[cranky108]

My first thoughts would be grounding. That's the issue I've seen as a common theme in high failure substations.

But I haven't seen that many failures in one place.

 

[burnt2x]

From the OP,
[q]"The siemens relay also saw the fault and issued a trip but to our surprise the interrupted currents did not even reach the setting."[/q]
Did you think before posting this? Relays don't activate if currents do not break-over the set pickup values and all other constraints compied with!

I guess we don't have a complete picture here, and I might just as well join the great guesswork of the day!
1. too large harmonics content on the load
2. bad/ inappropriate grounding
3. transformer sizing/ rating (K rating)
4. relaying problem
5. harsh transformer operational environment, etc.

Normally, we do modelling of the system (representation of the system on paper) and take off from the theoritical results, careful to temper the settings with other factors that comes witht the operations of the equipment.

 

[slavag]

Hello.
Burnt2x.
Is not so on the OP, but it is intresting point and I would like explain something.
"Relays don't activate if currents do not break-over the set pickup values and all other constraints compied with!"
In lot of digital/numerical relays you have:
Retardation time is also called overshoot time.
It's time what is needed to cancel the tripping if the measured signal goes below the resetting level.
Example:
If we have a setting time 100 ms and the retardation time of the relay is 20-30 ms, and we have a signal exceeding the set pick up level and the signal
goes below the the set level after 70-80 ms, and result is tripping. If the signal goes below the level before 70-80 ms, the relay has enough time to cancel the tripping.
Now Q's what is a value relay save in the fault report:
usually after 100ms and this value is below setting of relay. Is happend and often happend at the motor protection relay (at the start time). Problem is: thie issue isn't explain at any manuals.
Of course need check this issue with relay mnf.Possible other options, but need more more information about relay and DR of fault.
Regards.
Slava

 

[burnt2x]

Slava,
I kept my focus on the phrase I tried to quote (got it wrong)! If the DR from relay indicated values well below the settings, the scenario you just posted is not what the poster meant. The fault current did not exceed and then fell back as you tried to explain. "the interrupted currents did not even reach the setting"

Manuals explain trip initiation reset if the fault current goes below the setpoint, and you're dead right about that. Could it be that the relay was erroneously set?

 

[burnt2x]

Would like to add a little:
REF (Restricted Earth Fault) takes care of faults to ground.
Q: Is it not a protection engineer's usual design criteria to use definite time trip function?
I was thinking about the problem and realized that the normal current on the earth fault relays would be nil (no fault) and will suddenly appear (no slope whatsoever) when the earth fault occurs!
Granting the scenario you gave did happen, with the 100ms (normal delay inherent with EM relays) the damage could not have reached the point where the trafos gets toasted. Maybe achilufya needs to refer to the transformer damage curve to verify this item.

 

[slavag]

Hello Burnt2x.
At this case all possible.
Problem is, that Achilufya didn't back to us and didn't gave enough infomation. Maybe his company like pay for new trafos, I don't know . ;-).
We can continiue play with our ideas on the issue ;-).

Good Day and Good Week to you all.
Regards.
Slava

 

[achilufya]

Thank you very much for all the replies. Let me build up a good case(get more information)so the speculations can come to an end and i am quite confident your suggestions will go a long way.

 

[stevenal]

burnt2x,

REF relays only trip for faults within their protection zone. This zone is typically inside the transformer. The REF does not prevent all damage to the transformer, it protects the system from the faulted transformer and keeps the transformer damage to the minimum. The OP says the damage is one phase grounded on the secondary side, so the protection operated properly.

achilufya,

Are the five transformers operating in parallel? How closely matched are their impedances? Are they equipped with LTCs?