Transformer Inrush Curves 4

[111R]

I typically see transformer inrush plotted as a per unit value of normal transformer FLA current on a phase TCC at a specific point, such as 0.1 seconds.

I have also come across transformer inrush current curves. So, the inrush may be plotted as 10 per unit FLA at 0.1 seconds and 2 per unit FLA at 10-20 seconds.

What does this mean in normal applications? In a worst case scenario with opening and re-closing at opposite voltage zero crossings, does it mean that the inrush will be 10 per unit for 0.1 seconds and then decay down to 2 per unit after 10-20 seconds with infinite bus? Or, does it mean that if system fault current is limited to less than 10 per unit, the inrush will take longer to dissipate?

 

[Mbrooke]

Typically a substation has two or three transmission supplies. A trip and re-close only results in a 5 cycle voltage dip while the fault is occurring. With lines out for maintenance or during a storm a station may have only one 115kv or 230kv feed in service. A trip on that line will drop all the load for 15 seconds then close back in. Inrush is not just limited to the two or three 60MVA-80MVA units in that station but also the inrush of 34.5kv-12.47kv intermediate transformers. In addition to the normal hot load inrush.

 

[jrostron]

This sounds a bit like a FIDVR situation where the source will not support the load with inrush. In other places, it is largely from starting highly inductive loads like air conditioners or motors where the start current is many time the load current. this sounds like much more than just transformer inrush.

Assuming that this is the case, the choices are to shed load or kick in some large capacitor banks on the energized line upstream to compensate for these large inductive loads. A still another effective but more expensive solution is to put in StatComs for this situation.

Most likely for this contingency condition, shedding load is the best solution. That is dropping the 34kV circuits and starting them one by one.

Does this make sense as a solution?

 

[Bronzeado]

As far I understood, you have three paralel transformers of 230/34.5kV, 60-80MVA, feeding various "small" paralel transformers of 34.5/0.440(?)kV in series with the "big" transformers.
If it is correct, I would suggest you to use controlled switching devices to enerzige the big transformers and energize the small transformers sequentially with interval of at least 1 minute.

Cheers,

Herivelto Bronzeado

Cheers,

Herivelto S. Bronzeado
Brasília, Brazil

https://www.linkedin.com/in/heriveltobronzeado/

https://www.researchgate.net/profile/Herivelto_Bronzeado

 

[Mbrooke]

34.5kv-12.47kv, delta wye, each about 10MVA. 12.47kv then runs along overhead poles feeding transformers stepping down to 120/240, 120/208Y, 120/240D and 277/480Y.

 

[stevenal]

Perhaps we should agree on the lingo.
Per IEEE "parallel" transformers have primary windings in parallel and secondary windings in parallel. By extension, series transformers would have series arrangements of both sets of windings. A "tandem" or "cascade" arrangement is where the secondary of one transformer feeds the primary winding of the next.

To answer the question: No, we have never added the inrush curves of hundreds of distribution transformers together with that of the upstream substation transformers to see what transmission reclosing will do; nor have we implemented any kind of automatic feeder clearing and sequential reclosing for such an event. (Don't forget about the customer owned transformers further downstream) The transmission simply trips and recloses. The curves represent the absolute worst case for transformer inrush with voltage maintained, and are used for transformer primary protection, and not upstream protection. If hot load inrush caused a problem, we would review the event and adjust settings as needed.

 

[Mbrooke]

How is it that distribution transformer inrush can be ignored?

 

[davidbeach]

Distribution inrush can't ignored. But typical distribution practice has instantaneous blocked on manual close or auto reclose. The inrush isn't enough, for long enough, to cause time overcurrent to trip.

On the Transmission system, phase protection is much more likely to be distance rather than overcurrent. Transformer inrush generally doesn't present a low enough impedance to get into the zone 1 characteristic and doesn't trip.

I've never seen a load serving transmission system that reclose into any significant amount of transformation, certainly not 60-80MVA worth, maybe less than 40MVA total base rating, and then only with multiple contingencies.

But, my first exposure to this was a 230kV line to a wind farm with multiple large 230/34.5kV transformers and strings of 34.5/0.48kV and 34.5/0.69kV transformers. The 230kV line had tripped, but nothing at the wind farm. Initially couldn't get the line to pickup the transformers without tripping. There were some interim setting changes made, but the principal solution was to implement dead station tripping to trip the GSUs and the collector breakers so that the GSUs would be picked up one at a time followed by the collectors one at a time.

 

[Mbrooke]

I know- typically with all lines in service you wouldn't' be dropping 60MVA worth of load then picking it back up- but certain events (like Hurricanes, Ice storms, scheduled outages ect) can leave a station with only one supply.

 

[stevenal]

Mbrooke,

I never claimed you could ignore inrush. In your last post, you seem to be confusing transformer capacity with load.

David,

I just counted up 142 MVA of self cooled transformer capacity that our supplying transco recloses on per my understanding. The short line has never faulted that I'm aware, but they have no way of sectionalyzing our system beyond that line. We in turn supply an industrial load with 50 MVA self cooled capacity. We have reclosed successfully on this line.

 

[davidbeach]

Steve - Where? I'm seeing a bunch of connection points on the big system that say they're connecting to you, but no detail on your system.

But, like I mentioned, transmission protection is generally going to be less sensitive to reasonable amounts of inrush; hard to get into zone 1.

 

[Mbrooke]

Capacity with load? My point was that when a station is supplied by one line, and that line trips, all load connected to that station is dropped. Upon reclosing its picked back up.

Fortunately I'm hearing load damps out trafo inrush. 115kv-34.5, 34.5kv-12.47kv, 12.47kv-120/208Y... 3 layers of transformation before hitting the load.

My bad for any confusion and my mistake for throwing random numbers around. I apologies for muddying the waters.

 

[prc]

1) [tt][/tt]Bronzeado,thank you for your interesting paper on series sympathetic interaction.But have you ever noticed such phenomenon in field anywhere? I also believe actual situation in field will be different from lab tests due to voltage drop in line impedance.
2) Let us consider a 220/132 kV auto transformer feeding a city. Auto feeds to couple of 132/33 kv units and each 132/33 kV unit feeds few tens of distribution transformers. When the auto trips, all units are de-energised. My understanding is when the auto is charged again, inrush will not be the sum of inrush of all units and there is no need of energising down side transformers one by one. This is because the down side transformers will be seeing voltages less than rated values due to line drops from the inrush current. Probably at wind and solar farms this voltage drop will be low and hence increased inrush current value.

 

[stevenal]

I took the where question off line.

Mbrooke,

Magnetizing inrush is measured in per unit of FLA. The OP uses 10X and 2X. If load remains connected, I suppose you could figure total hot load inrush to be 11X and 3X. It doesn't move the curve very far on log-log.

 

[Mbrooke]

Just add trafo inrush together plus the load itself? Better than I assumed.

 

[stevenal]

Yes, unless you consider the all the transformers in cascade as you mentioned before. Actually you failed to mention all the door bell transformers and those within appliance power supplies, etc.

 

[Mbrooke]

There are those, but I'm more worried about R/AC compressors