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?

 

[stevenal]

Worst case slow decay to allow setting a relay or fuse above the curve and avoid tripping on inrush.

 

[cuky2000]

We suggest first to check if the linked thread provides some inside regarding the transformer inrush current.

It means that generally accepted assumed deterministic inrush current value. In a real case, this is a random phenomenon and varies depending of the instant of the supply voltage is applied to the transformer and the magnetization level of the transformer. The highest inrush current occurs at zero voltage crossing while the lower values happen at the peak source voltage (90[sup]o[/sup].)

a) After reclosing the transformer it is possible that the core is partially magnetized. The expected inrush may not be the highest even if the transformer closes at zero crossings.

b) During a short circuit the depress but recover in a short time after the fault is clear, t full value, and the inrush peak will depend on the instantaneous value of the source voltage after re-energize the transformer.

 

[111R]

Thank you. For clarity, I'm only looking at the inrush seen in an unloaded transformer (not a loaded transformer with a lack of load diversity upon energization).

Will the magnetization inrush ever stretch out to the 10+ second time range?

 

[davidbeach]

I've seen inrush go on and on and on. What you want is to pick up the transformer and all of its load; the resistance in the load is what damps out the inrush.

Transformer inrush is often categorized as somewhat random; sometimes it's this and sometimes it's that. Ok, in the right context that's true enough; one event on a given circuit will look different from the next event on the same circuit. But it's not random, not at all; it's very deterministic. The problem is that we don't/can't generally know the important variable in advance. That's the point on wave that the transformer(s) was deenergized. A collection of transformers that were simultaneously deenergized by opening of a single breaker (or the second to open of a pair of breakers) and that are simultaneously energized by the closing of a single breaker will present a waveform that looks like the inrush of a single transformer of the combined size of all of the individual transformers. Not random at all; just a lack of sufficient information to predict the result. But I digress.

I don't know about the 10+ sec time frame but I know I've dealt with inrush issues that persisted in to the 1+ sec time frame. For a protection engineer that's moving into the realm of eternity.

The worst inrush event I ever saw was trying to pick up a wind farm. Everything tripped at once, nothing opened in the interim, and everything tried to come back all at once. That included the 230/34.5kV transformers and all of the 34.5/xxkV tower transformers. Didn't work. Didn't work multiple times. Dead station tripping was implemented shortly there after so that the amount of inrush for any given breaker close was greatly diminished and controlled.

Count on it being long and high current and you'll never get into trouble. Try to get away with pretty typical values all of the time and you'll have a really bad experience every now and then.

 

[dpc]

Inrush is a complex event, with many variables. If your concern is setting the primary side overcurrent protection, then the time-tested rules of thumb usually work out. Usually. But if we make the goal being so conservative that you can absolutely guarantee there will never be a misoperation on inrush, you may be not be making the best compromise. Your assumptions on inrush will impact the protection of the transformer and potentially everything upstream of it. Protection and coordination are always in direct conflict.

 

[111R]

Thanks, that helps!

 

[Mbrooke]

Errr... what is the time tested rule of thumb?

 

[cuky2000]

To see how others consider this phenomenon, below is the result of a quick search from various sources that are in disagreement with the above statement.

[sub]• A stochastic signal such as inrush current has two important specifications: at a fixed time instant its value of harmonic content is a random variable, and as a function of time the random variables might be interrelated.
• The study of inrush current as a statistical phenomenon uses a Monte-Carlo technique. The inrush current magnitude is calculated as a function of the random variables switching angle and remanent flux.
• “Influence of random variables on the transformer inrush current,” in International Conference on Power Systems Transients. IPST ’95 Proceedings, Lisbon, Portugal, 1995, pp. 148–52 L.Pierrat & T. Tran-Q
• The residual flux for a 545- MVA transformer is measured to be slightly higher than 0.4 p.u. (worst case out of 10 random de-energization).[/sub]


The good news is that for most of us in the application engineers field, it is not critical if the magnetized inrush current behaves with random or deterministic characteristics.

 

[davidbeach]

If it was the truly random the the simultaneous energization of multiple transformers would look very different than the energization of a single transformer. It doesn’t. But the next time that group of transformers is energized it can look entirely different. That’s where the randomness comes it; the starting point produces the randomness. But every transformer that has the same starting point produces the same inrush, for that starting point. We don’t know, or have access to, the starting point, in general. Point on wave switching can greatly reduce the variability of the inrush.

 

[dpc]

Errr... what is the time tested rule of thumb?

Primary overcurrent device must clear a point on the TCC plotted at 6X to 12X the transformer full load current for 0.1 seconds. (0il-filled) Large dry-types can be somewhat higher - 12X to 15X. I use 10X for oil-filled and rarely had issues at that value. YMMMV.

 

[prc]

" the same starting point produces the same inrush" - May not; one more factor is there.Residual magnetism (magnitude and polarity) in the core (remanence) that depends on what point on voltage wave transformer was switched off previously.Controlled switching takes care of this residual magnetism too to result minimum inrush current.

Inrush comes down fast with a decay time constant(33% of peak value) of 4-10 seconds.This time constant depends on 2L/R where L is the air core reactance of excited winding and R resistance of winding. So for larger units, it may take more time for current to decay.

 

[Mbrooke]

YMMMV- Your Might Maybe More-so Vary?

 

[jrostron]

Yes indeed estimating inrush can be very complex but there are some simple rules of thumb that will assist. One such technique is in the link below.

https://www.southernstatesllc.com/applications/connecting-solar-to-the-grid

The amount of time for the resonance of the inrush depends on the presence of similar transformers on the same circuit. In this case they can resonant for quite some time. If the circuit is damped or if these resonant condition do not exist, it become damped somewhat quickly.

All switching of transformers leaves residual magnetism or remnant flux. Switching at maximum flux exaggerates the inrush while switching at a reverse flux minimizes it. Determining the polarity of the flux is critical to determining when to close. One technique exists that eliminates the need to do this. It is the use of pre-insertion resistors. This technique requires no special sensors or controllers to implement.

 

[Mbrooke]

Question- perhaps I should start I new thread so let me know if I should but just wanted to ask a similar question in regards to what David said...

Say you have a 115-34.5kv power transformer feeding a dozen 34.5kv-12.47kv transformers... what does the inrush look like for a trip and reclose on the 115kv side? Is it more prolonged? Equal in magnitude to the sum of the main and intermediate units or something totally different?

 

[Bronzeado]

Hi Mbrooke,

I hope the paper attached may help you in this matter.

Best Regards,

Herivelto S. Bronzeado
Brasília, Brazil

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

 

[Mbrooke]

Thank you- I will start reading it!

 

[jrostron]

Multiple transformers in series are more complex to analyze as they may have multiple residual flux conditions in the various cores. the resonance conditions can last for quite some time as the paper referred to indicates.

While point on wave closing is a theoretical way of mitigation of these issues, determining the right closing is difficult because of this issue of the residual flux conditions. We had a similar condition where this arrangement produced voltage drops of 13%. Upon implementing a properly sized resistor and resistor insertion time, this was reduced to only 3%. Additionally, the series resistor damped out these inner harmonics and there was very little residual voltage drop a few cycles later.

 

[Mbrooke]

Sure- but picture a radial transmission line or an open ended substation contingency. T line trips and recloses. The question is- do just let the line reclose picking all load back up or do you open 34.5kv breakers and re-energize substations one by one?

 

[jrostron]

This in general is a normal operation for reclosing in utility systems after a fault and clearing it. The harmonics and long inrush conditions that others are talking about is one where there is no load on the system. The condition you describe seems to be rather normal and inrush is high initially but should damp out rather quickly.

Please explain a bit of the issue and that you are trying to solve. I guess my assumption was that you had some renewable resources such as solar farms on the line. In that case, they must be disconnected before re-energizing the line.