Transformer Inrush Question from an ME 1

[fiberstress]

Hello,

I've read through several articles on transformer inrush, but I'd like to poke around the great EE minds here to see if I can get a physical explanation of why transformer inrush is only one direction (or polarity), (see the example waveform we captured in the link).

Much appreciated

 

[deltawhy]

I believe what you are looking at is a DC offset. It is thus not necessarily in "one direction" but the sine wave is "raised up" by exponentially decaying DC.

 

[dpc]

If you are asking why the inrush current is asymmetrical wrt to the 0 axis, it is called the dc offset as indicated by deltawhy. This is a function of the voltage phase angle at the instant that the transformer is energized. If the voltage is at (or near) maximum, there will be no offset. If the voltage is near 0, the offset will be maximum. This is due to the inductance in the transformer core and the fact that the inrush current wants to lag the voltage. The remnant flux in the core also can have an impact.

So the offset is not always in one directon and sometimes does not even occur. In a three-phase transformer, each phase will exhibit a different waveform since the three voltage phase angle will be different.

 

[fiberstress]

Forgot to add some background on this; it was recorded at the building main during closure of a disconnect to a critical equipment with a DC generator and motors, field controllers, and misc. loads but nothing was energized during this closure except maybe some lighting. We'd hypothesized maybe this was diodes or thyristors in the speed controllers of the dc motors, but it looked more like other example waveforms of transformer inrush.

 

[jraef]

No it's not really related to the load, it has to do with the magnetizing current for the transformer itself.

"Dear future generations: Please accept our apologies. We were rolling drunk on petroleum."
— Kilgore Trout (via Kurt Vonnegut)
For the best use of Eng-Tips, please click here -> faq731-376

 

[fiberstress]

My question is similar to the thread linked. In all the papers/articles I'm reading, nothing is truly giving an explanation of the physical phenomenon of DC offset. I understand the concept of the voltage phase angle at closing and it's effects on the lagging current/flux, but something scientists and engineers get out of touch with is that math is not the explanation, it's just man's best attempt at describing the phenomenon.
I know electricity is especially hard to describe physically, but can you guys give a good physical explanation? Does it have anything to do with response time of the current/flux (like the high dv/dt poking the flux in the rear with a hot iron telling it to get a move on)? My big question is that my waveform doesn't look like the dc offset stuff we saw in electronics 101 where the entire waveform shifts up or down. It looks like the positive peaks are amplified and the negatives aren't affected, so if this is DC offset during transformer inrush, how is this happening?

 

[cranky108]

Maybe a little different way of looking at this. We know there are reactive components in the power system circuit, and that those reactive components store and discharge energy with the power grid. That being said, that energy, at any instant in time is a DC value, but because that value is changing in relation with the power grid, it is forced to become a ac reactive component.

On inrush the current into unmagnitized wire that is the transformer, will reduce the voltage, which unbalances the equil trading of energy from inductive to capacitive, leaving a deficet which is fed into the new part of the circuit. The amount of inrush will depend on where in the power cycle you are at the closing of the switch. If you are near a zero energy stored point, you should experence zero Dc offset. The inrush will however still take place, just not the offset.

If there are any corrections, or additions, please make them. As I am not a teacher, I may not have said something correctly.

 

[fiberstress]

Cranky108, you're hitting me where I itch, thanks. So if that stored energy is positive, you'll get a graph like I att'd, and if it's negative, you'll get an inverse (we get this on the other phase that's near 0v but opposite dv/dt). Why would the negative current then be seemingly unaffected?

 

[fiberstress]

Here's a zoomed in picture of the waveform. Why does it seem like there's a threshold in the level of positive current after which it opens the gates and it surges?

 

[mikekilroy]

how did you measure this?
[ul]current transformer?
resistor in series with lead?
is it one phase of a 3 ph transformer?
primary or secondary?
can you include a voltage too on channel 2 of scope?
was scope on AC or DC input coupling?
what happened before the steady state 100App on left?
can you get pix starting with 0 amps?
[/ul]

Just some more data that may help understand....

 

[mikekilroy]

a sampling of google search of "transformer inrush scope pictures"

http://www.electrical-installation.org/enwiki/Transformer-energizing_inrush_current

https://www.google.com/search?q=tra...2gXXlYHwBQ&sqi=2&ved=0CDgQsAQ&biw=970&bih=576

http://relays.te.com/appnotes/app_pdfs/13c3206.pdf

 

[mikekilroy]

the 3rd reference above (relays.te...) totally explains the physics of what you see.

 

[fiberstress]

Here's the voltage waveform

 

[fiberstress]

And here is a zoom on the peaks

 

[fiberstress]

Mike, this was measured using an All Test Pro Online II analyzer with Summit Technologies HA1000a (0-1000Arms) CT's. Can't seem to find the ratio on any cut sheets.
It was taken at the building main, and the disconnect switch is fed through a main switchboard and it's branch circuit breaker.
I can't answer on the AC or DC input coupling, the analyzer is a blackbox, and unfortunately it only captured triggered events so I don't have any cycles before or after these.

 

[fiberstress]

...and I don't know which transformer it is, but likely a single phase 3kva for the operator cab lighting and an A/C unit, but could also be a 3 phase for the unit's controls (it's a large crane). So it seems like my opened gate on the up slope in my current waveform is the transformer saturating.

Does the reason for the inverse peaks of the current not being affected much have to due with the back emf returning and creating more normal impedance?

 

[jghrist]

Adding to what cranky108 said:
There is also a residual flux in one direction depending on the point on the voltage wave where the transformer was deenergized. This residual flux will add to the current if the flux is in that direction when it is reenergized.

 

[PHovnanian]

Does the reason for the inverse peaks of the current not being affected much have to due with the back emf returning and creating more normal impedance?

That would be due to saturation of the transformer core by the instantaneous value of the inrush current. Since this value is equal to the sum of the sinusoidal and the DC offset (the exponentially decaying component), it is higher (absolute value) on one side of the waveform than on the other. So the shape of the waveform is affected asymmetrically.

Its sort of a 'back emf' issue in that the magnetizing current creates a core flux which in turn produces this back emf. But once the core reaches saturation on each peak, very little additional flux and back emf is produced, resulting in that sharper spike on the top of the waveform.