×
INTELLIGENT WORK FORUMS
FOR ENGINEERING PROFESSIONALS

Log In

Come Join Us!

Are you an
Engineering professional?
Join Eng-Tips Forums!
  • Talk With Other Members
  • Be Notified Of Responses
    To Your Posts
  • Keyword Search
  • One-Click Access To Your
    Favorite Forums
  • Automated Signatures
    On Your Posts
  • Best Of All, It's Free!
  • Students Click Here

*Eng-Tips's functionality depends on members receiving e-mail. By joining you are opting in to receive e-mail.

Posting Guidelines

Promoting, selling, recruiting, coursework and thesis posting is forbidden.

Students Click Here

Jobs

"Zero offset" in the tranbsformer inrush current

"Zero offset" in the tranbsformer inrush current

"Zero offset" in the tranbsformer inrush current

(OP)
Hi folks,

I wonder if someone could give me a good physical explanation obout the "zero offset" that generally occur in the transformer inrush currents (please, see the waveform attached).

Best regards,

Herivelto Bronzeado  

RE: "Zero offset" in the tranbsformer inrush current

I'm not certain what you mean by "zero offset". Are you referring to the flat-topped waveform?

I'm just guessing here, bu that could be due to saturation. It would be easier to interpret the plot if the axis were labeled and more data was plotted. The input voltage for example. Also, some information on the system, transformer, and metering setup would help. Its possible that the CT could be saturating and the actual transformer input current waveform differs from what is being plotted.

RE: "Zero offset" in the tranbsformer inrush current

(OP)
PHovnanian,

YES, I am referring, in this case, to the flat-topped waveform. This "zero offset" can also be flat-bottoned.

Yes, it can be due to saturation but how can we explain that?

Remember that when  a transformer is energized the core saturates only during a part of the 60-50 Hz cycle.

My guess is that phenomenon (zero offset) may be caused by a "dynamic residual flux" (new term???) that remains in the core during the other part of the cycle in which the core is not saturated.

I called "dynamic residual flux" because it changes at every cycle. It is not the "residual flux" or "remanent flux" we are used to use. Do you agree?

Best regards,

H. Bronzeado
 

RE: "Zero offset" in the tranbsformer inrush current

The 'flat tops' that you see is the normal idling current, which is quite low in amplitude. The (in this case)negative inrush current is much larger and that makes the positive current look almost zero.

As you also can see, the DC offset is moving upwards. That can be the result of your CT's droop - it can also be the result of the inrush decaying. Usually it is a combination of both.

If you use a DC CT, you will only see the decaying inrush component.

Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

RE: "Zero offset" in the tranbsformer inrush current

(OP)
Hi Gunnar,

This phenomenon also appears when measuring the current in small tansformers (without CT).

It seems to be due to the "flux" (induction) that "leaks" the core which may leave some "residual flux" (dynamic) in that part of he cycle.

This "zero offset" is also seen in neutral current during transformer inrush, which indicates that it maybe caused by the "zero sequence flux". Does this make sense?

Regards,

H. Bronzeado  

RE: "Zero offset" in the tranbsformer inrush current

I will do some measurements on a small toroid transformer with a CT and with a shunt resistor. That will show the difference. My Saturday evening pleasure...

Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

RE: "Zero offset" in the tranbsformer inrush current

Sorry Brozo - someone just told me to do something else. I MAY be back - but no promises.  sad

Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

RE: "Zero offset" in the tranbsformer inrush current

(OP)
Thank you Gunnar, for your help!

Remember that my measurements were done in three-phase transformers.

Regards,

Herivelto
 

RE: "Zero offset" in the tranbsformer inrush current

Hello again. I was more or less forced to do a measurement on inrush. Took the opportunity to do the measurement I promised you.

I get almost identical results on this little single-phase transformer. The same flat tops

The explaation is that the inrush current is very large in comparison to the idling current. So, with normal scaling, it looks like it is zero. The impression is enhanced by the fact that magnetizing current is very small when flux is moving close to origin in the B/H diagram, which it is before flux has worked itself back to symmetrical flux. You can see the normal magnetizing current to the right in the recording.

Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

RE: "Zero offset" in the tranbsformer inrush current

(OP)
Gunnar,

Thank you very much for your effort in trying to find a reason for the flat zero offset in the transformer inrush current. It is really not so trivial.

I still thinking that this phenomenon is due to a unidirectional magnetizing of the core (each limb in a three-phase transformer) during the inrush.

Best regards,

Herivelto Bronzeado

RE: "Zero offset" in the tranbsformer inrush current

Yes. Unidirectional. No doubt. That is why it saturates in one direction and doesn't draw much magnetizing current in the other direction.  

Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

RE: "Zero offset" in the tranbsformer inrush current

I try to elaborate a little what Skogs said. When a voltage is applied to the transformer, a DC-component of current is "created", because the current of an inductance cannot change suddenly. In the pictures above, the DC-component is negative. The DC-component plus the negative half-wave saturates the core, so that the inductance decreases, and current in the negative direction increases. In the positive half-wave, the core is not saturated, so that the inductance is large, and current is small. Do you agree, Skogs?

RE: "Zero offset" in the tranbsformer inrush current

Yes. That is what I tried to say. And, in addition, the flux excursion in positive direction is very small so that the corresponding magnetizing current is also very small and looks like 0 A.

Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

RE: "Zero offset" in the tranbsformer inrush current

(OP)
Ijl and Gunnar,

Remember that the question is: why there is an "offset" in the "zero" of the inrush current (please, see the picture I posted at the begining).

Best regards,

Herivelto Bronzeado  

RE: "Zero offset" in the tranbsformer inrush current

OK. I did not understand that what you asked is why the 'zero' doesn't stay at zero. That, I find quite natural and expected. What COULD be difficult to understand is that the current stays at (or very close to) zero during the first half-periods.

First, the measurements I did were done without a CT. Instead, I used a shunt resistor so there should not be any discussion about the CT's contribution to zero offset.

Accepting that we do not at all measure with the same equipment and not with the same transformer, voltages or currents, we still get very similar results. Except that your deviation from zero during the positive half-periods is moving upwards more than mine are. I repeat what I said earlier about that:

"The 'flat tops' that you see is the normal idling current, which is quite low in amplitude. The (in this case)negative inrush current is much larger and that makes the positive current look almost zero.

As you also can see, the DC offset is moving upwards. That can be the result of your CT's droop - it can also be the result of the inrush decaying. Usually it is a combination of both.

If you use a DC CT, you will only see the decaying inrush component."

When you consider that the CT's L/R time constant is a lot shorter than that of the transformer, there is no question that the "zero offset" (using the words to describe the flat top's trend to move upwards) is a result of the CT core's flux slowly returning to steady state after the transient. A transformer can not, as you know, continue to produce a voltage with a DC component for any time much longer than its L/R time constant.

It is this adoption to the laws of physics that we see in your recording, which - if I understood you correctly - was made using a current transformer.

I do not think that there is any need to introduce new terms describing this behaviour. After all, CT's do not transform DC.

I haved zoomed heavily in the earlier recording. I also removed the voltage signal (blue) and the result shows how the positive current half-waves stay very close to zero The only change that can be seen is the positive magnetizing current increasing. The "zero offset" is not present and that is because no CT was used in this measurement.

I am sorry to be such a bore. But I needed to go into details to make this thing very clear.

Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

RE: "Zero offset" in the tranbsformer inrush current

OK, so we have answered the wrong question. The original question was "Why this non zero value occurs?", in the graph of the first message. I am not quite sure, if I understand the question correctly, even now, when reading it the second time. Can you give us some background to the question, please.  Why do you think that it is strange that the current is not zero?

 

RE: "Zero offset" in the tranbsformer inrush current

So, based on the angle of closing, we have a large dc offset in the actual current.

I think everyone agrees it is not unusual for the current to cross zero.  But if it does, we don't have any explanation for why there is a flat spot.

A proposed explanation follows.

Start with previously hypothesis that the offset is so bad and saturation so bad that the actual current on the other side of zero is miniscule in comparison, and looks like zero for all practical purposes.

But there is a deviation from that behavior and I think that's what you're asking about.  The deviation in your terminology would be a small positive offset of the predominantly negative waveform. But I would prefer to think that the negative dc offset in the displayed waveform is decaying below the negative offset in the actual waveform, which makes the whole waveform shift toward negative.

Why would there be a diference between actual and displayed? It would be some kind of hi-pass filtering effect that does not transmit dc well.  One reason could be CT error.  A CT cannot transmit pure dc (it's a transformer after all).  If you apply a step on the primary, you will see a step on the secondary which will then decay with Lm/R time constant where Lm is magnetizing inductance of the transformer.  The fact that the "positive offset" your terminology (seems to increase) would be consistent with this.  There may be other filtering effects by the instrumentation used for the capture.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Correction/clarification in bold

Quote (electricpete):

If you apply a step on the primary (Step means: i=0 before t=<0, I=Idc for t>0 0), you will see a step on the secondary which will then decay with Lm/R time constant where Lm is magnetizing inductance of the current transformer and R is the combined resistance of CT secondary winding and series-connected secondary wiring/devices.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Another  correction in bold:

Quote:

But there is a deviation from that behavior and I think that's what you're asking about.  The deviation in your terminology would be a small positive offset of the predominantly negative waveform. But I would prefer to think that the negative dc offset in the displayed waveform is decaying below the negative offset in the actual  waveform, which makes the whole waveform shift toward positive.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Here is further discussion of expected CT error which will cause dc component of output to decay below dc component of input, with time constant Lm/R

Assume:
linear magnetizing branch
purely resistive burnden.  R represents this plus the CT winding
all quantities referenced to the secondary


    I1->          Vm       I2->
=========================================
                 |                      |
                 Lm                     R
                 |                      |
=========================================

all quantities referenced to the secondary:
I1 = Im + I2 = V2/(L*s) + I2 = (R*I2)/(L*s) + I2 = I2* [1 + R/(Ls)]
I2/I1 = 1 / [1 + R/(Ls)] = s / [ s + R/L]

This transfer function has a zero at the origin and a pole at s = -R/L. For typical installation L > > > R, and the pole -R/L is very close to the origin.... perhaps at 0.001 sec^-1. The pole and zero being so close together means that their effects will cancel for all (complex) frequencies except those very close to the pole zero pair.    So we have a hi-pass filter with unity gain for all except the most slowly varying signals.

Whether or not an actual decaying dc gets component properly transmitted depends on the time constant of actual decaying dc signal compared to the  Lm/R time of the CT circuit.  

If the actual signal has time constant much faster than the CT, it will be well represented.   As the time constant of the actual signal approaches or exceeds that of the CT, an error will occur.  (displayed decaying dc component decays below actual decaying dc component).
 

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Attached is a spreadsheet that I used to simulate what I described.  It is same as the simple single-phase model spreadsheet from the other thread except I added a simulated CT driven off the main current.

I played around the parameters and I could begin to get the waveform to look like yours (see plot tab).

But to do this I needed to do 2 things that seem unrealistic:
1 - I used a ct time constant of 1 sec.  I think typical would be much higher.  If I had made it higher, the effect wouldn't show up in this initial time window.  That makes me wonder if there is some instrumentation filtering effect you're seeing other than the ct (like ac input coupling)
2 - I pushed the iron very far into saturation (even without the dc offset... if you let the simulation run to steady state the waveform is very very distorted).  I had to do this to get the positive portion of the peaks to be relatively small.  I couldn't get them flat without pushing it even further into saturation, which just doesn't seem realistic.   I am left really not having a good explanation for the perfectly flat portion of the curve.  Maybe there are some 3-phase flux interaction effects that need to be considered... just hard to juggle those in your head.

Again you can tweak the parameters and try for yourself if you have the interest. This time I have turned auto-recalculation back on so the spreadsheet should be easier to use than the last one.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Actually if I use the same solution but graph it different...zoom in a little on the time axis (as attached), it makes those positive peaks look flatter and now looks even closer to your waveform.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Something I just realized - I built the main magnetic model with a motor in mind (gapped core).    In that case the flux vs current curve would be well represented by two distinct linear regions as shown... because at low flux the reluctance is dominated by the airgap, at high flux permeability becomes constant at mu0.  There is only the transition area where it's not a straight line.   It is not good representation for a transformer primary winding at all since that has no airgap.   Whether fixing that aspect would improve the agreement (better predict the flat spot)... I'm not sure.  I'll have to think about that.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Pete. It would be nice if you could do your internal thinking before you post. As it is now, I have a problem seeing what you are trying to say.

There are a few mechanisms that produce the waveform we see:
 
1. The transient DC component that is dependent on where on the sinewave you switch on. Worst angle is 0 and 180 degrees.

2. The fact that a saturated core doesn't produce any (or very little) counter-EMF, which makes the primary current grow to values that usually are 20 - 100 times larger than idling (magnetizing) current.

3. The fact that magnetizing current is very low in the opposite direction when core is saturated in the other direction. This can easily be seen from a B/H graph for a "narrow" magnet material.

These three facts explain why the positive excursions in Bronzo's recording start at a very small value. So small that they look like zero A in comparison with the inrush current.

The next fact, and this is what makes the "offset" move in positive direction (and is also what Bronzo thinks is a new and abnormal phenomenon), is:

4. The CT has a rather low short constant and the output voltage from a CT cannot contain a DC component over any time that is long (say five times longer) compared to the time constant. So, the output signal moves so that its mean value is zero after a number of cycles. It is this 'adjustment' that we see as an increasing "zero offset" in Bronzo's original posting.

My question now is: Do you have an other explanation? I really can't say from your latest eight postings.

Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

RE: "Zero offset" in the tranbsformer inrush current

Maybe we should follow the advice from BigInch:

"The problem isn't finding the solution, its trying to get to the real question."

and Alehman:

"The engineer's first problem in any design situation is to discover what the problem really is."

The original question was
"Why this non zero value occurs?"

The simple answer is that when a voltage is applied to something that conducts electricity, the current is nonzero.

But what is the real question, the real problem?

RE: "Zero offset" in the tranbsformer inrush current

ijl - You are free to ask for further clarification.  I have already asked for clarification in the other thread.  In my mind the question is: why is there a flat spot.

Gunnar - your explanation is exactly the same as the one I have outlined between this post and the last.  I didn't realize you had said anything about CT-related decay of the dc component - that was the new part I was discussing here.

There were two reservations I had expressed - the first is gone and the 2nd remains:

1 - Earlier I expressed a reservation about whether time constant close to 1 second (needed to recreate this) was possible.  This morning I googled CT time constant which led me to 6.14 below that suggests there can be a wide range of typical time constant of CT/burden and 1 second is certainly plausible for some.
http://pm.geindustrial.com/faq/documents/general/ger-3973.pdf

2 - Another reservation I had... for the magnetic model I was using I could not get close to this waveform without setting things up such that the transformer would be very far into saturation even after the actual dc component decayed (not realistic).  I am not sure whether this is a result an incorrect magnetic model (gapped core) or whether there is some 3-phase transformer effect (not modeled) that helps explain this aspect.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Clarification in bold:

Quote (electricpete):


In my mind the question is: why is there a nonzero flat spot

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Quote (Bronzeado):

This phenomenon also appears when measuring the current in small tansformers (without CT).
But I assume you used a clamp-on. Clamp-ons see this same behavior.

There was a lot of discussion in this thread:
thread238-217182: ability of clamp-on probe to see exponentially-decaying dc current

I tried to summarize some aspects in my attachment in that thread 19 Oct 08 15:12, which I have also attached here for convenience.

The relevant figure for this discussion is Figure 2 (on page 5).  It shows a response of a smaller AEMC clamp-on probe to a signal that is known to be a step-increase in dc (no ac component).   The response was step increase followed by exponential decay with time constant around 0.06 seconds.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Gunnar - my apologies for repeating things you had already said.  I think you gave a very good answer.

Also ijl I agree the wording of question is very strange.  I (we all including you) just try to explain the shape of the posted curve until it makes sense to our satisfaction.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Yes Pete. This thread has been winding in many directions. I think we brought it on track finally.

Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

RE: "Zero offset" in the tranbsformer inrush current

If anyone is interested, I was able to finally recreate the posted waveform pretty closely with reasonable values of saturation (knee of the curve at 70% of applied voltage).  The graph is shown in tab plotsheet, attached.

To get it to work, I just need to decrease the "sharpness" parameter of the magnetic circuit (the high sharpness associated with gapped core was throwing it off).

I conclude that Gunnar's explanation (which it took me longer to get to) was very much on-target.   It is the exact same thing I modeled... basic model shown in tab ProblemDescription, transformer magnetic characteristics in tab MagChar, and other input parameters in tab "Main".  (input parameters are now listed on the top of the graph also).  There was no need for modeling any three-phase behavior to recreate this waveform.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Me too, see the attached picture, simulated with ATP. I think that the question is, are the tops really flat in the original picture (of the first message) or is the resolution of the picture too coarse. The attached picture just confirms what has been said earliler. That is, current is large, when saturated, else small. But the small current is sinusoidal, not flat.

RE: "Zero offset" in the tranbsformer inrush current

Isn't it nice to have reality confirmed by two simulations!? smile

Thanks to both of you.

Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

RE: "Zero offset" in the tranbsformer inrush current

Reality is in the eye of the beholder.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

I think the important aspects were well covered by Gunnar early.

We said the flat appearance is result of scale and that is certainly true and probably the most relevant explanation for the flatness.  But there is a smaller effect that contributes to the flatness that I also wanted to mention - the low-excitation magnetic characteristics of the iron (that's something I stumbled onto when I started out with wrong model that did not properly reflect the low excitaiton characteristics).  It contributes to a waveform that is flat-topped whenever the part is excited in the very low part of the B-H curve.

And when we look at the current half-cycle that has opposite polarity of the dc off-set, it certainly is operating in the low-excitation portion of the curve so should exhibit this slightly flat-topped behavior.

The terminology used in the spreadsheet was Leffective which is just a parameter that is proportional to differential permeability (differential permeability is the instantaneous slope of the B-H curve).  Electrical steel has low differential permeability at low excitation, gradually increasing to peak differential permeability near the knee, then decreasing again for higher excitation.

Attached I have modeled simple sinusoidal excitation (no offset) in the very low excitation portion of the curve where differential permeability (Leffective) increases with current.  The results shown in the plotsheet tab is a curve which is not sinusoidal, but instead relatively flat accross the top and very steep (compared to sinusoidal current of same magnitude) at the zero crossings.

An interesting note - i think most people are familiar with the non-linear effects at high excitation but not as many familiar with behavior at low excitation.  It has probably become intuitively obvious for most that at high excitation we get a spike near the peak of the exciting current waveform.  In a simplistic way, we can view this as the opposite... flat spot near the peak of the exciting current waveform at low excitation.
 

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

I should mention I have neglected hysteresis, which becomes more important in the low-excitation portion of the curve.  It is by no means an exact simulation, but illustrates a point that I believe is correct:  The permeability is not constant at low excitation and we expect some distortion of excitation current, and that tends to make a flatter top.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

(OP)
Hi folks,

I was away from the internet last weekend so I could not participate in the discussions.

I will prepare something to you all in order to clarify my original question and give you more insight on this phenomenon.

I will be back soon.

Regards,

H. Bronzeado    

RE: "Zero offset" in the tranbsformer inrush current

(OP)
Dear Gunnar, ijl and electricpete,

Please, find attached the simulation results of three-phase transformer inrush in which we can see the phenomenon of "zero offset" in the inrush currents.

We can see, also, that this phenomenon seems to follow the "dc decay" of the neutral current (Fig. 9e), which in turn follows the "dc decay" of the magnetic flux outside the core (9i). In my opinion this flux ("zero sequence flux") is the responsible for the core unidirectional (dc) magnetizing, causing the "zero offset" in the inrush current. Does this make sense for you?

Best regards,

Herivelto Bronzerado

RE: "Zero offset" in the tranbsformer inrush current

You are right - it looks like what is attached is result of a model simulation, so the plotted currents likely don't reflect the error in the dc component that we talked about.   

So, the explanation that we proposed is not the only one... there can be other things at work such as whatever is at work in this model (I didn't study it).

No doubt the exact shape can vary with different configurations of primary winding (delta, ug-wye, grounded wye) and core (3 leg or 5 leg) among other things.

Thanks for posting that.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

RE: "Zero offset" in the tranbsformer inrush current

Bronzeado,can you put up the complete paper you referred to or atleast title,author etc.

RE: "Zero offset" in the tranbsformer inrush current

prc

Lower left says IEE Poc.-Sci. Meas. Technol., Vol. 141, No. 6, November 1994

That is probably where you can find the complete paper.

Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

RE: "Zero offset" in the tranbsformer inrush current

I agree also that this must be neutral current because (holding a straightedge vertically on the graph), it looks like all 3 phases experience a positive-direction flat-spot at the exact same time.  

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

Red Flag This Post

Please let us know here why this post is inappropriate. Reasons such as off-topic, duplicates, flames, illegal, vulgar, or students posting their homework.

Red Flag Submitted

Thank you for helping keep Eng-Tips Forums free from inappropriate posts.
The Eng-Tips staff will check this out and take appropriate action.

Reply To This Thread

Posting in the Eng-Tips forums is a member-only feature.

Click Here to join Eng-Tips and talk with other members!


Resources