How to set residual/remanent flux/magnetism? 4

[krunom]

Hi, I am carrying out some switching measurements on a one-phase transformer and it would be very important that I could set the same residual flux before each measurement.

For instance, could I use DC-source so that through the winding flows DC-current that would set residual flux to some value? Furthermore, if I decrease the DC-current to zero, would the residual flux decrease on zero also?

Namely, it is not clear to me does the flux (after aplying DC-current) depends on residual flux which existed before the applying?

P.S. I sent this question on Magnetic forum as well, however this question is also "electrical" and this forum is more visited and i am really desperate, so please dont be angry

 

[Skogsgurra]

Decreasing DC to zero does not reduce residual flux to zero. You need to "circle" the B/H diagram fom maximum flux down to zero.

A very efficient way of doing it is to put a capacitor across primare (or secondary) winding when switching AC off.

Chose the capacitor so that you get a decent frequency and at least around 20 periods before oscillations decay down to noise.

Make the frequency a few hundred Hz. May also work with thousand, but eddy currents will reduce the defluxing action if frequency is too high.



Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[krunom]

Thanks for the answer! But how about setting some fixed value of flux by applying DC current to the winding, i.e. it is not clear to me does the flux (after applying DC-current) depends on residual flux which existed before the applying of DC-current?

 

[itsmoked]

If you applied a DC current of a moderate percentage of the transformer's nominal current, you would get the same residual magnetism each time - regardless of what the previous residual was.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[krunom]

Thanks itsmoked for the clear answer

However, Skogsgurra said that if I would decrease applied current to zero the residual would not decrease to zero and however I feel that this is in a some kind of contradiction with your answer.

Namely, could please anybody explain how is it possible that with applied DC current I would get the same residual magnetism each time (regardless of what the previous residual was), but if I decrease the DC current that the (residual) flux would not decrease... :-| Or maybe it would decrease each time on some non-zero value?

 

[itsmoked]

Well... It's all a bit confusing because you are really talking about a pile of iron made up of millions of little magnetic domains. Each domain is what ends up pointing in some direction. The vector sum of them all ends up being the resultant residual magnetism of your transformer.

Skogsgurra was passing on a typical method of how one can randomize the millions of domains effectively.

If you want no resulting detectable field you need to apply a decaying sinusoid to the core. The result is that the various layers of domains from the outer regions on down into the inner regions all get pointed in different directions as the sinusoidal B field decays. The end result will be each layer pointing in varying directions for a sum of zero. (generally)

However, this is kind of hit and miss. One time it will be zero. The next time maybe not quite zero. This is because when you setup his clever scheme of an LCR circuit that will ring and thereby provide a nice decaying B field, just how well it works, would depend on when you interrupted the AC power.

If instead you just apply a DC current you will never get zero residual but you will get all the domains lined up and consistent. However you must make this DC B field strong enough to effect the entire bulk core or you will have the outer reaches still left in whatever orientation the last large AC B field left them in.

Watch a movie of a degaussing procedure for a CRT television. It is this same method in a planar situation using a large electromagnet and the human hand providing the decaying alternating B field by slowly removing the electromagnet in a spiral manner.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Bronzeado]

Krunom,

There exist a method to set a residual flux by using a DC current. I used this method 25 years agi to measure the saturation curve of a single-phase tranformer (66.6MVA, 230/26/26kV). Unfortunately, I don´t know where is my old notes.

Howerver, depending on the size of the transformer, the easiest way to set upp a value of residual flux in a transformer is to apply a decaying sinusoid voltage on the winding that wound the core, as suggested itsmoked and, in another way, Skogsgurra.

If you are "playing" with a small transformer in a lab, just use a "variac" and apply a voltage from zero, go up and return to zero. In this case the residual flux will be zero.

Best regards,

Herivelto Bronzeado

 

[krunom]

Thanks for all advices and I understood the method with decaying AC excitation. However, I would like to know one more detail about method with applied DC-current.

Namely, if I apply some current of a moderate percentage of the transformer's nominal current I would direct the domains in one direction. So far, so good

However, after that should I remove the DC-excitation (current) abruptly? If yes, I would have a great problem with induced voltage according to Faraday's law. If I use a diode connected anti-parallel to the winding, I would not have a problem with induced voltage, but the DC-current would not be remove abruptly...

So, should I remove the DC-excitation (current) abruptly?

 

[itsmoked]

I shan't think it would make a great deal of difference. Do whatever works for you electrically. Once the domains are DC aligned reducing the current slowly or rapidly would not un-align them. If you end up with an arc then likely you are going to get some ringing.. Which may result in varying amounts of un-alignment.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

Look at a hysteresis diagram of and transformer steel. Applying suitable DC current will always leave you on the same point of the hysteresis curve. AC will put you in the center, off the locus of the curve. Take your pick of the method to use.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Skogsgurra]

krunom

You shall not remove the DC at all if you want a specific DC induction level.

When you remove the DC, you "slide back" to the coercivity flux that corresponds to the DC induction you just had. It is only if you know what the coercivity flux is (or if you can measure it) that you can work without DC bias.

If you want to measure any property of the core with a well defined DC flux in the core, you need to use a DC current source with a high internal impedance. Using a large inductor in series with the DC source is probably the most practical way to have that bias.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[prc]

krunom,can you explain the test you are planning and the need for a fixed residual flux?
Use a diode connected anti-parallel to winding -what is meant anti-parallel?I thought you are planning to connect across the winding.

Skogsgurra, I believe by removing the DC current, the residual flux will not vanish immediately.It will take quite long time for that to happen.

In power transformers, we have to handle residual flux.One is during switching surge test where we apply surge of approx 200/500 microsec duration.If we apply 100 % magnitude wave, then we apply wave of approx 60 % magnitude for demagnetisation before applying next 100 % wave.

Another occasion is to check magnetisation current at low voltages.(230 V AC) If there is residual flux in core, ( that is the case after a fault tripping or winding resistance measurement with Dc current) excitation current values will be high and misleading.Core is demagnetised by applying a DC current ( root2 times the rms magnetisation current of transformer)for some time and then reverse the polarity of Dc current.Then again DC current of say 80 % is applied and the procedure is repeated. Then apply 60 %.40 % ETC. This used to work.

 

[itsmoked]

Everything you state prc rings true. You are correct the magnetism will not 'go away'. It will remain, effectively, forever.

And even if the field drops back to coercivity flux that is where it will stay. That is why you want to put a large part of the transformer's current rating thru it, to get everything up to and past the coercivity point.

It is interesting to hear that you guys preset the cores on your transformers..

Keith Cress
kcress -

http://www.flaminsystems.com

 

[krunom]

Thanks prc for the reply.

I am doing some measurements on a laboratory set-up, i.e. I am switching on the network that comprise the transformer. It would be very important that I would be sure that the residual flux is the same at the beginning of each measurement, i.e. before each switching on.

Therefore, between measurements I should carry out some procedure to magnetize the core on same level. This level could be zero, but it is not necessary.

So, the most important thing is that the residual flux is the same before each measurement.

 

[Skogsgurra]

Then you should use zero. Anything else will only complicete things and make the results less useful.

A comment for prc: Yes, the flux goes back to the residual flux corresponding to the DC flux that was in the core just before switching current off. Have a look at a typical set of hysterisis loops (there are such sets in almost any text-book on electricity) and see how the flux "slides back" towards residual flux (H=0). Depending on what the actual level was when switching off, you land on different remanent flux values. Also, it doesn't take long time at all. Do not confuse with a permanent magnet losing "power" over time. Not the same thing.

Also, the switching from +100% to -100%, followed by 80% and -80% is not used very often and I am not so sure that it works. The whole idea is to gradually reduce the flux so that ever diminishing loops are followed. That can be done either by using a decaying oscillating current as described in my first post. Or swiching +100%, -90%, +80%, -70% or a similar combination of lower and lower currents with alternating polarities. That technique has proven itself for decades in magnetic filters for feed water in boilers. One filter, that Siemens used to deliver, did that in .8% decrements (using an 8 bit D/A).

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[krunom]

Thanks to all of you for your answers. In the meantime I managed to talk about this problem with my ex-faculty-professor. He is "an expert" regarding the transformers and he told me that if I want to have a zero residual flux I should use decreasing AC excitation. He said that by using decreasing DC excitation I wouldnt become zero residual flux each time because I dont know what was the initial residual flux, i.e. in this case the initial residual flux has a significant impact.

 

[itsmoked]

That should not be the case if the DC flux is great enough for the first step. But the AC would still be the best bet.

Keith Cress
kcress -

http://www.flaminsystems.com