Transformer Strange Inrush Current 2

[Tstuhl]

Hello all,

I wanted to gather thoughts on a particular trend that has been observed. The trend is as follows:

When a transformer is isolated and then energized, the longer the transformer was in isolation for, the higher the inrush current peak magnitude. Effectively, if the transformer was energized after a few days in isolation then the peak is lower than if it had been in isolation for several weeks.

I see this as a very interesting trend to try to explain. The thoughts of my team are aligned with it having to do with how a transformer can gradually lose the residual flux in its core, and somehow having less flux in the core begin connected to a higher inrush current upon energization. The literature I have come across on the subject does not support this idea though, as the inrush current peak should be smaller as the residual flux in the core decreases.

In addition to this, I found some nifty discussions on prefluxing of a transformer core before energization. This effectively requires the core to already be saturated before energization, and it has shown to significantly reduce the peak of inrush current in simulations if point-on-wave switching is used as well. I was thinking that this might harbor some kind of explanation for the trend, but I am not sure how a core would remain saturated for up to days after isolation.

Let me know what you think about how to explain this trend.

Thanks,
Tstuhl

 

[MatthewDB]

When the residual flux is high, it can help you or hurt you depending on where you energize the line. I wouldn't expect de-magnetized to be always better, I would just expect it to be more consistent.

 

[itsmoked]

It's all about whatever residual is there and point-on-wave energization. Because there is such a large possibility for alignment or anti-alignment between the residual and the point-on-wave I would think you'd need about 30 energizations to really have an honest detectable statistical result. Could it be you're just thinking you're "seeing a trend" after a couple of cycles? Often I chase a theory but find little to back it, subsequently it turns out the theory is junk which is why I'm finding nothing to back it.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[prc]

In case you want to study about the Transformer Inrush current, the best tutorial is CIGRE Technical brochure No 568 published in 2014 by WG C4.307. This 125 page document covers all about inrush current and mitigation techniques. Inrush is not depend on time of keeping under de-energization alone, but several other parameters- point on voltage wave when switching is done, position of winding energized etc, etc. The residual magnetism in a core may remain not weeks ,even months gradually decaying away. So you cannot link magnitude of inrush linearly with the time of keeping in de-energised condition.

 

[Tstuhl]

Most of what I have been researching has been related to how POW switching, residual flux in the core, and several other parameters affect the inrush current. I never did find any dependence on how long the machine might be isolated for, but due to the "trend" as well as the constant pushing from those above me I have been meticulously trying to find some kind of explanation. Since I have aggregated enough opinions from this thread and other people, I think I will conclude that it probably is something that is happening due to chance. Plus, I never did get a get an explanation of how many times this was actually observed, so it really could be a case of just having a small sample size and some bad luck.

Thank you for the feedback guys!

Thanks,
Tstuhl

 

[RodneyDixon]

Would agree with all the previous comments particularly . prc: The residual magnetism in a core may remain not weeks ,even months gradually decaying away. So you cannot link magnitude of inrush linearly with the time of keeping in de-energised condition.

Why you are seeing differences in results are primarily due to the residual flux (point on voltage waveform at which transformer was de-energised) and the point on the voltage waveform which you are energising the transformer.

If you were to carry out multiple tests you would invalidate your initial thoughts on denergisation state time being a factor.