Please take a minute to review and accept our Terms of Use.
Welcome to the PLECS User Forum, where you can ask questions and receive answers from other members of the community.

Many technical questions regarding PLECS are answered on the Technical Solutions page of our website. Tutorial videos, specific application examples, and pre-recorded webinars are available on our YouTube page. Please follow us on LinkedIn for the latest Plexim news.

3 phase transformer model in magnetic domin

0 votes
176 views
Hello,

I have tried to build a 3 phase transformer in PLECS magnetic domin. The trasformer model works so far well, and I would like to look the magnetic character more in detail.

The flux rate of 3 legs in magnetic domain are symmetrical, but after the integral, the flux are no longer symmetrical. (It contains a DC portion because of the constant C I think, as shown in the attachment). Since then, the flux density we got for 3 legs is quite different and it also depends on the specified phase of the voltage source. If the saturation or hysteresic effect of the core material is also considered, it makes a huge difference for the flux density and magnetic loss for 3 legs  (as shown in the attachment).

My question is, how can I get this DC portion in flux and then in flux density eliminated?

Thanks in advance!
Sara
asked Nov 22, 2020 by Sara.Z (14 points)
Are you able to post your model here?
Hi, attached is a model with linear core.
Sorry, I just took a look at your model. You should consider this like a set of series inductors connected to an ideal three-phase voltage source- without any dampening impedance you will actually expect imbalanced currents unless you provide initial current values to force the balancing. The same is true with your magnetic circuit, where you will need to include initial MMF values for the core elements in each leg to ensure the DC flux portion is eliminated. I hope this helps!

1 Answer

0 votes
Thank you for sending the new model. I see that you modified the PLECS Hysteretic Core library component to include an initial MMF. That is not necessary in this case and should not be done as it is very tricky to get the balanced values correct. In the attached model I use the standard block and while the flux imbalance is clear if you then run the model as is, a few adjustments should fix the issue. First, the solver tightness (relative tolerance) should be much smaller due to the small time constants in the system- I changed it from the default value of 1e-3 to 1e-7. This ensures better accuracy and error handling and makes sure the state trajectories are on the right path from the beginning of the simulation. Then you will see that the dampening you need to diminish the initial imbalance is occurring, but just very slowly. I adjusted the simulation time to 5 seconds to see the behavior over a longer period of time and increased the R1 winding resistance value to 5 ohms. The flux waveforms do then just about settle by the end of the simulation, but of course you can experiment with different time constants as well. I hope this helps, but let me know if you have further questions.
answered Jan 21 by Kris Eberle (1,240 points)

Hey Kris, thanks again for your helpful answer. 

Recently, I am trying to build a model of a real 3 phase isolating transformer in magnetic domain, and both eddy current losses and hysteresis losses are considered. (related to How to calculate the magnetic resistance for the transformer simulation in the Plecs Magnetic Domain? - PLECS User Forum (plexim.com)) Whereby considering both the speed and accuracy, I have only divided the entire transformer into 5 layers.

The simulation runs so far well, but now the biggest problem is, it runs very slowly. Maybe you have some ideas to bring it faster or maybe a multicore computer could help? Thanks in advance. Could it be possible that I contact you via email for more detailed information?

Hi,

You are welcome to send your inquiry to our general Support email (support@plexim.com) and then I or someone else here can answer it as soon as possible. If you do so, please provide your model file.

Thanks!

Kris
...