How to obtain a bode plot from a Power Electronics Circuit in Simulink

Dear Sir,

I am simulating a power electronics circuit. It is a Buck converter which is controlled with Peak Current technique. I want to obtain the bode plots of some variables. My circuit contains transistor which are switching element. So I try to follow this: https://es.mathworks.com/help/slcontrol/ug/visualize-bode-response-of-simulink-model-during-simulation.html but I cannot obtain the bode graph of a group of elements. It only gives me the bode plot of each component independently. Thank you very much for your time.

[Álvaro Morales Muñoz]

I discover how to do that with https://es.mathworks.com/videos/frequency-response-estimation-68884.html but I am still having problems:

The bode plot become weird after a half of the switching frequency. That do not make sense, because I am not using any digital system. I attach an example: I am simulating a buck converter, with 100 kHz of switching frequency. I only show the part where the weird response starts. Figure 1 shows the expected Bode graph (Vout/duty):

And the bode plot obtained from linearization the system using simulink (Vout/duty). Figure 2:

The weird behaviour is expected at 628000 rad/s, not at 314000 rad/s.

For the last image, I use "Sinestream":

Could anyone tell me what is happening? Again, thank you very much for your time.

Answer 1

[Siddharth Jawahar]

Hi Alvaro,

You are on the right track of using frequency response estimation for such swithing systems. However, the experiment settings for needs to be carefully set up to get good results. For example, we would recommend a peturbation amplitude of 5-10% of the nominal value. 0.001 seems too low of peturbation amplitude for duty cycle and resultant plant response would be indistinguishable from noise in the system. I would recommend looking at this video to follow some best practices while estimating frequency responses:

Best,

Sid