Average modelling of (MOSFET) Half-bridge

, ,
1

Hi,

we want to model a CLLC resonant converter, using average modelling on the high and low voltage side of the full bridges. I saw the “Half-Bridge” block in Plecs with “Sub-cycle average” configuration, but I am not sure if this is the right approach for modelling my full bridges using average modelling. We want to speed up the execution time of our simulation by NOT calculating every switching event, instead using the average output voltages of every half bridge.

Can you put me in the right direction please, if I can use this Half-bridge blocks or do I have to model the full bridges using controlled voltage and current sources?

Thank you for your help and with best wishes,

Lars L.

2

Hi Lars,

I may not appreciate the full context of your question, but I see two points embedded in your query:

  1. How does one create an average model of a half bridge in PLECS?

You can use a half-bridge block from the PLECS library in sub-cycle average configuration. You can also create your own model based on controlled voltage and current sources based on the literature. The sub-cycle average models have integrated diodes that can model things like natural commutation better than conventional “DC transformer” based models. See the following paper for more details.

  1. how does one create an average model of a CLLC.

I would argue that averaging the half-bridges is not an appropriate technique to derive an average model for a CLLC, as the behavior of interest is the resonant behavior within a switching period not observable using average models for the half-bridges.

A common technique is First Harmonic Analysis (FHA), although other methods are still under active research, as FHA has known limitations. To create a basic FHA model, one would create a controlled voltage source whose input is a variable-frequency sinewave representing the first harmonic of the switching waveform.

In short, I would suggest a bit of research on an appropriate average modeling technique, and to reach out if you have questions on how to implement such a model.

Regards,

Bryan