I have a PLECS model of an open-loop 3.5V - 12V boost converter. My goal is to calculate heat losses as the switching frequency increases.

Here is a summary of the problem as I understand it:

1. I could choose the highest switching frequency I can imagine, but there is a threshold where losses are too great to justify that frequency
2. I need to justify the sizing of my inductor
3. The switch I'm using has a V_ds of 12kV. I don't believe that matters, but I could be wrong

My losses don't change significantly no matter my combination of inductor size and switching frequency.

I'm certain that there's something I'm not understanding, but all available examples of designing a boost converter that I can find select the switching frequency before selecting the inductor.

Some problems I'm having are beyond my understanding, like efficiency not scaling with switching frequency when I perform a sweep using the same inductor sizing.

Here is my table of inductor sizing and switching frequency, if you fancy playing around with my model.

Inductor Size (H)	Switching Frequency
6.88657407407407e-05	20000
2.75462962962963e-05	50000
1.37731481481482e-05	100000
9.18209876543210e-06	150000
6.88657407407408e-06	200000
5.50925925925926e-06	250000
4.59104938271605e-06	300000
3.93518518518519e-06	350000
3.44328703703704e-06	400000
3.06069958847737e-06	450000
2.75462962962963e-06	500000

 

Vds of 12kV? really? Whatever it is, conduction losses will be far greater than something sensibly rated. Conduction losses will probably dominate over switching losses, making the experiment pointless.
Switching losses will depend very much on your gate drive circuit and also whether you can achieve either zero-voltage or zero-current switching. The shorter the transition time between on and off the lower the switching losses.
At a rough estimate, the power dissipated in the switching device is equal to half the voltage multiplied by half the current for the duration of switching. So it approximates to VIf/4t
where f=frequency and t=switching time.