Can be a good help the next article:

https://www.fairchildsemi.com/application-notes/AN/AN-7520.pdf

Dear Prof. Bogdan

Thank you for your support. I have already read that article, Fairchild models

"are based on equations developed by curve fitting laboratory test data".  My question was related to calculating the power loss using only the data available on the datasheet.

Thank you in advance

Well, for the IGBT the total loss in one switching cycle is is the sum of the energy Eon (switch on) + Ef (in forward state) + Eoff (switch off). Ef can be calculated from the current and the resistance of the IGBT in on-state. Good datasheets will have curves for Eon and Eoff, but the one you have linked to does not. A "back of the envelope" way to estimate the Eon and Eoff losses is to assume a linear increase in current and a linear decrease in voltage during the ton and toff switching times. The Eon and Eoff power losses are then simply the integral of the current*voltage curve over the time, ton and toff respectively. The switching time and the actual shape of the current and voltage curves do however depend on the actual setup, load, parasitic inductance/capacitance and so on, so you should always measure when you have your setup in the end. Nevertheless, the estimates should be enough for you to design for example a cooling solution, just be sure to leave some headroom. As you will see in the datasheet and in the application notes the listed times are not always the times where the actual switching happens. Manufactures can use different parameters for the listed times, some may be delay times, i.e. the time before the IGBT starts to conduct. So refer to the switching characteristics in the datasheet to get good estimates for ton and toff, don't just pick the parameter with the closest name.

I have attached an image showing the power losses in an IGBT, and an application note from ABB that explains this more in detail.

For the diodes I have attached an application note from IXYS.

Hope this is of some help :-)

Dear Kristian

Thank you very much for your answer.

Dear kristian,

Thank you so much for your explanation that helped me a lot.