Hello,
- What is the recommended method for calculating inverter input power, output power, total semiconductor losses, and efficiency in PLECS?
For the input power, use a Probe block to probe the Source power signal of the DC input source. Calculating the active and reactive output power of a three-phase system requires some additional signal processing. You can find examples in our power-distribution demo models, such as the Cascaded H-Bridge Converter.
The semiconductor conduction and switching losses are best calculated using the Switch Loss Calculator block. Add all MOSFETs and diodes belonging to the inverter and select Total loss. If you would like to inspect the instantaneous switching and conduction losses, use a standard Probe block and drag the semiconductor device of interest onto it.
When using losses from thermal descriptions, calculate the efficiency as
\eta
=
\frac{P_\mathrm{in}-P_\mathrm{loss}}{P_\mathrm{in}}
=
1-\frac{P_\mathrm{loss}}{P_\mathrm{in}}.
You should not calculate the efficiency as P_\mathrm{out}/P_\mathrm{in} in this case. Losses calculated from the semiconductor thermal descriptions are evaluated in parallel with the electrical circuit and are not drawn from the electrical domain. Therefore, an efficiency calculated only from the electrical input and output powers does not include these thermal semiconductor losses.
The Buck Converter with Thermal Model demonstrates how to use the Switch Loss Calculator and calculate efficiency from the input power and thermal losses.
- Should the input/output power and losses be averaged over a specific number of fundamental cycles to obtain accurate efficiency results?
Yes. Use Periodic Average blocks.
For an AC system, averaging over one fundamental period is normally appropriate. In your case, with a fundamental frequency of 50 Hz, the averaging time is 20 ms. You can use the same averaging time in the Switch Loss Calculator.
- Are there any particular settings required in the thermal models or loss calculation blocks when performing a load sweep?
No special settings are required for the load sweep.
However, you should verify that the simulated currents, blocking voltages, and junction temperatures remain within the range covered by the lookup tables in the thermal descriptions. Outside these ranges, PLECS extrapolates the loss data.
If you also want to include ohmic losses in the thermal model, you can extend the Heat Sink component so that it encloses the corresponding resistive components. The Heat Sink absorbs the losses dissipated by the components within its boundaries. For a resistor, these losses are calculated as i^2R, or equivalently u^2/R.
Use a Heat Flow Meter connected to the Heat Sink to measure the total losses dissipated by all components within its boundaries.
The general relationship between electrical components, thermal losses, and the Heat Sink component is described in the manual under Thermal Concepts.
- Is a continuous decrease in efficiency with increasing output power expected under certain operating conditions, or could this indicate an issue with the model or measurement method?
In your model, this behavior is expected.
I simulated your model and connected a Thermometer to the heat sink. After 1 second of simulation time, the heat-sink temperature exceeds 240 °C. Note that semiconductor losses are temperature-dependent and generally increase as the temperature rises.
You can inspect the thermal behavior in more detail using a Probe block. Drag an individual MOSFET or diode onto the Probe block and enable the following signals:
Conduction lossSwitching lossJunction temperature
This allows you to identify the contribution of each semiconductor and compare their individual junction temperatures. Note that the directly probed switching-loss signal consists of energy impulses in joules. Use the Switch Loss Calculator to obtain the corresponding average switching-loss power in watts.