Videos > Designing a 3-Phase Inverter in Ansys Simplorer and Coupling with Maxwell FEA for a 160kW PMSM
Aug 1, 2025

Designing a 3-Phase Inverter in Ansys Simplorer and Coupling with Maxwell FEA for a 160kW PMSM

Hello everyone, this is Batan from Ozen Engineering, Inc. In this video, I will show you how to couple your electric motor into the Ansys Simplorer to generate your inverter fatigue excitation.

Introduction

Before diving into the details, I have created a blog for step-by-step guidance on this program. You can follow along to check all the steps I used to create this model. In this video, I will provide a high-level overview of what I have done, how it works, and how you can utilize it in the Ansys environment.

Creating a 3-Phase Inverter

To create a 3-phase inverter, we have two options:

  • From the grid
  • From a battery, as in electric vehicles

In this circuit, we have a voltage source to replicate the battery in the vehicle or any project, and switches to represent the IGBTs or MOSFETs. We then have three-phase connections to our electric motor.

Simulation Setup

Here, I use resistance and inductance as extra parts of the electric motor since my Maxwell model uses stranded windings. When using stranded windings, Maxwell cannot calculate the resistance and inductance of your windings. Therefore, we add external resistance and inductance to the circuit:

  • Phase resistance: 50mΩ
  • Phase inductance: 5μH

To get feedback from the phase currents, I assign ammeters for each phase. These phases are connected directly to the ABC to DQ conversion. We need to know the rotational position of the electric motor while it is spinning, so we assign a rotational angular velocity source to provide rotation to the motor.

Measuring Torque, Velocity, and Rotational Angle

To measure these parameters, assign the necessary blocks and make the connections as shown. After obtaining the angular position of the electric motor, convert the mechanical angle to electrical for use in the controller. This involves:

  1. Summing the first angle with 7.5π/180 to correct the rotor position shift.
  2. Multiplying the value by 4, the number of pole pairs of the electric motor.
  3. Adding π to the electrical angle for ABC to DQ conversion.

PI Controllers and Conversion

In this model, I use PI controllers for each ID and IQ current control. You can also create a controller to set the reference input signal based on torque or speed conversion. For example, for 3000 RPM modeling, I use:

  • ID: -7.37
  • IQ: 38.2

Optimize these values based on your electric motor characteristics. After setting up the PI controllers, couple them into the DQ and ABZ conversion to convert signals into SVPWM. Here, we use:

  • Switching frequency: 1000 Hz
  • Dead time: 1E-06
  • VDC as VBUS

Integrating with Ansys Maxwell

To achieve the FEA model in Ansys Maxwell, copy your electric motor into the same project as Simplorer. In Maxwell, right-click and choose design settings. Add the following settings:

  • Enable transient to transient link with Twin Builder

This provides the electric motor model to Ansys Simplorer. To add this transient model, click TV Builder, add component, Maxwell component, and choose the transient code simulation setting. Ensure the link type is Maxwell transient to transient and select your simulation for the electric motor.

Connecting the Electric Motor

Once the component is on your workspace, drop it anywhere in your circuit and make the connections as shown. I connected the electric motor as star connections and coupled the phases directly with my circuits. This completes the circuit for the battery FAT inverter.

Grid FAT Inverter

For the grid FAT inverter, modify the circuit by adding 3-phase AC sources instead of a single DC source at the entrance of the 3-phase conversion. Assign the RMS value as a parametric value (grid_v), change the frequency to 60Hz, and set the phases to 250, 120, and 0 for each phase.

After adding AC sources, couple them into a 6-pulse rectifier. Add a filter capacitor (1μF) to the end of the circuit. Use a voltmeter (Vm1) to name the value. Change the Vbus value to Vm.v for the SVPWM block to generate signals for the switches.

Running the Simulation

Complete all settings and run your simulation by clicking the analyze button. It may take 5 to 20 minutes based on the model's complexity. After completing the simulation, return to your Maxwell model to generate any desired results, such as torque values or loss calculations.

Conclusion

This system-level coupling can be used for various applications, such as thermal simulations of your electric motor and inverter. I will create a blog to show how to generate a Q-equivalent circuit model for faster simulation results. I highly recommend visiting the blog for more details and steps.

Thank you for watching this video. Please contact us at Ozen Engineering, Inc. for more information.

[This was auto-generated. There may be mispellings.]

Hello everyone, this is Batan from Ozen Engineering. In this video, I will show you how you can couple your electric motor into the Ansys Simplorer to generate your inverter fatigue excitation.

Before going into the details of this process, I would like to show you that I created a blog for step-by-step guidance of this program. Here you can go through and check all the steps I created this model.

In this video, I will give a high-level presentation of what I have done and how it is working and how you can use it and utilize it in the Ansys environment. Now let's start with Ansys Simplorer first. In order to create a 3-phase inverter, we have two different options.

The first one, it can be from the grid, as we can see from this example, or it can be from a battery, like in electric vehicles. In this circuit, we have the voltage source to replicate the battery in the vehicle or in any project, and there are switches to represent the IGBTs or the MOSFETs.

And then we have three-phase connections into our electric motor. Here I use resistance and inductance as an extra part of the electric motor since my Maxwell model or electric motor is using, as you see here, the stranded windings.

When this option is stranded windings, that means Maxwell is not able to calculate the resistance and the inductance of your windings. In order to simulate this resistance and inductance in the circuit, we have to add external resistance and inductance into the circuit.

Here I assign 50mΩ for the phase resistance and 5uH for the phase inductance. After that, in order to get the feedback from the phase currents, I assign ammeters for each of the phases. And these phases are connected directly to the ABC to DQ conversion.

But here, as you see, we have to know the rotational position of the electric motor while it is spinning. To do that, we have to assign one rotational angular velocity source to provide the rotation to the electric motor.

In order to measure the torque, velocity, and rotational angle, we have to assign these three blocks and make the connections as shown. After we are having the angular position of the electric motor, we have to make this mechanical angle into the electrical conversion to use in our controller.

To do that, we are summing the first angle with 7.5 pi divided by 180. We are doing that because in our rotational settings, I added an initial value of 7.5 degrees to shift the rotor position in terms of stator position.

After correcting this shift angle, we have to multiply this value by 4, which is the number of pole pairs of the electric motor. And lastly, for the phase conversion, I am adding 1 pi into the electrical angle to be used in ABC to DQ conversion.

After here, I connected the phase currents into this conversion to achieve the ID and IQ currents. Here, in this model, I use PI controllers for each ID and IQ currents controlling.

Normally, you can also create one controller to set this reference input signal based on your torque or speed conversion. Here, in this model, I assign manual values. For example, for 3000 RPM modeling, I would like to have –7.37 ID and 38.2 IQ values.

And the PI values I basically tried and achieved the optimum value for my electric motor. You should optimize and modify these values based on your electric motor characteristics.

After we are having these PI controllers, we have to couple these PI controllers into the DQ and ABZ conversion to convert the signals into the SVPWM. Here we have the switching frequency of 1000 Hz and the dead time is 1E-06 and we are of course using the VDC as VBUS.

After completing the circuit, the main thing that is achieving this FEA model in Ansys Maxwell. In order to do that, you have to copy your electric motor into the same project of your Simplorer and go to your project in Maxwell and right-click and choose design settings.

And you can add the following settings here. From the pop window, you can go to advanced product coupling and you need to enable transient to transient link with Twin Builder. This will provide this electric motor model into the Ansys Simplorer that you can use it.

In order to add this transient model into the Ansys Simplorer, you can click TV Builder, add component, Maxwell component, and since it is a transient code simulation, you have to choose this setting and from the pop-up window, you have to make sure that the link type is Maxwell transient to transient and you chose your simulation for your electric motor.

After that, you need to make sure your solution is also correct. Once you click OK, you will see that the component is on your workspace. So you can drop it into anywhere that you would like to use in your circuit and make the connections as shown in here.

Here I connected the electric motor as star connections and coupled the phases directly with my circuits. That's the whole circuit for the battery FAT inverter.

For the grid FAT, the only difference that we need to modify, instead of having a single DC source in the entrance of the 3-phase conversion, In this version, we have to add 3-phase AC sources.

Here as you see, I assign the RMS value as a parametric value as grid_v and change the frequency as 60Hz and make the phases 250, 120, and 0 for each of the phases. After having these AC sources, you have to couple this one into the 6-pulse rectifier.

That's an ideal rectifier, you can find different components in Ansys Simplorer component libraries. And to filter this 6-pulse rectifier, we have to add a filter capacitor to the end of the circuit. Here I use 1uF for this circuit.

In order to use this value into our circuit, we have to add 1Vm and name it as Vm1 or any voltmeter names you want. The only difference that we will have here, instead of having this DC link voltage into our settings, we have to come and change this Vbus value as Vm.v.

In this way, the SVPWM block will use this method to generate the signals for the switches. After completing all the settings, you can simply run your simulation by clicking the analyze button. It may take from 5 to 20 minutes based on the complexity of your model.

After completing the simulation, you can go back to your Maxwell model that you actually generated the inverter FAT excitation and generate any results that you want to have. For example, for your reference, I created torque value of these results.

And for the loss, as you see, I also generated pure ZE core loss and I compare it with the inverter FAD FAA model. As you see, the difference is 25%, which is significantly important for electric motor calculations.

Here you can also see the phase currents of the electric motor with inverter FAT excitation. You can use these results for any of your simulation requirements and your system-level co-simulations. You can also do the same simulations in grid FAT simpler.

Here it may take more than the battery FAT simulations since it will generate the DC voltage from the AC voltage sources, and it will take each of these data sets into this SVPWM, and it will delay the simulation time significantly. This is the end of this blog.

You can use that system-level coupling for different applications. For instance, you can couple your I-Spec model into the Ansys Simplorer, and you can generate to create the thermal simulations of your electric motor and the inverter in the same system level.

I will also create a blog that shows how to generate a Q-valent circuit model of this FA model in order to achieve the simulation results way faster than the FA model. I also highly recommend you to go through the blog for any details and the steps that you would like to have in this blog.

Thank you for watching this video. Please contact us at https://ozeninc.com/contact for more information.