Struggling with Custom Motor Integration and PID Tuning in IPMSM Velocity Control

Hello MATLAB Community,

I am currently working on customizing the IPMSM Velocity Control example to incorporate my own motor parameters. Despite several attempts, I have been unsuccessful in achieving the desired control performance. Here are the specific motor parameters I am using:

• Pmax: 60,000 W (Maximum power)
• Tmax: 80 Nm (Maximum torque)
• Ld: 0.0008 H (Stator d-axis inductance)
• Lq: 0.001 H (Stator q-axis inductance)
• L0: 0.0005 H (Stator zero-sequence inductance)
• Rs: 0.01245 Ω (Stator resistance per phase)
• psim: 0.06 Wb (Permanent magnet flux linkage)
• p: 4 (Number of pole pairs)
• Jm: 0.0012 Kg*m^2 (Rotor inertia)
• Back EMF constant: 0.233 V/(rad/s)

I have tried using the Step Response Optimizer and the "Tune Field-Oriented Controllers Using Closed-Loop PID Autotuner Block" to adjust the P and I values for D, Q, and the speed control loops, but I have not been successful.

Could anyone provide guidance or suggestions on how to properly adjust these parameters? Specifically, I am looking for advice on:

1. Properly integrating my motor parameters into the IPMSM Velocity Control example.
2. Effective methods for tuning the PID controllers for the D-axis, Q-axis, and speed control loops.

Any help or insights would be greatly appreciated.

Thank you!

[Sam Chak]

Hi Alex, have you used some functions from the System Identification toolbox for estimating system model such as state space or transfer function?

[Alex]

Thank for the suggestion! Can you recommend examples or tutorials to speed up my process?

[Finn]

A good start for tuning your pid controller, whould be to tune it from your motor parameters. You can use two methods. I would recommend you to tune the current controllers with the amplitude optimum. After calculating the PID-Controller gains you can look at the step responses (d-Axis and q-Axis seperatly). To exclude any coupling resuting from back-emf you can lock the rotor, so it don`t rotate when you apply a q-current. When the current loop shows good behaviour you can tune the speed controller with the symmetrical optimum.

This two methods will give you a decent control performance. If you want to optimize, you can go on and for example use the closed loop pid tuner.