Electric motor torque control based on speed interaction/instability

[MiguelS]

TL;DR Summary

Problems trying to induce a load torque to one motor through another motor, due to motor #1 having a torque control based on speed, and motor #2 having a speed control.

Hi everyone

I'm trying to exert a controlled load into an electric motor with another motor (having first motor spin at a lower speed than second motor, exerting a torque).
Both are connected through a shaft, and we're measuring speed and torque at the shaft.
Motor #1 has a torque control based on speed commands (by speed command differences exerting acceleration, going into Torque = Inertia*Acceleration equation, passing through a PI controller).
Motor #2 has a speed based control.

When Motor #1 tries to exert torque to Motor #2 lowering its speed (spinning at a lower speed than Motor #2), Motor #2 sees this speed decrement and exerts torque to try to bring back the Motor to its commanded speed.
This is driving the control very unstable and unable to keep a controlled torque, due to this interaction.

Has anyone experienced something similar, where you would try to control torque to a motor, but due to the control being speed based, both motors enter a control loop interaction?

Any torque control schemes you would recommend, based on speed commands? (We can modify Motor #1 torque control algorithm and tune it/change programming, but not Motor #2)

Please let me know your thoughts or any question/comment.

Thank you very much : )

 

[jrmichler]

If I understand correctly, your setup is like this:

Just two motors connected by a common shaft, with no other load. Your description is unclear, but my understanding is as follows:

1) You want to drive Motor #2 at a specified speed.
2) That speed is constant (or relatively constant) when the load changes.
3) You want Motor #1 to apply a specified torque to Motor #2.
4) That torque acts to slow Motor #2 down, requiring more torque from Motor #2.

If all of this is correct, then you apparently have a motor test stand. It's unclear if you are testing Motor #1, Motor #2, or both motors. If all of this is correct, then we need to know what types of motors you have. Are they induction motors with VFD drives, or some type of servomotor with servo controllers? Are they industrial motors or those hobby motors that the Arduino people experiment with? Some make and model numbers would be very helpful.

 

[Baluncore]

You have not specified the type of motor.

To eliminate hysteresis in a positioning system, two DC electric motors can be operated together in parallel, to position the load through two shafts driving the same final bull gear, or a shaft from two ends. DC motor torque is proportional to motor current. The two parallel motors are operated with a fixed DC offset current. For small loads there is a fine weight balance, while for larger loads the motors work together in the same direction.

To do the same with two AC induction motors, you must account for the torque to slip curve, and manage the motors by driving them with two different frequencies, offset by a fixed AC frequency, rather than being offset by a fixed DC current. That will balance small loads in sensitive opposition, while for larger loads it operates the motors in parallel.

In either case, you operate only one control system, driving the two motors from the same one PID controller, but with a fixed offset DC current, or fixed offset AC frequency.

 

[onatirec]

This is how we test motors on our dynos here - speed control on the dyno motor and torque control on the test motor. I don't see why you would base your torque control on speed differentials by some Torque = Inertia*Acceleration equation. This seems inherently very noisy and would need a lot of filtering to smooth out the torque. It's not clear what type of motors these are, but i'd recommend controlling torque independently on motor 1 by, for instance, controlling the current directly. You could use feedback from your torque transducer to dial in the current/torque on motor 1, but keep it separate from the small variations in speed.