2
\$\begingroup\$

I would like to construct an electronic virtual 3-phase motor for testing VSD/servo drive systems in the workshop, where the actual motor is not available for testing (e.g. because it is on a running machine, or it hasn't arrived yet, etc).

Let's assume it's a squirrel-cage induction motor with fixed ratings (1hp) for now. Of course, it would be easier to obtain such a motor and just wire it to the drive, but where I'm heading is a set of electronically configurable motors for testing many different systems.

  1. Can I do this by connecting appropriately sized inductors and resistors between the drive and a 3phase inverter? I.e. would this be able to produce an accurate model of such a motor, assuming smart enough control of the inverter? There would be voltage test points so that the inverter controller can model the motor accurately. See below:

schematic

simulate this circuit – Schematic created using CircuitLab

  1. If simulated encoder feedback is added, can this architecture also be used to model permanent magnet 3phase servomotors? If no, what would need to be modified? If yes, would the inverter controller need to produce complex (non-sinusoidal) waveforms?

  2. To make an electronically configurable motor, can I switch different inductances and resistances in and out using contactors / relays?

\$\endgroup\$
2
  • \$\begingroup\$ Without "regeneration", be prepared to turn 1 kW to heat. \$\endgroup\$
    – greybeard
    Commented Jun 2, 2023 at 7:09
  • \$\begingroup\$ Yes good point; I am going to use a smart 600VDC power supply that can feed back into the grid. Unfortunately it'll have to be 3ph but the DUT is typically 3ph, so this will not likely undermine the device's utility. \$\endgroup\$ Commented Jun 2, 2023 at 23:24

1 Answer 1

0
\$\begingroup\$

So, it turns out my Google Fu is a bit rusty. This kind of device is called a Motor Emulator, and large versions of them are used extensively in the automotive industry for testing EV car parts and can model VSDs and servos, with encoder feedback.

Here are some academic papers on the topic:

http://www.iri.upc.edu/people/riera/VPPC10/vppc2010.univ-lille1.fr/uploads/PDF/papers/RT6/95-14197-final.pdf

https://eprints.whiterose.ac.uk/114042/

https://repository.lib.ncsu.edu/handle/1840.20/38865

Following on from those publications and my SPICE simulations, the basic idea of the circuit is right for the stated objectives, but:

  1. The series resistors are not needed and not helpful. Inductance is definitely required to stabilize the current flows between DUT and emulator. Some systems employ LCL circuits instead, but they require special attention, and a simple inductor is more robust.
  2. Switching inductances in and out is not required and not helpful, as we do not need to try to match the physical motor's inductances; that is the job of the inverter/current controller. An excessive series inductance limits the current slew rate and thus the ability to emulate highly dynamic systems.
  3. I drew BJTs; in fact, fast switching (e.g. MOSFET) is needed because the control loop required to emulate the motor electrical system is very fast (hundreds of kHz).
  4. Both voltage sensing (at the location shown above), and current sensing (through the inductors) is required. The sequence is: DUT outputs voltage->emulator reads voltage->emulator updates electrical model state and sets current target->current through inductors is controlled->DUT adjusts voltage->......
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.