VFD Motor Parameter Auto-Identification

[Jieve]

Hello,

I was wondering if anyone could shed some light on what actually happens during the stationary and rotational motor parameter auto-identification that takes place when commissioning drives. The Siemens G120 units I've been working with apply something like 20V on and off to the motor (1hp, 400V) in stationary mode, then run it at around 1200 rpm for around 30 secs with 3-4 quick accelerations in the rotational ID mode. What parameters specifically is the drive measuring during each of these processes?

In sensorless vector mode, the drive does both the stationary measurements and the rotational measurements when I select that both should take place, but in V/hz mode it only does the static measurement. I guess the drive doesn't need any measured rotation parameters in V/hz mode, it just calculates the V/hz slope based on the motor data input on commissioning. This seems somewhat weird though, as the curve becomes non-linear at lower voltages.

Finally, should the motor always be decoupled from the load during this process? Or connected to it (assuming no damage can be done)?

Thanks for any responses.

 

[itsmoked]

Hi Jieve. You don't really want to know.

It's pretty daunting stuff. In recent years drives have migrated their control implementation into what's called Field Oriented Control or FOC. FOC uses some mind bending matrix calculations called the Clarke and Park transforms. The transforms convert the instantaneous complex motionary aspects of a running motor into a static non-rotating orthogonal model that can be monitored and manipulated using faster math and ignoring voltage before being inverse transformed back to the rotational realm and applied to the windings via the output stage of the drive. The values they need to carry out FOC successfully are;
Number of motor poles
Motor R[sub]r[/sub]
Motor Magnetization Current
Motor R[sub]s[/sub]
Motor Ls[sub]d[/sub]
Motor Ls[sub]q[/sub]
Motor Rated Flux
Probably a few I forgot too.

Then there about about 10 more things the drive has to find out to carry out its task, these include voltage offset calibrations of all three phases and current offset calibrations of all three phases to account for the actual hardware differences found between production drives.
Next all the Proportional, integral, and derivative aspects of tuning the PID need to be discovered, they depend a lot on the motor's inertia and available torque.

The short answer is: The motor's resistance, inductance values, and inertial values.

The difference between vector and volts/hertz is that in all modes not volt/hertz the drive must gather all the above info because it is running FOC. In the V/Hz mode the drive knows nothing about a rotating motor and simply synthesizes a 3 phase power output that includes a simple V/Hz voltage modification control.

The motor should generally be disconnected since the drive desperately wants to completely understand the motor correctly so it can carry out FOC. Having a bunch of stuff hooked to the motor shaft can skew the drive's understanding of the motor's variables.

Actual PID tuning of the "systems" may warrant having the load connected, however some drives cheap out and lump the FOC and PID stuff into the same "tuning" and make it hard to do the right thing. If the drive truly has a PID tuning mode verses a motor 'understanding/FOC' mode then often you'd be better off having the load attached for PID tuning.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Jieve]

Keith,

great answer, thanks for that very detailed and informative explanation!

 

[itsmoked]

Glad I could help in this case.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[singapore11singapore]

Hi,
When you do a standstill rotation one of the parameters to set is the motor temp...P652 "motor ambient temp"...it is very important you tell it the correct value, as this has a huge effect on the calculated values.
The four spikes you see towards the end of the rotational test....is caused by P1959 (0).0....the "enc test active"...it is basically ensuring your encoder is following the actual as to be expected....
Regards Singapore11

 

[itsmoked]

Thanks for that excellent point singapore.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Jieve]

Singapore,

How significant of a difference are we talking regarding ambient temperature? I've noticed upon commissioning with the commissioning wizard, it doesn't ask for motor ambient temp. This is something you really have to dig through the parameters to find, which seems weird if it's that important. You're saying this parameter should be changed to the actual ambient temp immediately before starting the motor identification process, after commissioning with the panel or "starter" software?

Also, the acceleration pulses take place upon rotation identification in sensorless vector mode, so without any encoder. I had assumed that this was some type of inertia measurement.

Thanks.

 

[Jieve]

Hi,

never got a response to my last post and am still wondering about the ambient temp setting, can anyone shed some light on that?

Thanks

 

[singapore11singapore]

Hi,
We initially found the same.....so left the temp at default value....but only later when testing on our test bed we found that inputting the correct temp value gave us smoother reactions during running. Therefore, in our case....a drilling application, we call the chair and ask them what temp these see on the motor temp display and enter that into Starter as the ambient temp.
For turning measurement with encoder, I set P1959 (0)(rotating measurement configuration to 7H...which activates moment of inertia, saturation characteristic, and enc test only....the rest of the selection are Off.
In general, all always have encoders.
I will do a standstill test and then rotate in encoderless whilst monitoring the encoder output. Run like this for a bout 20 mins to warm the windings and then do the rotating test with encoder.
Regards
Martin