Sensorless control of PMSM

[soundaryaParticipant]

I am getting error after completion simulation time.My simulation time running is very slow.Even it is not reaching to 1 sec. By the its reaching its showing an error that due internal problem Matlab have to end now.I will attach one word file and my simulink model .

Attachments:

You must be logged in to view attached files.

[soundaryaParticipant]

I am sending my model.

Attachments:

You must be logged in to view attached files.

[cKKeymaster]

The model runs slow due to strong chattering. Infact many zero crossings are ignored as they cross the default settings. While this may not always mean the simulation results are wrong, they may well be so. Be careful in interpreting results. Also, its suggested to use ode15s, as this implicit solver might be better suited in situations like this.

I couldn’t reproduce the Matlab crash that you are experiencing. It could be system or version specific. Try simulating on a different computer and a later version. I could produce results in two versions 2011a & 2012b. Picture of results attached.

• This reply was modified 10 years, 7 months ago by  cK.

Attachments:

You must be logged in to view attached files.

[soundaryaParticipant]

Thank you.I tried by changing solver and even i removed Enable zero crossing but still there was problem memory allocation.I didn’t try in other systems .I will try for it. And i am sending one more model in which i reduced the complexity.If it is possible please try to simulate it and attach results in workspace……..and attach to my mail.

Attachments:

You must be logged in to view attached files.

[cKKeymaster]

You shouldn’t disable zerocrossings. They are critical for fidelity of simulation. If you are running into memory problems try to minimise signal-logging/plotting to only signals of utmost importance. Or, you can try running simulation multiple times and log only a few signals in each run – for this you shouldn’t have any randomness in the model. Alternatively, a computer with more memory should help.
Will try to post sim results here.

[cKKeymaster]

Even w39 has chattering problem. Looks like ode23t and ode23tb are more suited here (empirical observation). Always use adaptive zero crossing algorithm.
A good example where a solver can yield incorrect results – ode113.

Didn’t simulate until end, as the simulation was too slow. Results until 0.06s attached.

[cKKeymaster]

Attached.

Attachments:

You must be logged in to view attached files.

[Author]

Posts