I worked with this demo too.
Because this problem is quite simple, some easy analytics can be used to start the identification.
In short, the load torque (x axis on your data) Tload is at equilibrium equal to the machine torque Tem=K*I minus the friction torque Tfrict. You can use this equation to identify the blue curve I=(Tload+Tfrict)/K.
(In this case I consider the viscous friction = 0.)
Also the electrical voltage equation U=R*I-K*W holds and can be used for the green curve W=(U-R*I)/K
If you use 2 torque load points, it gives you 4 equations to identify K, R and Tfrict, and if the modeling hypothesis are right the 4th should be consistent.
(be careful if this data was obtained using a battery and relay, the R value corresponds to Rmachine+Rrelay+Rbattery)
I think it is faster than setting an optimization, even though you can still try it after.
Ok.... I've got the hang of it now!
Here is the procedure, how I set a DC motor parameter
Set signal source and reset friction paremeters
Run Simulation and plot current graph
Amplify the ampere value 119 times
Now, I can create the current graph i want
(1.8 is default value)
Using W=(U-R*I)/K formula to find resistance
When ampere is 5.5, Power is 0
So, the resistance value is 12/5.5
Set resistance value and plot power graph
Finally, complete whole performance curve
Well.. Actually, that is not the final step...
I tune the friction value to set the efficiency
I think I've finally done it
thank you so much
here is an example for: How you do it with Amesim optimization.
You can use the Demo qthelp://lmsimagine.lab/ame_dir/demo/Platform/Opti
As you can see I used the simple least squares as objectives.
=> You can find more details in the demo documentation of check valve optimization.
What helped me a lot: