Your workaround is a good way to get around it -- however, it makes it such that an initial condition of zero velocity will get you "stuck" in that position no matter how large your power is.
This is something you will have to deal with because your model uses power instead of torque. If you were supplying torque instead, you could directly integrate that using T = J*w.
Some other approaches you could try, depending on your problem:
- Make sure that your shaft has an initial "idle speed", so you cannot have a speed less than a certain value.
- Apply a small minimum power value at zero velocity. That is, instead of the zero-velocity case yielding zero power, it is able to push the shaft a little bit to get it started.
- Sebastian
