I have been designing a model of an already existing actuator which is run in FOC with an SVPWM inverter and cascade speed-current control. Losses like eddy current or hysteresis, or switching losses are kept ideal. The motor which I am using is a PMSM.
As I must rely on the real hardware, I am using the same configuration datas for the controllers. Now I am having trouble with my model, because the results are not really acceptable since the actual speed lies above my target speed whenever I use saturation blocks for the speed controller. The range of them lies within rated torque and 0 and the speed stays above the desired speed. When I set up the range between rated torque and -rated torque, I am having oscilations around the target speed. The output of the speed controller lies then within the range of my current saturation block. I keep producing huge current spikes whenever the motor reaches the reference speed. How to avoid them?
BTW: Since there is no feedforward compensation integrated in the digital controller of the hw, I didnt implement one in my actual model. I read that in FOC u just inject the q-term of the FF compensation and dont inject the d-term. Is that right?