This master's thesis discusses the control of a 6-phase permanent magnet synchronous machine (PMSM). The introduction presents the development history of multiphase machines, their applications of use, reasons for their use, and a brief comparison between 3-phase and multiphase machines.
The central part focuses on the mathematical model of dual 3-phase PMSM, which is the base of our double 3-phase PMSM model for simulations in Matlab and Simulink. First, we developed the inverter model which controlled PMSM model using two individual current control. Then we implemented Vector Space Decomposition (VSD) control, which decouples the machine on a single flux/torque producing dq subspace and a single non-flux/torque producing dqz subspace.
Both control techniques were compared based on responses of current PI controllers. Decoupling the machine into two subspaces with VSD control enables optimization of current PI controllers, reducing the instability of controllers with two individual current control. The simple relationship between two individual current control and VSD control allows retaining the optimized current controllers from VSD control and incorporate them in two individual current control.
Poor response of optimized current PI controllers with optimized parameters at higher rotational motor speeds called for a self-tuning PI controller feature that retains fast and yet stable response even at higher rotational speeds.
Torque control in VSD is controlled only through current controllers in dq subspace. Because this subspace is decoupled from the dqz subspace, optimization of current PI controllers in dqz subspace doesn't affect the motor torque output. Current controllers in dqz subspace only define how fast and stable is the power-sharing between 3-phase winding sets.