Finite-control-set model predictive control of axially laminated flux-switching permanent magent machine with extended voltage space vectors

Publication Type:
Thesis
Issue Date:
2018
Full metadata record
The Flux-switching permanent magnet machine (FSPMM) has recently attracted considerable interest for high performance drive applications due to their high torque and high power density features. The laminations of traditional FSPMMs are radially laminated, i.e. steel sheets are laminated perpendicular to the shaft axis. Due to the nonlinear magnetic path, the radial laminations can have serious partial magnetic saturation at the edges/tips of stator teeth or rotor poles. The rated frequency of FSPMMs is usually much higher than traditional rotor-inserted PM machines at a given speed. In this case, the core loss of FSPMMs becomes evident especially beyond the rated speed, which leads to decrease of output power, torque/power density and efficiency. The reluctance motor with axially laminated rotor has received growing interest in recent years. This type of motor can achieve a higher torque density compared with segmented rotors and flux-barrier rotors. In this thesis, an axially laminated flux-switching permanent magnet machine (ALFSPMM) with HiB grain oriented silicon steel stator and rotor cores is proposed. The HiB silicon steel features high permeability and low specific core loss, and as a result, the total power loss of proposed motor is much lower than the conventional FSPMMs. The detailed fabrication procedures are presented. The theoretical characteristics of ALFSPMM are calculated by 2D finite element method (FEM). Experimental measurements of the prototype machine are presented to validate the FEM calculation. On the machine control side, the direct torque control (DTC) is one of the most popular control algorithms. It features simple structure and fast dynamic response. However, the performance of DTC in terms of torque and flux ripples and drive system efficiency is unsatisfactory since the voltage space vector (VSV) is selected heuristically. Recently, the finite-control-set model predictive direct torque control (FCS-MPDTC) has been developed as a simple and promising control technique to overcome these problems. The FCS-MPDTC still suffers from relatively high torque and flux ripples due to the limited number of VSVs. This thesis proposes a novel FCS-MPDTC with an extended set of twenty modulated VSVs, which are formed by eight basic VSVs and twelve extended VSVs by modulating eight basic VSVs with fixed duty ratio. To mitigate the computational burden caused by the increased number of VSVs, a pre-selective scheme is designed for the proposed FCS-MPDTC to filter out the impractical VSVs. The drive system efficiency is also investigated. The theory and simulation are validated by experimental results on the ALFSPMM prototype.
Please use this identifier to cite or link to this item: