The multi-phase open-winding induction motor and its adaptive H-bridge multi-phase inverter system have received extensive attention due to their advantages of small torque ripple, strong fault tolerance, and easy power capacity expansion. This paper analyzes the modeling of the multi-phase open-winding motor system and speed sensorless control technology to achieve high degrees of freedom control and low switching frequency characteristics in a twelve-phase, large-capacity, open-winding motor system. The simulation and experimental verification are conducted to enhance the operating performance of the low-switching-frequency multi-phase open-winding motor system.

The twelve-phase open-winding induction motor system and its equivalent three-phase simplified model are established. The equivalent three-phase full-order observer model and its speed estimation method are presented. The full-order observer used in speed sensorless control has the advantages of low control bandwidth requirements and a wide range of observation speeds. However, the switching frequency of the H-bridge large-capacity inverter supporting the ship’s multi-phase open-winding induction motor is low, which inevitably increases the digital discretization error of the full-order observer. This paper derives the full-order flux observer models based on the forward Euler method, the simplified second-order discretization method, and the proposed Adams fourth-order discretization method. Then, the steady-state error and observer stability are compared using the F-norm and pole diagram. Theoretical analysis reveals that the full-order observer, based on Adams' fourth-order discretization method, achieves the best discrete accuracy and stability of the observation system while minimizing the computational complexity of the digital control system.

A simulation model and a test platform for a twelve-phase, 25 kW open-winding induction motor with speed sensorless control have been developed. The results show that, compared with the forward Euler method and the simplified second-order discretization method, the observation results for speed, current, and flux based on the Adams fourth-order discretization method are almost consistent with the actual values. Through the speed sensorless closed-loop speed regulation test and load mutation test, it is further verified that the full-order observer based on Adams' fourth-order discretization method exhibits good speed regulation and load-carrying capacity, which can achieve better dynamic and steady-state performance under both extremely low and high-speed conditions. The full-order observer discretization method can provide technical support for applying speed sensorless control technology to low switching frequency multi-phase open-winding motor systems.