Solid rotor induction motors have a solid body rotor, which leads to low efficiency and power factor, and currently, the rotor is mostly optimized by slotted and squirrel cage structures. A generalized multilayer analytical model for different rotor structures is established, which can consider the effects of rotor eddy currents and saturation, based on which a generalized equivalent circuit model is established. The effects of number of slots, depth of slots, width of slots, squirrel cage material and end ring thickness on rotor impedance, torque and rotor losses are analyzed. On this basis, with efficiency, power factor, starting torque and starting current as the optimization objectives, and the number of slots, slot depth, slot width, squirrel cage material and end ring thickness as the optimization variables, the optimization schemes of slotted rotor and squirrel cage rotor are obtained by using the three-dimensional finite element method. The theoretical analysis is verified by finite element simulation and prototype experiment, and the results show that the electromagnetic parameters of solid rotor induction motors with different rotor structures can be accurately calculated using the universal magnetic field analytical model and the universal equivalent circuit model with an error within 5.8%. Slotted and squirrel cage rotors can effectively improve the motor power factor and efficiency, but this will lead to a decrease in starting performance. For the optimization function established in this paper, compared with the smooth rotor, the performance of the squirrel cage rotor is improved by 6.08%, which verifies the accuracy and validity of this paper and the optimization design scheme.

Accurate loss evaluation of high-frequency transformers (HFTs) under non-sinusoidal excitation is essential for the optimal design of dual-active-bridge (DAB) converters, yet remains challenging due to spatially non-uniform magnetic flux distribution across the core and spectrally rich harmonic content in the winding current. This paper proposes a dual-refinement electromagnetic loss analysis method that combines a topology-aware region-partitioned improved generalized Steinmetz equation (IGSE) with a harmonic-resolved Dowell model. The ferrite core of a shell-type HFT is partitioned into three magnetically distinct zones—core legs, side pillars, and yokes—with the IGSE applied independently to each zone using zone-specific flux density waveforms derived from the single-phase-shift (SPS) control model, thereby capturing spatial non-uniformity in core loss. For winding loss, the current is decomposed via fast Fourier transform, and per-harmonic AC resistance is computed up to the seventh order with frequency-dependent skin depths, accounting for the spectral contribution to copper loss. A 300 V/6 kVA ferrite-core foil-winding HFT prototype was fabricated, and the proposed method was validated through COMSOL finite element simulations and experimental measurements using open-circuit and short-circuit tests. Results show that the calculated core loss and winding loss agree with measurements within 10–15% and 8–10%, respectively, demonstrating improved accuracy over the conventional uniform-IGSE and total-RMS Dowell approaches.

Aiming at the chattering issue in speed observation for sensorless control, this paper proposes a sliding mode observer based on an improved double-power reaching law for high-performance speed estimation in five-phase induction motors. Traditional constant-rate reaching law observers exhibit significant chattering, while the double-power reaching law, though offering certain “variable-gain” adjustment effects, still has limited chattering suppression capability. To address this, the paper introduces a state variable related to the stator current into the conventional double-power observer, further enhancing the ability of the sliding mode gain to vary with the system state. This approach effectively suppresses chattering while maintaining convergence speed. The stability of the observer system employing the new reaching law is proven using Lyapunov stability theory, and the value ranges of key parameters are determined. Simulation results demonstrate that, compared to traditional constant-rate reaching law and conventional double-power reaching law observers, the proposed improved method significantly reduces speed observation chattering and effectively enhances the observation accuracy of the observer.