Wolfgang Gruber

Flux-switching permanent-magnet machines have gained popularity because their distinctive design places both the permanent magnets and the windings in the stator. This configuration not only improves the structural integrity of the rotor under centrifugal stresses but also makes cooling of the stationary permanent magnets easier and more efficient compared to rotor permanent-magnet structures. These machines are highly versatile and well-suited for applications requiring bearingless motors, including cleanrooms, medical equipment, and pharmaceutical production. However, because the magnetic core must accommodate both the windings and the permanent magnets, it experiences saturation, resulting in performance constraints. This paper proposes a novel bearingless partitioned-stator flux-switching permanent-magnet slice motor architecture. To support this concept, the study focuses on unaddressed challenges in the electromagnetic design of bearingless slice motors, particularly those related to core saturation and the optimization of torque and force performance. By separating the stator into sections containing permanent magnets and copper, iron core saturation is mitigated, resulting in improved torque density. Machine performance is further improved through winding optimization, where various winding schemes are evaluated. The proposed design and underlying principles are validated using 2D and 3D finiteelement analyses of the optimized configurations. Experimental validation confirms the theoretical findings, providing a thorough understanding of the operational characteristics of the motor.

This manuscript presents a reinforcement learning (RL) agent method to optimize the geometry of a coaxial magnetic gear using a 2D finite element magnetic (FEM) simulation. The proposed optimization algorithm aims to improve the maximum torque within given boundaries of the magnetic gear geometry by adjusting parameterized radii. A linear actor–critic gradient algorithm is implemented, where the actor learns a policy to adjust and discover the values of five geometric parameters of the magnetic gear model, and the critic evaluates the performance of the resulting designs. The RL agent interacts with an environment integrated with a 2D FEM simulation, which provides feedback by calculating the total torque of the new geometry discovered. The optimization algorithm uses a greedy exploration method that uses the total torque as a reward system, which the RL agent aims to maximize. The results obtained for the magnetic gear optimization demonstrate the effectiveness of the proposed RL algorithm, which can be applied to automate multiparameter geometric optimization using artificial intelligence systems.

In this paper, a magnetic-geared motor with an integrated bearingless high-speed rotor is proposed. Active two-axis suspension forces are provided to the suspended high-speed rotor by three-phase suspension windings. The principles of magnetic suspension and torque generation are presented. In addition, the suspension characteristics are evaluated in 3-D finite element analysis. A test machine was built, and both magnetic suspension and rotation were verified. As a result, the principle of a magnetic-geared motor with a magnetically suspended bearingless high-speed rotor has been confirmed for the first time.

A very simple and therefore wide-spread way to control motor speed and torque is to regulate the stator voltage by pulsewidth modulation (PWM). Depending on the modulation type, number of inverter levels, and modulation index of the PWM, various spectral components occur in the supply signal and inject ripple current in the motor phases. This current distortion can cause significant additional losses in the stator and rotor core, permanent magnets, and windings of an electric machine. Therefore, it is important to consider these losses in an early design stage. In this work, an efficient calculation method for harmonics of multilevel PWM signals and generated phase currents with different modulation types is presented. The PWM losses are calculated via 2-D finite element method analyses. The calculation and simulation workflow is realized within the software tool SyMSpace and applied to a three-phase permanent magnetic synchronous machine (PMSM). Loss measurements at different PWM frequencies and load points are conducted on a 3-phase PMSM. Since simulation and measurement results show good accordance, the proposed methods could be validated.

In recent years, a wide variety of motors have been used, but when manufacturing motors with different structures, it is necessary to make major changes to the winding manufacturing equipment, which causes an increase in cost. Therefore, the authors propose to manufacture the motor winding using a flexible PCB. In the proposed method, if there is no big difference in the structure and dimensions of the substrate, a winding with a high winding density can be easily manufactured at a relatively low cost. In this paper, we have manufactured a small magnetic levitation motor using windings manufactured using flexible PCB technology and investigated the effect of manufacturing error.

This article investigates and compares two popular combined winding configurations for bearingless machines, the multiphase (MP) and dual-purpose no-voltage (DPNV) windings. Each of these winding configurations has the ability to independently create suspension force and drive torque using the same stator coils. The article develops analogous bearingless motor force and torque models, space vector equivalent circuits, and a current regulation architecture for each combined winding approach. A comprehensive comparison is made between the two combined winding topologies by using these analogous models and through an experimental study. It is shown that, from a motor design perspective, these combined winding topologies are equivalent for certain design conditions, but that each winding places very different demands upon the drive power electronics. The article concludes that the optimal choice of winding configuration depends on the machine parameters and the expected loading and drive cycle. When compared in terms of total current stress exerted on the bearingless drive, the MP winding is preferred when over half of the coil current is used to produce torque rather than force. In terms of the drive's volt-ampere rating, the DPNV winding is shown to be advantageous for high operating speeds and low torque values.

This work deals with a compact and efficient permanent magnet excited synchronous machine to be used as paraglider towing winch. A battery supply with a voltage of only 48V is chosen for mobility and safety reasons. A rough machine design is first carried out analytically. After predesign, the final machine geometry is genetically optimized by 2D finite element method simulation using the software tool SyMSpace. After manufacture, the drive is commissioned with field-oriented control and the expected machine characteristics are compared with measurement results. Finally, also two self-sensing rotor angle estimation methods (one for the lower and one for upper speed range) are implemented and evaluated as back-up system to the installed rotor angle encoder.

The nonlinear radial forces relationship regarding phase currents and rotor angle demands a nonlinear control structure for bearingless motors. Most state-of-the-art control algorithms of bearingless motors are based on the assumption of linear current to force/torque relationship, and only the nonlinear relationship to the rotor angle is taken into account. However, in some types of bearingless motors, especially in motors with reluctance-type rotor, suspension forces are quadratic functions of the phase currents, which means that, in addition to linear terms, quadratic terms also contribute significantly to force generation. In this paper a novel control structure is presented, which allows controlling a bearingless motors featuring significant nonlinear current to force/torque relationship. DSP implementable solution was found by numeric solution followed by the 2nd order Taylor approximation. To check the functionality the suggested control structure was implemented with a bearingless flux-switching 12/10 motor.