Conference Agenda

The sizing of Axial Flux Machines (AFM) remains a non-trivial task, mostly due to 3D geometry. Generally, 3D FEA software is used for a comprehensive analysis of such machines, which may require long calculation times. In the context of pre-sizing, the use of FEA software is not the most relevant option, analytical models being more suitable due to shorter computation times which makes possible fast analysis of various geometries. This paper proposes a fully analytical model for calculating the open circuit magnetic field of a slot-less Axial Flux Permanent Magnet (AFPM) motor. Based on mirroring techniques, the combination of 2D equations in two normal planes to calculate the 3D magnetic field is the novelty proposed in this paper. Sensitivity analysis of geometrical parameters can be easily performed which makes it a perfect tool for pre-sizing process. A fair comparison with 3D FEA software (ANSYS Maxwell 3D) based on a set of motors identified from technical literature highlighted the performances of the model regarding calculation speed and accuracy. The limitations of this model are also shown and discussed.

Machine learning methods are an interesting tool for automatic fault detection in electrical machines, based on the measurement of a physical quantity of interest, but need large datasets to be accurate. As there might not be available measurements for some loads or defects, numerical simulations based on the Finite Element (FE) method may be used as an alternative for dataset generation, but numerous solving of nonlinear magnetostatic problems coupled with a dynamical circuit equation system to represent the load leads to huge computational times. Then, Model Order Reduction (MOR) approaches based on the Proper Orthogonal Decomposition (POD) coupled with the interpolation method Energy Conservating Sampling with Weight (ECSW) enables to reduce the computational time. In this communication, we propose a methodology to construct a reduced model of a generator in the nonlinear case, valid on a wide range of operating points. The reduced model is evaluated in terms of accuracy on local quantities of interest and speedup, compared to a FE model.

The paper presents a new fast hybrid reluctance network for the analysis of surface-mounted permanent magnet machines under load during the preliminary design step using optimization methods. The proposed hybrid reluctance network model takes into account the effects of magnetic saturation through equivalent current densities, expressed by Fourier series. It combines analytical methods and reluctance network to accurately predict the magnetic flux densities, the linked flux, the voltage under load condition and the torque waveform. The proposed hybrid reluctance network represents the field generated by the stator windings and the stator saturation effects taking into account the permanent magnet permeability. The results are validated against the finite element method, showing strong agreement with the numerical solution and more accurate results than those obtained with the methodologies available in the literature.

This paper presents a multi-static reluctance network (MRN) to carry out the modelling electromagnetic behavior of a surface Permanent Magnet Synchronous Generator (PMSG). The main objective of this work is to propose a model offering a good compromise between accuracy and simulation time. The rotational motion modelling is achieved by considering the displacement of the magnetomotive sources representing the permanent magnets of the machine. In this way, the reluctances of the air gap do not change during the rotational motion and eliminates the need of node reconnection techniques. The results obtained with the proposed methodology are compared with the Finite Element Method (FEM), and the results have strong agreement.

This work proposes an analytical model to determine the parameters needed for the sensorless vector control of coreless axial flux permanent magnet machines. A reluctant model is used to compute the flux linkage constant while the phases resistances and inductances are calculated from coils geometry. The model is computationnaly light and can be easily embedded in a low-cost microcontroller to tune the control law. The approach was validated with finite element analysis and experiments. The domain of validity of the model was explored through a sensitivity analysis. Experimental tests were conducted on a 1 kW wind turbine axial flux generator to validate the proposed approach. By enabling embedded model-based controller tuning from the generator geometry, this study intends to contribute bridging the gap between magnetic computations theories and practical high performance control of electrical machines.

The shaftless pump-jet propulsor achieves superior speed and efficiency in autonomous underwater vehicles (AUV) due to its compact, low-noise, high-performance integrated motor and blade design. This study optimizes the collaboration between the blades and the motor, employing Computational Fluid Dynamics (CFD) for fluid and thermal analysis and Finite Element Method (FEM) for electromagnetic modeling. Through multi-field coupling and parametric design, the research systematically verifies the propeller's performance, with experimental results demonstrating that it can attain speeds up to 26 knots. This work represents a substantial contribution to the advancement of underwater propulsion technology.

The development of electrification has led to increasing demands on the performance of electrical equipment. And increasing attention has been paid to the performance of magnetic materials in electrical equipment. However, the current magnetic material property data has not been able to form a centralized management platform from the user's point of view, which is not conducive for practitioners to find, compare the property data and subsequently compute . This paper introduces an open access magnetic material database for electrical equipment from its composition, structure, operation mechanism and core functions. The material property data in the database covers a wide range of materials and a wide range of measurements, which can provide material property data support for the electrical industry.

This paper investigates the effect of magnetization state (MS) variation of low coercive force (LCF) magnets on the variable-leakage-flux (VLF) property in a newly developed hybrid-variable-flux permanent magnet machine (HVF-PMM) by adopting a coupled solution model (CSM), which combines the frozen-permeability method (FPM) with a nonlinearity-involved parallelogram hysteresis model for LCF magnets. By employing the CSM, the PM flux linkage at different MSs is separated from the main magnetic flux linkage, and the VLF characteristic under different MSs can be obtained and analyzed. On the other hand, the individual impact of different types of magnets on the magnetic saturation of the leakage flux bypass can be thoroughly analyzed, considering the magnetic interaction between the two types of PMs. In this case, the effect of MS variation on the VLF property can be identified and accurately quantified, thereby offering the possibility for realizing the trade-off between torque capability and flux regulation range. Finally, a prototype is experimentally validated to verify the theoretical analyses.

Amorphous alloys show great potential for application in the stator of electric machines due to their low iron loss characteristics. However, Amorphous alloys have a low saturation density and are prone to density saturation under high load conditions, which limits the performance of the machine. In this paper, a stator structure with a hybrid core of amorphous alloy and silicon steel is proposed to balance low iron loss and high load performance. Meanwhile, a polynomial method is used to establish the shape of the splices of the hybrid core material and to explore the effect of the geometry of the splices on the performance of the machine. Moreover, an agent model based multi-objective optimization approach is used to optimize the spliced structure to address the difficulty of optimization due to the large number of shape parameters at the splices. The results show that hybrid core machines are optimized to reduce the reluctance and improve the flux distribution to improve the machine performance compared to amorphous alloy stator machines. Finally, a prototype was built and the significant effect of the optimized hybrid core structure in enhancing the performance of the machine was experimentally verified.

Variable flux memory machines (VFMMs) embrace the benefits of adjustable air-gap magnetic flux and wide speed range. The key characteristic of VFMMs refers to the utilization of low coercivity force permanent magnets (LCF PMs), which can be either re-magnetized or de-magnetized by applying pulse currents to manipulate magnetization states (MSs) with minimal loss. However, the finite element method (FEM) is needed to obtain the specific parameters under different MSs when the analytical method is used to quickly analyze the characteristics of VFMMs, significantly increasing the complexity of the analysis. Therefore, this paper first elaborates the MSs adjustment mechanism through a measured numerical hysteresis loop model, and derives a function related to relative permeability versus magnetic field changes. Further, a nonlinear model linking magneto-motive force, magnetic reluctance, and pulse current is derived using a simplified magnetic circuit. The developed numerical model of LCF PMs is then integrated into a magnetic circuit model of a dual-layer VFMM. Consequently, the analytical predictions are obtained and compared with FEM results to validate the effectiveness of the proposed LCF PM model. The detailed model description and experimental validation will be provided in the full paper.

An automatic loader for feeding electromagnetic railguns at high repetition rates is currently being investigated at the French-German Research Institute of Saint-Louis (ISL). The principle of the loader is based on acceleration by induction similar to the functioning of a coil launcher. The major advantage of this principle is the compact design of the loader and, since no contact between the static and the moving part is needed, an almost frictionless injection of the launch package into the railgun is feasible. The challenge of this work consists in the design of the loader by performing two steps of numerical simulation. In a first step the basic parameters of the loader and its dimensions are determined by means of a fast-running numerical in-house code written in FORTRAN, allowing design optimization over high number of iteration loops. The resulting launcher dimensions as well as the time-dependent current profile are introduced in the second step into a 3D FEM model (LSDYNA), in which the mechanical and electromagnetic solvers are coupled. This step allows verifying the mechanical consistency of the loader subjected to magnetic pressure, as the materials models of the involved structural parts take into account elastic, plastic and strength behavior. In a third step, the most relevant solutions were tested on the railgun facility of the medium caliber launcher RAFIRA (25 mm square caliber) at the ISL laboratory. We could show that the numerical results are in good agreement with the outcomes of the experimental work, thus increasing the confidence in the two-step numerical approach.

Inter-turn short circuits in the stator winding constitutes a critical fault in induction motors, significantly impacting their operational reliability and efficiency. Although in the literature inter-turn short circuits are commonly considered as single phase faults, in this article, short circuits in all three phases of the machine are analyzed. The proposed approach accounts for detailed electromagnetic modelling and the asymmetry caused by multi-phase faults. The presented results can be used for advanced diagnostic techniques for multi-phase inter-turn short circuits.

Inter-turn faults in the stator winding represent a critical failure in induction motors, significantly affecting their reliability and efficiency. Although extant studies mainly focus on faults within a single phase, there is a paucity of publications presenting the results of simulation calculations or measurements of inter-winding faults in more phases of the stator winding. The present study aims to address this lacuna by presenting the results of a continuous wavelet transform of the stator winding current waveform, with consideration of inter-turn short circuits in all three phases.The results obtained provide insights into the complex fault dynamics and their impact on machine performance, thus forming the basis for improved diagnostic techniques and deepening the understanding of fault mechanisms in induction motors.