Conference Agenda

Non-destructive testing for electrically conductive objects is usually carried out by impedance measurements using eddy current coils. Analytical integral impedance models for these type of problems are well-known in the literature. These integral models typically have a long calculation time. Therefore, series expansion models have been developed to speed up the calculation time by using the so-called Truncated Region Eigenfunction Expansion. This method is typically used to model single coils which consist of one single winding to generate and measure electromagnetic fields. Nevertheless, using a separate excitation and measuring system in an eddy current testing sensor is often advantageous. This paper applies the truncated region eigenfunction expansion method to a cylindrical excitation coil involving two vertically and coaxially symmetrical pickup coils placed next to a coated sheet metal. Furthermore, the advantages of the presented analytical model in terms of calculation time improvement and their applications with regard to the parameter estimation of coated steel sheets are shown. In addition, the new series expansion model offers much better error control because the number of series elements can be specified.

Optimization of power transfer efficiency is an important way to increase the performance of a wireless power transfer system. Coil shapes have played a profound role in the performance of wireless power transfer, especially with the coupling coefficient, output power, and energy transmission. Spiral coils can provide a better coupling coefficient over various distances between the transmitter and receiver coil. Therefore, it is important to design a reasonable coil shapes to improve the transfer performance of the system. Achieving optimum coil geometry, which maximizes the transfer quality factor of WPT system, requires proper selection of optimization method. A direct search method called the generalized pattern search algorithm (GPS) was used to optimize the mutual inductance between the transmitter and the receiver coil. A GPS method was used to optimize the coil, according to the developed objective function. The procedure relies on coupling a GPS algorithm to a mathematical model to perform an automated optimization design of the WPT system. The results show the optimum geometry designed for the WPT system operating at 85 kHz, with transfer quality factor of 60.42, can achieve maximum efficiency of 96.22 %.

Accurate calculation of temperature field of outer rotor PMSM is particularly important, as it determines the overall performance and operational stability. Due to complex structure of the winding part in stator slot, the temperature field encounters huge challenges. For this reason, this paper conducts research on the winding part in stator slot. Firstly, in order to reduce difficulty of modeling, based on the principle of constant thermal resistance, the windings in motor slot are equivalent to large conductors, which improves the calculation accuracy and shortens the calculation time at the same time. Secondly, based on finite element simulation calculation, the feasibility of the models before and after the equivalence is verified, and the influence rules of factors such as heat flux density and boundary conditions on temperature field distribution of windings in the slot are analyzed. Finally, a 24-slot and 20-pole starter motor model was taken as research object, and the temperature field experimental tests under different working conditions were carried out to verify the effectiveness of the method.

Predicting power losses due to eddy currents in electrical conductors is an important aspect of electrical engineering research. These losses can reduce device efficiency and even

cause damage in the case of excessive overheating. The issue becomes more serious when dealing with high-speed electrical machines that operate at extremely high frequencies. Permanent magnet machines designed for transportation and electric propulsion account for the majority of this category of devices. Due to the high thermal sensitivity of rare earth permanent magnets, it is critical to develop reliable and robust models for predicting eddy power loss and thus preventing per-manent magnet thermal demagnetization. This work is concerned with this research issue, and a hybrid model for calculating losses in permanent magnet machines was established. It is based on combining the numerical computation of the magnetic field using the finite element approach with another analytical method for computing losses due to induced currents in permanent magnets. The model is applied for a permanent magnet synchronous machine

In order to suppress the vibration and noise of the surface mounted permanent magnet synchronous motor (SPMSM), an analytical calculation model of electromagnetic force based on the optimized magnetic pole shape is proposed and verified by the finite element method. Secondly, the coupling simulation analysis of magnetic vibration and noise of SPMSM with magnetic pole optimization is carried out, and the magnetic pole optimization scheme with minimum vibration and noise value of the motor is obtained. Finally, the noise and sound pressure level of the motor during operation is obtained through vibration and noise experiment, which verifies that the optimized motor can effectively reduce the vibration and noise of the motor under the premise of ensuring the electromagnetic performance.

The issue of low demagnetization fault diagnosis accuracy and the over-reliance on prior knowledge and human expertise in diagnostic methods for permanent magnet synchronous motor(PMSM) used in electric drive systems., this paper adopts a two-seat permanent magnet synchronous motor(PMSM) for 22-pole 24-slot electric aircraft as an example to study the dynamic identification method for demagnetization faults. The finite element method is used to simulate the motor's electromagnetic field and calculate its no-load back electromotive force, and an advantage in feature learning by combining residual network (ResNet) and convolutional neural network (CNN) is proposed. A multi-attention fusion residual convolutional neural network (MAFResCNN) with a squeeze-excitation module (SEM) and a convolution block attention module (CBAM) is constructed to classify the motor's no-load back electromotive force using wavelet transform (WPT). Experiments show that the method can not only intelligently detect the motor's demagnetization fault but also estimate the degree of demagnetization with high accuracy, with judgment error within an acceptable range.

Magnetic target localization is widely used in biological magnetic navigation, vehicle tracking, and buried object detection. Mathematical models and error analyses can help improve the positioning accuracy of magnetic targets. Existing research has overlooked the influence of the position and posture of a magnetic target, and lacks coupling analysis between multiple factors. Considering the complete position and posture parameters, this paper establishes a more comprehensive multi-factor mathematical model. Using this model, we identified the inaccurate area of the positioning algorithm and the compensation direction for magnetic field measurement errors, clarifying the impact of various influencing factors on positioning accuracy. The positioning experiment inside the magnetic shielding room (MSR) showed that the deviation between the simulation results and the experimental results did not exceed 2.9%, proving the accuracy of the mathematical model and error analysis.

With wider and wider application of wireless power transfer (WPT) devices, self-resonant ones are attracting more and more attention thanks to their merits such as high-power densities. However, the simulation of self-resonant type WPT devices is facing new challenges like taking into account the stray capacitance, the appearance of different types of conductors, the need to simultaneously consider the electric field in conductors and insulating objects and the magnetic fields in the whole domain. For the self-resonant WPT devices, typically they are usually composed of solid conductor layers, stranded Litz coil, dielectric substrates and possibly a magnetic core. As a result, one need to consider the complicated electromagnetic fields resultant from combined effect of full electromagnetic material parameters and different types of conductors. To this end, in this work a benchmark testcase is proposed for checking the numerical result by finite element analysis. Different settings by changing the material parameters will be carried out and measurement will be also implemented to rigorously check the accuracy of the numerical scheme regarding the nodal field values.

The electromagnetic induction phenomenon is ubiquitous and engineers are willing to exploit it for many purposes, ranging from contactless power or information transfer to energy harvesting. Simple transformer characteristics, like tranformation ratio or leakage fluxes, become elusive whenever the intended inductive device departs significantly from a conventional electric transformer with well-defined wound coils and closed magnetic cores. The purpose of this paper is to show that the domain of validity of the T-shaped equivalent circuit is actually rather large, to provide practical applicability conditions for it, and to review its essential and irreducible limitations.

The operator learning method (OLM) can directly predict the results of numerical analysis problems without going through the tedious procedure of traditional numerical methods. However, the studies on OLM so far have only focused on improving the accuracy and speed of single-domain problems. This paper proposes a multidomain OLM to solve electromagnetic field problems for practical engineering applications. The direct boundary integral equation is initially employed to solve the value and its gradient of the boundary and interface. Then, the multidomain problem is transformed into multiple single-domain problems. The OLM is employed to solve the problem of each single domain. As an example, the proposed method is used to solve the 2D electrostatic box and multi-domain electrostatic box problems, and the results are compared to the boundary element method. Finally, the proposed method is employed to solve TEAM Problem 22, and the result is compared with ANSYS/Maxwell.

During the optimal design of a large size permanent magnet (PM) electrical machine, its electromagnetic performance and the PM anti-demagnetization capability should be iteratively analyzed with FEM. If the PM has been processed with grain boundary diffusion (GBD), its magnetic property varies inside it. This will lead to high modeling complexity and computation burden. Therefore, it is proposed to model the PMs as a homogeneous equivalent material when calculating the machine overall performance, avoiding excessive detailed modeling of the PM within the large machine model. Then, the magnetic field intensity in the homogeneous PM is extracted, and applied in the actual GBD PM to analyze its demagnetization situation. The GBD PM has a much smaller size than the whole machine, therefore can be re-modeled as a nonuniform material for FEA. Thereby, both computational efficiency and accuracy can be ensured.