Conference Agenda

This paper proposes a general pattern of spoke-type interior permanent magnet (IPM) machines. The general pattern can represent different rotor structures and can be applied to automatically determine optimal rotor structures. Various rotor structures, including symmetrical and asymmetric rotor core topologies, can be obtained by changing the design parameters of the general pattern. A non-dominated sorting genetic algorithm II (NSGAII) is adopted in the optimal design to search for optimal rotor configurations automatically. With the aid of the optimization program, an asymmetric spoke-type rotor structure with better performance is obtained. The analysis and comparison are conducted between a conventional machine and a proposed asymmetric machine. The finite-element simulation results show that the optimal design performs 7.7% higher output torque due to the magnetic-field-shifting (MFS) effect while the total permanent magnet (PM) volume remains the same. The torque ripple of the proposed structure is also reduced by 82.1% simultaneously.

A novel general pattern of permanent magnet (PM) arrangement for the optimal design of magnetic gears (MGs) is presented in this work. The proposed general pattern can generate five common PM arrangements and a novel PM arrangement with a combination of Halbach array and spoke-type array. The common PM arrangements include surface-mounted, consequent pole, Halbach array, simplified Halbach array, and spoke-type. By using this optimization pattern, the PM arrangement with high torque and low torque ripple in MGs can be obtained within a short computing time. To exhibit the merit of this optimization method, this general pattern is applied to a reluctance magnetic gear (RMG), which is a novel type of MGs suitable for high-speed power transmission systems. The transmission torque and torque ripple of the RMG model are derived by using finite element analysis (FEA) of magnetic field distribution. Meanwhile, the design optimization of this model is conducted by using a genetic algorithm (GA) method. The merit of this general pattern is revealed by the comparison of the proposed RMG and a conventional surface-mounted RMG model. The proposed RMG in this study can transmit 138% higher torque than the conventional ones. The PM volumetric torque density also increases by 18.89%. Moreover, the torque ripples of both high-speed and low-speed rotors are less than 1.5%, which is considerably low in RMGs.

As the electromagnetic devices developing towards systmatization and multi-function, optimization designs of electromagnetic device are often involved with large scale of variables. An improved fireworks algorithm is proposed to solve the electromagnetic optimization problems with large scale of variables. To locate the promising area in the vast variable space, a gradient-based spark generation rule is developed in the generation of new sparks; and an exposion magnitude control strategy is proposed to blance the exploration and exploitation. Mathematical function with 30 variables and a linear antenna array with 24-elements are used to testify the effectiveness and efficiency of the proposed method. The results demonstrate that the improved fireworks algorithm can find the global optimal solution with less iteration compared with the original method.

The weight of magnets limits the technological progress and wider applications of the cyclotron. A high temperature superconducting (HTS) ironless proton cyclotron has been developed in China Institute of Atomic Energy (CIAE). In this paper, a fast vectorized integral equation method was applied to calculate the main magnetic field in the median plane of the cyclotron. With little computational cost, the primary layout of the coils can be obtained based on the genetic algorithm (GA). To minimize the phase error, the function between the currents of the trim coil system and phase error was also established. Therefore, it is feasible to use the linear programming (LP) method to complete the production of the main magnetic field which satisfies isochronous acceleration requirements and avoids resonance crossing.

To address the inefficiencies of exiting multi-objective robust optimization methodologies in application to metamaterial designs, an improved multi-objective GA is proposed. To accelerate the solution speed of the original GA in finding both high quality solutions and distributing them uniformly, two polynomial approximation-based move operations are introduced. Moreover, some dominant techniques including the construction of the relationship between different objective functions, and the relationship between the objectives and the design variables are also investigated. Also, an adaptive response surface model is presented to efficiently quantify the robust performance of a solution. The numerical results have demonstrated the feasibility and merits of the proposed methodology.

The technology of Unmanned Aerial Vehicle (UAV) has reached in the last years a maturity which allows their use in a number of relevant civil missions. The reduction of the requisites for take-off is essential to extend the operating scenario of UAVs. In this digest the authors describe the design of an electromagnetic launching system for a light weigh UAV based on a Double Sided Linear Induction Motor (DSLIM).

The length of the launcher and the average thrust force are assigned. The design, based on a multistage configuration, is performed in two steps. In the first one, a semi-analytical model of the linear motor is established and the choice of the most relevant parameters is performed. The design is subsequently refined by the use of a more accurate model based on the Finite Element Method (FEM).

The tubular linear oscillating machine can achieve reciprocating movement directly without auxiliary transmission mechanisms, which is widely used in the Free-piston energy conversion system. In order to achieve lower thrust ripple and eliminate eddy current effect, circumferential silicon steel stack is usually utilized in the stator as universal design. However, it results in assembling difficulties. In this paper, a radially-split lamination tubular linear permanent magnet oscillating machine with auxiliary configuration (RSL-TLPMOMA) is investigated to achieve low thrust ripple and assembly convenience simultaneously. Firstly, the laminated coefficient expression of split-laminated is deduced, the effect of laminated number and radius ratio on laminated coefficient is discussed, the optimal combination of the laminated number under different radius ratios are given. Secondly, a subdomain conversion method is used to predict the magnetic field in the linear machine with consideration of the auxiliary-teeth and end effects. Subsequently, the electromagnetic performance under no-load condition is predicted and the finite element method is utilizing to validation. Finally, an RSL-TLPMOMA prototype is manufactured and test. As a whole, the analytical process offers universal guidance in the initial design for a linear oscillating machine.

This paper presents a new topology optimization method for IPM motors, considering current conditions using a convolutional neural network (CNN) to reduce optimization time. The multitask CNN estimates the multiple torque performances from the current conditions and cross-sectional material distribution, and applies it to the topology optimization process. As a result, the average torque and harmonic components of the target motor improve by 5.6% and 68%, respectively, under multiple current conditions. Furthermore, the optimization time is reduced by 74% using the proposed method, as compared to the conventional methods.

This digest presents a new optimal design methodology for power transformers. The approach employs the manufacturer's design procedure which is defined and solved as a multiobjective problem. The nonsorting genetic algorithm NSGA-III is used to find the optimal solutions. Afterward, a finite-element (FE) electromagnetic model is constructed using the geometry and data corresponding to the optimal solution. Next, the FE transformer model is used with the design of experiments to generate a surrogate model. Finally, another multiobjective optimization is carried out using the surrogate model. Numerical results of the proposed methodology are given.

The spoke type permanent magnet synchronous generator (PMSG) has difficulty in magnetizing compared to other types of permanent magnet synchronous generators due to its structural characteristics. Magnetization performance is an important performance directly related to demagnetization and mass productivity of permanent magnets. Therefore, this paper proposes a spoke type PMSG rotor shape design considering magnetization. In addition, the barrier was designed asymmetrically to reduce cogging torque, which is one of the important performance specifications, and the performance of the spoke-type PMSG model proposed to improve magnetization and cogging torque was evaluated through finite element analysis (FEM).

Recently, the increase of microplastics (MPs) in ocean have been recognized as one of ocean pollution listed in SDGs. MPs are fragments of plastics with a diameter smaller than 5 mm. Since MPs absorb harmful chemicals while floating in ocean, small fish which eats the absorbed MPs may fall into bad digestion and poor health condition. To solve the ocean pollution problem, some MP collection methods have been proposed, such as collection by filter, buoyancy, ultrasonic, and low magnetic field. However, none of them are suitable for MPs collection from ocean due to low processing ability. While, recently, the performances of high-temperature superconducting (HTS) magnets have been improving. HTS magnets which can generate a field higher than 30 T are available relatively easily. It is expected that MP collection performances are improved by such high field superconducting magnets.

In this paper, we propose a conceptual design for MP collector from seawater using high magnetic field. To evaluate the performances of the device, we developed a fluid simulation coupled with magnetic field analysis. The simulation results indicate high magnetic field improves the performances. Finally, an ideal superconducting magnet is optimally designed with a Genetic algorithm.

This paper proposes a fast topology optimization based on the deep neural network (DNN) which predicts the current-dependent motor torque characteristics from its cross-sectional image. The trained DNN is shown to provide the current condition which gives

the maximum torque under an assumed motor control method. Using the proposed method, topology optimization is successfully performed with reduced number of field computations.

Traditional heat transfer fluids, such as water or oil, usually show low thermal conductivity. Suitably engineered nanofluids have been considered to potentially circumvent this issue. Magnetic nanofluids (MNFs) are a new class of such improved performance heat transfer fluids. MNFs exhibit both the fluid and magnetic properties, and consequently the heat-transfer rate of MNFs can be enhanced and controlled by external magnetic field, thus, MNFs have found several applications in heat-transfer processes and attracted much attention in recent years. The performance of heat exchangers based on MNFs can be optimized by suitably designing a field map aimed at improving their thermal properties. In this study, using a multiphysics finite element model, combining Maxwell equations with the equations for the heat transfer in fluids in a pipe, an optimization of magnetic field source is aimed, to enhance the capabilities of the heat exchanger.

The 2-pole 6-slot permanent magnet synchronous motor (PMSM) is an attractive option for ultra-high-speed (UHS) applications due to their high manufacturability and balanced magnetic force. UHS PMSM should be designed with short stack length and end coil height due to rotor dynamics issue. Also, due to its high frequency, it must be designed considering stator iron loss, rotor eddy current loss, and AC copper loss. With the same pole-slot combination, the performance varies greatly depending on the coil pitch, so this paper compared the size and efficiency of the 2-pole 6-slot UHS PMSM according to the coil pitch. For reasonable comparison, an optimal design was performed for each coil pitch, and to reduce computation time, the optimization was conducted in two stages: size minimization and efficiency maximization. It was confirmed that in terms of total axial length, 1-coil pitch was the shortest due to the advantage of short end coil height, even with the lowest winding factor. On the other hand, 2-coil pitch has good trade-off relationship in terms of efficiency and size.

We present an optimal design of a surface-mounted permanent-magnet motor with a duct structure for a turbo compressor using a novel analytical computation method. Furthermore, to increase the power density and the cooling effect of the system, we employ a duct with a rotor overhang structure. Although the design enhances the performance of the turbo compressor, multiple 3D finite element analyses (FEAs), which are complicated and time-consuming, are necessary during the design stage to determine the overhang and duct lengths. Alternatively, analytical models may be used. Conventional conformal mapping facilitates 2D electromagnetic field analysis; in contrast, 3D electromagnetic field analysis can be performed using the proposed analytical method referred to as z-axis double conformal mapping. The optimal design of the turbo compressor is achieved by combining the proposed analytical method and optimization algorithm. The accuracy and computation speed of the proposed analytical method are verified via comparison with 3D FEA results.

This paper proposed a novel surface-mounted permanent magnet (SPM) machine. The permanent magnet (PM) of the machine is divided into two sections. There is a magnetic pole with a large pole-arc coefficient in each segment to reduce the cogging torque and tooth-harmonics EMF of the SPM machine. Viewed from the axial direction, the magnetic poles of the larger pole-arc coefficients on the two segments are symmetrically distributed along the center of the machine, which effectively reduces the radial magnetic pull force caused by unequal pole-arc coefficients. A comparative study of the proposed machine and other SPM machines (conventional machine, machine with skewing slots, machine with segmented pole) was carried out. The research results show that the proposed machine has smoother torque performance, while the problem of radial force imbalance introduced by the asymmetric rotor structure is overcome.

This paper is a comparative analysis study on topologies of magnetic geared motors for the application of railway vehicle traction systems. The magnetic geared motor is a system in which a magnetic gear and a permanent magnet synchronous motor are mechanically combined, and it consists of a dual rotor that rotates at different speeds. Due to these characteristics, it has the advantages of high-power density and lightweight compared to the existing traction system. However, magnetic geared motors have completely different characteristics and power split depending on the topology, and it is important to select an appropriate topology depending on the application. Recently, some studies that reflect this on industrial or electric vehicles have been conducted, but there is no research on this for railway vehicles. Therefore, in this paper, the suitability of each topology is analyzed to derive the final model considering the design specifications of application for an urban railway vehicle traction system.

In the electromagnetic design stage of PM motors, the consideration of reducing torque ripple is as important as the improvement of motor efficiency. In this paper, the electromagnetic performance of the SPM motor with non-uniformly magnet spacing is analyzed and calculated in detail. The Maxwell tensor method is used to explain the mechanism of mitigating cogging torque, and it will obtain the optimal non-uniformly magnet spacing more quickly than finite element parameterization. According to the change rule of winding flux linkage, the winding factor of the proposed motor are calculated, which describe the principle of harmonic generation fundamentally. The magnetic field is affected by many factors and the demagnetization curve of motor materials are nonlinear, so a multi-objective optimization strategy based on the genetic algorithm and the finite element calculation is selected to acquire the best parameters.