Conference Agenda

Overview and details of the sessions of this conference. Please select a date or location to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).

Session Overview
PD-M4: Optimization and Design
Friday, 19/Jul/2019:
10:50am - 12:40pm

Session Chair: Kazuhiro Muramatsu
Session Chair: Gang Lei
Location: Patio 44-55

Show help for 'Increase or decrease the abstract text size'

Optimal Design of Field Winding of Dual-excited Synchronous Condenser

Guorui Xu, Chaolong Luo, Yang Zhan, Ran Liu, Aamir Nazir

North China Electric Power University, China, People's Republic of

The Dual-Excited Synchronous Condenser (DESC) can enhance the ability of absorbing reactive power. A DESC which has two sets of asymmetrical field windings on the rotor is proposed based on a Traditional Synchronous Generator (TSG). The new field winding structure may change the distribution of the magnetic field and the saturation of the rotor core. However, the new rotor structure is expected to give better electromagnetic performances. This paper proposes several rotor structures with different number of slots and slot pitch of q-axis field winding. The air-gap flux density, magnetic field distribution and saturation under different rotor structures are compared. A rotor structure which can enhance fundamental flux density and reduce harmanics is obtained. The results are verified by a 10-kW DESC model machine. This study provides the theoretical basis for the design and manufacturing of large DESCs.

On Efficient Use of Deep Learning for Acceleration of Topology Optimization for Rotating Machine

Shuhei Doi, Hajime Igarashi

Hakkaido University, Japan

This paper discusses the effective use of deep learning for acceleration of topology optimization, which is applied to an electric motor. The classifiers with respect to the torque characteristics are constructed through training of a convolutional neural network (CNN). Then minimization of the iron loss considering the constraint for the torque characteristics is performed with aid of the trained CNN. The threshold for execution of FE analysis posterior to the CNN evaluation is shown to give a significant effect on the performance.

Optimized Design, Analysis of High Gear Ratio Magnetic Gear for High Speed and High Power

Se-Yeong Kim1,2, Do-Kwan Hong1,2, Tae-Woo Lee1,2

1University of Science & Technology, Republic of (South Korea); 2Korea Electrotechnology Research Institute, Republic of (South Korea)

Magnetic gears are non-contact, which has advantages such as reduction of maintenance cost, improved reliability, overload protection, noise and vibration reduction. Therefore, we optimized high speed, high power magnetic gear which can replace mechanical gear. This optimized magnetic gear model can be applied to various fields. This model is 81kW, 60krpm, and gear ratio is 10:1. The outer diameter was determined considering the torque, and the stack length was adjusted to reduce the outer diameter. Also, to reduce the torque ripple, we changed the shape of inner permanent magnet. The optimization model was proposed using orthogonal array and the response surface methodology, and was selected considering torque ripple and power density. In the loss analysis of this optimization, permanent magnets were divided in the axial direction to reduce eddy current loss. And structure and vibration of the optimization model based on 3D magnetic analysis were verified.

Application of Opposition Teaching Learning Based Optimization for 5G Communication Systems Antenna Design

Achilles Boursianis1, Sotirios Goudos1, Traianos Yioultsis2, Katherine Siakavara1

1School of Physics, Aristotle University of Thessaloniki, Thessaloniki, Greece; 2School of Electrical and Computer Engineering, Aristotle University of Thessaloniki, Thessaloniki, Greece

To address the complex optimization problem of antenna design for 5G wireless communication systems, we enhance the Teaching Learning Based Optimization (TLBO) algorithm with Opposition Based Learning (OBL). In this paper, we obtain new TLBO algorithms with different OBL variants. The different opposition schemes are evaluated in order to find the variant with the best performance. Numerical simulation results show that the current optimum opposition scheme outperforms the other algorithms and achieves the highest average ranking. Moreover, the COO-TLBO variant is applied in the design of aperture-coupled E-shaped patch antenna. Preliminary results demonstrate very satisfactory values of return loss in the desired frequencies of 5G wireless communication systems.

Multi-objective Optimization of Benchmark TEAM Problem using Gradient based Approach and Adaptive Weight Determination

Namhee Ryu, Minsik Seo, Seungjae Min

Hanyang University, Korea, Republic of (South Korea)

This paper proposes new approach to find Pareto optimum solutions of benchmark TEAM problem by using gradient based optimization algorithm and adaptive weight determination scheme. Air-cored multi-turn winding problem that has conflicting objective functions are considered, and the solutions are obtained by using the sensitivity information of the objective functions. With the adaptive weight determination scheme, the values of different weights are iteratively updated during the optimization process, and evenly distributed solutions are gradually found in the objective space. To evaluate the effectiveness of the proposed approach, obtained solutions are compared quantitatively with the solutions of non-dominated sorting genetic algorithm (NSGA-II), and it is confirmed that more diverse and uniform solutions could be obtained with the proposed approach.

Topology Optimal Design of Near-Field Plate Applied in Wireless Power Transmission System

Xiu Zhang1, Honghao Sun1, Xin Zhang1, Zhihan Zhang1, Weinong Fu2, S. X. Niu2

1Tianjin Key Laboratory of Wireless Mobile Communication and Wireless Power Transmission,Tianjin Normal University, China, People's Republic of; 2Department of Electrical Engineering, The Hong Kong Polytechnic University

Wireless power transmission system has attracted many researchers’ attention from all over the world. However, the contradiction between the power transmission efficiency and the power transmission distance limits the application of the wireless power transmission system. In this paper, in order to solve the problem, a novel near-field plate is designed and applied in the wireless power transmission system. The simulation results indicate that the near-field plate can focus electromagnetic wave to improve the performance of wireless power transmission system. In order to obtain the optimal result, the topology optimization method is applied to design and analyze the near-field plate. After analysis, the optimal structure of the near-field plate is obtained and the power transmission efficiency of the wireless power transmission system achieves maximum value.

Design and Analysis of a Three-Degree-of-Freedom Linear Oscillatory Actuator

Akira Heya, Katsuhiro Hirata, Takahiro Matsushita, Yoshinori Kono

Osaka Uuiversity, Japan

Oscillatory actuators are applied for various applications such as entertainment systems, consumer products, etc. In addition, the oscillatory motion is expanded to multiple degrees of freedom. A three-degree-of-freedom linear oscillatory actuator (3-DOF-LOA) has been developed. Our previous 3-DOF-LOA needs five sets of coils for a three-degree-of-freedom motion, and forces on each axis are interfered with each other. In this paper, we propose a novel 3-DOF-LOA with 3 sets of coils, and its forces on each axis are not interfered with each other. The magnetic structure and operating principle of the proposed 3-DOF-LOA are described. The force characteristics are calculated using a magnetic field analysis, and the feasibility of the proposed structure is discussed.

Optimization of a Linear Flux-switching Permanent Magnet Machine in Rail Transportation using Pareto Set Persuading Algorithm

Keyu Guo1,2, Ke Wang1, Qiongxuan Ge1, Yaohua Li1, Peilong Wang1,2, Huihuang Wang1,2

1Institute of Electrical Engineering, Chinese Academy of Sciences, Beijing 100190 China; 2School of Electronic Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing, 100049

In this paper, a flux-switching linear permanent magnet machines (FSLPMM) with 12 slots and 11 poles (12s11p) used in urban rail transportation is researched. The structure of FSLPMM is firstly depicted and parameterized. Then a multi-objectives optimization algorithm called Pareto-set persuading algorithm (PSP) is used to optimize the performances of FSLPMM. In the optimization progress, the normal force and thrust force of FSLPMM are set to be two objectives, while the power factor and efficiency are set to be non-linear constraints. Finally, the effectiveness of PSP on solving optimization problem of FSLPMM is validated by comparing the optimization results of PSP with that of non-dominated genetic algorithm-II (NSGA-II).

Sensitivity Analysis Based Optimization of General Magnetic Gear Patterns

Yuan Mao, Shuangxia Niu

The Hong Kong Polytechnic University, Hong Kong S.A.R. (China)

Magnetic gear (MG) has many advantages compared to mechanical gear in terms of torque density, power efficiency and stability. In order to improve the working efficiency of the magnetic gear, there is no stoppage for the development of its topology in recent years’ research. This paper proposed an optimization method of a general topology pattern of MGs. This general pattern can cover all possible type of coaxial surface mounted MGs. In this general pattern, the inner layer and outer layer can be unipolar permanent magnetic-ferrite array or air-ferrite array or bipolar permanent magnetic array. The modulation ring can be permanent magnetic-ferrite array or air-ferrite array. Taguchi method combined with Analysis of variance (ANOVA) is introduced to evaluate the influence of each parameter in MGs. The generic algorithm (GA) optimization method is adopted in the study.

Continuum Sensitivity-Based Optimization Method for IPMSM with Shape Constraint of Permanent Magnet

Seung Geon Hong, Il Han Park

Sungkyunkwan University, Korea, Republic of (South Korea)

This study proposes a continuum sensitivity-based design optimization method to obtain optimal design of interior permanent magnet synchronous motor. Compared to finite difference method, the continuum sensitivity analysis significantly shortens the computational time when design variables are numerous. We combine the continuum sensitivity with the shape constraint so that both shape and size can be designed simultaneously in shorter time. To demonstrate the feasibility of this method, design optimization was tested to V-shaped IPMSM rotor design. In addition to IPMSM design, the proposed method is expected to be effective in various design problems that require shape and size to be designed simultaneously.

Stepwise Optimal Design Strategy of an Interior Permanent Magnet Synchronous Motor for a Fuel Cell Electric Vehicle

Jae-Gil Lee1, Dong-Kuk Lim2, Hyun-Kyo Jung1

1Seoul National University, Korea, Republic of (South Korea); 2Ulsan University, Korea, Republic of (South Korea)

This study presents a stepwise optimal design strategy (SODS) which effectively utilizes the magnetic equivalent circuit (MEC) and the finite element method (FEM), and an optimization algorithm for the design of an interior permanent magnet synchronous motor for fuel cell electric vehicles (FCEV). The proposed strategy respectively uses the MEC, the FEM, and the optimization algorithm by dividing the entire design in to the initial, detailed, and the optimal design stage. The proposed design strategy can be widely used as a design technique of electric machines since it can not only reduce computational cost but also accurately consider the diverse design cases.

Optimize Least Square SVM Datasets for Eddy Current NDT Sensors

Mohamed Chelabi1, Tarik Hacib1, Yann Le Bihan2

1Laboratoire d'électronique et d’électrotechnique industrielle (L2EI), Faculté des sciences et de la technologie, Univ. Jijel, Algérie; 2Laboratoire GeePs, Génie électrique et électronique de Paris Supélec, UMR 8507 CNRS, Univ. Paris 06, Univ. Paris-Sud, France

This paper presents the optimization of data base that used by SVM to the resolution of inverse problems of electromagnetic non-destructive testing (NDT). The goal is the defect size characterization in a conductive nonmagnetic plate from the impedance variations measurement, by the novel combination of the Least Square Support Vector Machines (LS-SVM) and Finite Element Method (FEM). The LS-SVM is a statistical learning method that has good generalization capability and learning performance. COMSOL multi-physics based on FEM is used to create the datasets required to train the LS-SVM. A good agreement is obtained between the numerical results and the experimental ones. Teaching Learning Based Optimization (TLBO) is proposed for LS-SVM parameter tuning.

Development of Novel Dielectric Cover for Millimeter-wave Band Waveguide Slot Array Antenna

Keiichi Itoh1, Kazuma Takita1, Masaya Kumata1, Hideaki Matsuda1, Masaki Tanaka1, Hajime Igarashi2

1National Institute of Technology, Akita College; 2Hokkaido University

We propose a novel dielectric cover structure for the 76 GHz millimeter-wave band waveguide slot array antenna. The proposed cover is consists of the dielectric hemisphere and the dielectric layer. The convergence effect of the hemisphere is expected to improve the gain of the antenna. The purpose of this study is to develop the proposed dielectric cover using FDTD (Finite Difference Time Domain) method. By finding the appropriate design parameters, the antenna gain with the proposed cover increases by about 1 dB as compared with that without the cover. From the E-plane far field radiation patterns, the convergence effect of the proposed cover is confirmed.

Asymmetric Magnetic Field Generation for Spherical Motors

Ho-Joon Lee1, Jun-Hui Won2, Seung-Taek Oh2, Ju Lee2

1BIST(Busan Institute of Science & Technology), Korea, Republic of (South Korea); 2Hanyang University, Seoul 04763, Republic of Korea

This study analyzes asymmetric magnetic field generation for a unique surface permanent magnet (SPM) spherical motor with a rotating stator field and an oriented permanent magnetic field. Accordingly, the mismatched harmonics of the two fields cause torque ripple. Torque characteristics of the spherical motor are analyzed for each state of rotation and positioning angle. Finally, we compare experimental data to 3-dimensional finite element method simulations. The proposed SPM spherical motor realizes three degrees of operational freedom and is expected to be useful to many robotic applications.

Development of Encoder-Decoder Predicting Search Process of Level-set Method in Magnetic Circuit Design

Ryota Kawamata1, Shinji Wakao1, Noboru Murata1, Yoshifumi Okamoto2

1Waseda University, Japan; 2Hosei University

Finite Element analysis (FEA) of magnetic field generally requires a lot of calculation time. Especially, Design optimization method such as the level-set method with FEA results in enormous calculation time to find more global solution. In this paper, we propose a novel method of precisely reproducing the conventional optimization steps by means of Convolutional Neural Network (CNN) and Long Short-term Memory (LSTM). The developed method enables us to implement high speed search of solution, which means the possibility of effectively optimization with various initial shapes and conditions. Finally, we evaluate calculation time and computational accuracy of the proposed method by using a magnetic circuit design model.