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
Session
PC-A3: Optimization and Design
Time:
Thursday, 18/Jul/2019:
2:20pm - 4:10pm

Session Chair: Luiz Lebensztajn
Session Chair: Shuhua Fang
Location: Patio 44-55

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

Focus on the Concept of Imaginary Parametric Optimization Curve Applied to the Preliminary Design of Induction Motors

Frederic Wurtz, Nicolas Moisson-Franckhauser, Lauric Garbuio

Univ. Grenoble Alpes, CNRS, Grenoble INP

This paper outlines the importance of the early steps of the design of engineering systems. To address it, we propose “Imaginary Machine Design” (IMD) as a global design approach. The fundamental principle of IMD is a reformulation of the initial design problem, initially formulated as a discrete optimization problem in a real design space, toward a continuous and differentiable optimization problem in a design space of “Imaginary Machines”. We will more especially focus on the concept of “Imaginary Parameterized optimization Curve” applied to the Preliminary Design of Induction Motors. We will show how it helps to understand the shape of the searching space, during the “problem setting phase”, before searching real discrete solutions in “the problem solving phase”



Simple Analytical Model for Parameter Estimation of Permanent Magnet Synchronous Motor

Min-Ro Park1, Hyeon-Jin Park1, Hae-Joong Kim2, Ho-Young Lee3, Myung-Seop Lim4, Jung-Pyo Hong1

1Department of Automotive Engineering, Hanyang University, Seoul 133-791, Korea; 2Rotating Machinery Center, Korea Testing Certification, Gunpo-si 435-823, Korea; 3Technology Convergence Group, Korea Institute of Industrial Technology, Daegu 711-883, Korea; 4School of Mechanical Engineering, Yeungnam University, Gyeongbuk 38541, Korea

This paper proposes an analytical model for the parameter estimation of the permanent magnet synchronous motor (PMSM) considering the magnetic nonlinearity. The proposed analytical model consists of a simple equivalent magnetic circuit (EMC) and equations. At this time, the simple EMC compensates for the limitation of the equation, with which is difficult to consider the magnetic nonlinearity. From the proposed EMC, the stator reluctance is calculated, which reflects nonlinearity for parameter calculation. Then the parameters are calculated mathematically, reflecting the calculated stator reluctance. Finally, the proposed analytical model is verified via finite element analysis (FEA). As a result, the maximum errors of the back electromotive force and the inductance are smaller than 2 and 5%, respectively.



A Multi-objective Topology Optimization Methodology and its Application to Electromagnetic Actuator Designs

Yilun Li1, Shiyou Yang1, Zhuoxiang Ren2

1Zhejiang University, China, People's Republic of; 2Sorbonne University, France

An engineering topology optimization (TO) problem is generally formulated as a multi-objective optimization (MOO) one. However, most of the exisitng methods for multi-objective TOs transform the multiobjcetives into a single-objective one. In this paper, a multi-objective topology optimization (MOTO) methodology based on a hybrid MOO algorithm to integrate the Non-dominated Sorting Genetic Algorithm II (NSGAII) and Differential Evolutionary (DE) algorithm is proposed. The framework of the proposed hybrid MOO algorithm is elaborated, and its performance is evaluated by solving typical test functions. To validate the proposed MOTO methodology, it is applied to the topology optimization of an electromagnetic actuator. The numerical results have demonstrated that a set of novel topologies with multiple objectives improemnt can be obtained.



A Multi-Objective Topology Optimization Methodology based on Pareto Optimal Min-Cut

Meng Xia1, Shiyou Yang1, Jan Sykulski2

1Zhejiang University, China, People's Republic of; 2University of Southampton, U.K

One bottleneck problem in multi-objective topology optimizations which is still open to address is the overwhelming heavy computational cost coming from the finite element analysis embedded in fitness computations in the iterative procedure. To solve this problem, a novel methodology based on Pareto optimal min-cut for multi-graph is firstly proposed. Through searching for the Pareto optimal graph cut in each iteration, the proposed method finds a Pareto optimal solution to the TO problem in the current iteration. According to the numerical results as reported, the proposed method is able to achieve a good trade-off between an acceptable computational cost and a promising performance parameter.



Impact of the Machine Model on the Optimal Design of Induction Machines

Aswin Balasubramanian, Osaruyi Osemwinyen, Floran Martin, Anouar Belahcen

Aalto University, Finland

Machine model plays an important role in determining the optimal design of an induction machine. This paper analyses the optimization problem of an induction motor with an iterative semi-analytical machine model using particle swarm optimization technique. The optimized machine is then simulated by a finite element model for further investigation of the machine performance.



Design and Analysis of an improved Large Single Sheet Tester measurement system under stress

Yu Dou1, Yongjian Li1, Shuaichao Yue1, Jianguo Zhu2

1Hebei University of Technology, China, People's Republic of China; 2School of Electrical and Information Engineering, University of Sydney, Sydney

In electromagnetic devices, mechanical stress may influence the magnetic properties of the core materials. The performance of the devices should be therefore decreased by the stress effect. In order to systematically analyze the effect of stress on the magnetic properties, an improved large single sheet tester (LSST) measurement system under stress has been designed and performed. The cross specimen with multi-slots and sensing method are optimized to eliminate the additional eddy current losses in the LSST method and improve sensing accuracy. The same experiments are also performed on Epstein frame to verify the accuracy of the magnetic properties measurement system. Also, the magnetic properties under uniaxial and biaxial stresses of non-oriented silicon steels 50JN470 are systematically discussed in alternating magnetization.



A Topology Optimization Method Based on Pseudo Gray Image Segmentation Using Level Set Method

Meng Xia, Shiyou Yang

Zhejiang University, China, People's Republic of

The level set method for topology optimization often suffers from a poor convergence rate and is heavily dependent on geometry mapping. To overcome these inefficiencies, a new method by transforming the sensitivity information into a pseudo gray image and then segmenting the corresponding image using the level set method is proposed for topology optimization. The proposed method can readily control the convergence speed by adjusting the contrast of the pseudo gray image. According to the numerical results as reported, the proposed methodology is able to significantly enhance the solution speed without any compromise on the solution quality.



Branch and Bound Algorithm based on Meta-model with Error Prediction for Computational Electromagnetics

Reda El Bechari1,2, Stéphane Brisset1, Stéphane Clénet1, Frédéric Guyomarch1, Jean-Claude Mipo2

1L2EP, France; 2Valeo EEM, France

Meta-models proved to be a very efficient strategy for optimization of expensive black-box models, \textit{e.g.} Finite Element simulation for electromagnetic devices. It enables to reduce the computational burden for optimization purposes. Combining meta-models with a branch and bound strategy will lead to high fidelity global solutions. But, the effeciency of these algorithms relies on the estimation of the bounds. In this work, we investigated the error prediction error given by meta-models to apply such approach and applied it to electromagnetic benchmark problems.



Comparison of approaches for Optimization of Electromagnetic Devices using Finite Element Method

Reda El Bechari1,2, Stéphane Brisset1, Stéphane Clénet1, Frédéric Guyomarch1, Jean Claude Mipo2

1L2EP, France; 2Valeo EEM, France

In the context of optimization using finite element method (FEM), many issues arise, mainly numerical noise due to re-meshing when dealing with geometric parameters furthermore the computational time of the overall optimization can be prohibitive. In this communication, we compare different approaches to deal with this type of problems. Some are direct approaches for which optimization methods are coupled directly to FEM simulation and other approaches aim to replace the expensive FEM simulation with cheap meta-models that reduce the overall time of optimization. As treating different approaches, a comparison protocol is proposed and an electromagnetic optimization benchmark is treated.



Optimization of operating conditions considering output of Dual-Winding 6phase PMSM for EPS application

YOUNGWOO NOH1, Jun-Hui Won1, Seung-Taek Oh1, Ju Lee1, Chan-Bae Park2, Ho-Joon Lee3

1Hanyang university, Korea, Republic of (South Korea); 2Korea National University of Transportation, Korea, Republic of (South Korea); 3Busan Institute of Science and Technology, Korea, Republic of (South Korea)

There is an increasing demand for safety due to the commercialization of autonomous vehicles. In such situation, it is difficult to satisfy the high requirements of ISO26262, which is a functional safety standard specification of automotive components, as an existing system. Especially, in electric-power steering system, this functional safety study is proceeding one step further. In the existing system, if an electrical fault occurs in a motor or motor driver, a safe state called ‘fail-safe’ or ‘fail-silence’ is targeted through a detection and reaction mechanism. In other words, the existing EPS system pursues ‘no assist’ as a safe state.

However, in the autonomous driving system, a safe state of the fail operational level should be targeted through a detection, isolation, and recovery from the failure. In order to achieve this safe state, several redundant systems have been adopted. In other words, two redundant system configurations constitute a fail operational system in which one side fails and the other side maintains normal operation. Especially, in the electric power steering system, since the steer-by-wire system is introduced from the viewpoint of autonomous driving, the importance of redundancy design concept is increasing more and more. The safe state changes to point of ‘half-delivery’. For the design of this redundancy concept, dual winding motor is mainly applied to fail-operational EPS system. This paper investigates the case where the currents of each winding are the same or different in a dual winding motor and is proposed for optimal operating conditions through simulation and experiments



Multi-Physics Optimum Design of a BLDC Motor for a Special Application Using Different Core Materials

Morgana de Lemos1, Renato Carlson1, Nelson Sadowski1, Thiago Bazzo2, José Fabio Kolzer2

1GRUCAD/UFSC, Brazil; 2UTFPR, Brazil

This paper presents an optimization oriented multi-physics model of design using a deterministic optimization algorithm (quadratic sequential programming). The model makes use of equivalent reluctance circuits to represent the nonlinear magnetic behavior of the core and an equivalent thermal circuit to determine the temperature in the motor parts. Besides these, electric, geometric and economic models are used. The model allows to develop the optimal design of a BLDC motor with two distinct operating points, one said nominal, and another said maximum. This operational requirement is verified using a simulation with MatLab. In addition, the possibility of using Soft Magnetic Composite (SMC) in this project is also analyzed, comparing the SMC design with the classic solution using laminated steel. For this, a Pareto curve is obtained with optimal solutions for both materials allowing the designer to determine the compromise between cost and efficiency that best meets the requirements of the application.



Design Optimization of Rotor Structure in Synchronous Reluctance Motor to Improve Torque Characteristics in Several Driving Conditions Using Topology Optimization

Yoshifumi Okamoto1, Yuki Yamashita1, Reiya Suzuki1, Hiroyuki Kaimori2, Shinji Wakao3

1Hosei University, Japan; 2Science Solutions International Laboratory, Inc., Japan; 3Waseda University, Japan

Interior permanent magnet synchronous motor (IPMM) is widely applied to many industries. However, the cost of IPMM is generally higher than the motors without permanent magnet. On the other hand, because the synchronous reluctance motor (SynRM) which doesn’t use the permanent magnet, the cost of SynRM is lower than that of IPMM. On behalf of its advantage, the torque characteristic is worse than that of IPMM. To improve torque characteristic of SynRM, many structures of rotor core are designed and applied to SynRM. Therefore, the design method using topology optimization (TO) innovates the structure of rotor core of SynRM. In this paper, the design method of rotor core in SynRM is investigated. Since the SynRM is normally driven in several driving condition, the performance of SynRM should be enhanced at every condition. Therefore, the rotor structure is comprehensively optimized using TO at multi-current condition.



Squirrel Cage Optimization of Synchronous Motor with Multi-Segment U-Type Permanent Magnet Rotor

Maria Sofia Pechlivanidou, Antonios Kladas

National Technical University of Athens, Greece, Greece

The paper presents a methodology for the optimization of squirrel cage in a multi-segment U-type permanent magnet rotor of a synchronous motor. The adopted approach is based on sensitivity analysis of the geometrical parameters and 2D FE modeling. Following this process, both high starting torque and synchronization capability in case of grid supply as well as low eddy current losses in case of inverter supply are achieved.



The improvement of an Electromagnetic Acoustic Transducer using TLBO Algorithm

Houssem Boughedda1, Tarik Hacib1, Yann Le Bihan2, Hulusi Acikgoz3

1University of Jijel, Algeria; 2GeePs - Group of electrical engineering - Paris, France; 3Engineering Faculty, Karatay University Konya, Turkey

The Electromagnetic Acoustic Transducer (EMAT) is a recent ultrasonic technique that generates and detects ultrasonic waves in the conductive material without any physical contact compared to traditional ultrasonic techniques. The main downside of this technology is the low signal that generates inside material by EMAT. Therefore, the aim of this work is to improve the sensibility of the EMAT, through designing an optimal structure of EMAT using a new heuristic algorithm called Teaching Leaning Based Optimization (TLBO). A forward model was built in the COMSOL Software based on the Finite Element Method (MEF) to simulate the electromagnetic analysis and the mechanical analysis of EMAT system, in order to evaluate the Lorentz force and the mechanical displacement that produce at the surface of the specimen under test. TLBO algorithm is used to solve the inverse problem in order to deduce the optimal geometrical parameters of EMAT, which generate the extreme value of the Lorentz force and the high vibrations inside material. The obtained results show the efficiency of the optimal sensor, and encourage further work on this subject.



Particular Coupled Electromagnetic, Thermal, Mechanical Design of High Speed Permanent Magnet Motor

Dimitrios Tsiakos, Antonios Koronides, Christos Krasopoulos, Maria Sofia Pechlivanidou, Antonios Kladas

ICCS-National Technical University of Athens, Greece

Permanent magnet motors constitute a favored choice for electric drive applications in particular in high speed ranges due to their increased efficiency, compact rotor structure and relative robustness. The respective design involves analysis of coupled electromagnetic, thermal and mechanical phenomena imposing respective limitations. Direct coupling of such a multi-physics problem results in computational onerous optimization schemes; in order to tackle this difficulty an innovative weak coupling design procedure is introduced considering interdependence of the phenomena involved during both preliminary and final design stages foreseen. The proposed methodology has been applied in a surface mounted permanent magnet motor case.



Analysis of Eddy Current Loss Reduction considering Magnet Gear Shape using Response Surface Methodology

Min-Seok Kim1, Eui-Jong Park2, Sung-Ho Lee1, Yong-Jae Kim2

1EV Component & Materials Group Korea Institute of Industrial Technology, Gwangju, Korea; 22Department of Electrical Engineering, Chosun University, Gwangju, Korea

A guide was installed on the outer rotor to reduce the eddy current loss of the magnet gear. However, the addition of a guide to the magnet gear reduces the torque. To solve this problem, a pole piece shape is proposed. For the pole-piece design, the results were analyzed through the electromagnetic field finite element analysis, and the relationship between the design variables and the objective function was analyzed using the Box-behnken method. Therefore, the proposed pole piece design solves the problem of torque reduction caused by guide installation. Also, the applicability was confirmed by comparing the basic model with the transmission torque characteristics.



Novel Strategy applying Distance-based Adaptive Inertia Weight on Particle Swarm Optimization for Designing IPMSM

Jin-Seok Kim1, Byeong-Kwan Son1, Dae-Woo Kim1, Jong Wook Kim2, Yong-Jae Kim3, Sang-Yong Jung1

1Sungkyunkwan University, Korea, Republic of (South Korea); 2Donga University, Korea, Republic of (South Korea); 3Chosun University, Korea, Republic of (South Korea)

In this paper, a novel strategy of distance-based adaptive inertia weight is proposed for rapid and accurate performance of particle swarm optimization(PSO). Based on the distance of each particle, the search area is divided into outer and inner search areas. When the particles are located in the outer search area, the inertia weight is set to the basic(minimum) value for wide exploration of particles to find another global minimum point. On the other hand, when the particles are located in the inner search area, which is the average distance circle, the inertia weight is changed by the distance between the global minimum point and position of particles to guarantee intensive exploitation. To verify the effectiveness, the proposed algorithm is compared to established adaptive inertia weight(AIW) PSO with three test functions: the Branin, Goldstein and Rastrigin functions. Moreover, the suggested algorithm is applied to the design of an interior permanent magnet synchronous motor to search for the optimal design in terms of the on no-load characteristics, such as total harmonic distortion of back electromotive force under the condition to achieve the target average torque.



Optimal Design of a Three-Coil Wireless Power Transfer System Using Memetic Differential Evolution

Xin Zhang1, Zhaohui Shang1, Lei Wang2, Xiu Zhang1, Weinong Fu3, Shuangxia Niu3

1Tianjin Normal University, China, People's Republic of; 2Institute of Biomedical Engineering, Chinese Academy of Medical Sciences and Peking Union Medical College; 3Department of Electrical Engineering, The Hong Kong Polytechnic University

The technology of wireless power transfer (WPT) has broad applications especially in implantable biomedical devices. Because of the limitation of the size of the receiver coil, how to lengthen the power transfer distance is crucial in biomedical applications. Three-coil WPT system has the potential to achieve greater distance and higher efficiency than traditional two-coil WPT system. This paper studies a novel three-coil WPT system with a receiver coil and two transmitter coils. In order to maximize the efficiency of the three-coil WPT system, a memetic differential evolution (MDE) algorithm is proposed and analyzed. In the three-coil WPT system, based on the Biot-Savart’s law, the electromagnetic property of the square coil is analyzed using finite element method. Then the structural design of the system is optimized by the MDE algorithm. The simulation and experiment results show that the receiver coil can receive about dozens of millivolt, which is more efficient than two-coil WPT system.



Comparison of different meta-model techniques for torque optimization constrained by mechanical strength in automotive high-speed Spoke-type IPM machine

Xiaoyu Huang1,2,3, Morvan Ouisse1, Abdelkader Benabou2, Raouaa Bahri3, Jean-Claude Mipo3

1Univ. Bourgogne Franche-Comté, FEMTO-ST Institute, CNRS/UFC/ENSMM/UTBM, Department of Applied Mechanics, 24 rue de l’épitaphe, 25000 Besançon, France; 2L2EP / Université de Lille, 59655 Cité Scientifique, France; 3Valeo - Equipements Electriques Moteur, 2 rue André Boulle, 94000 Créteil, France

This paper focuses on the torque optimization of a high speed Spoke-type interior permanent magnet (IPM) machine for automotive powertrain application. The optimization target is to balance the torque performance at low speed and durability of rotor strength at maximum speed. Firstly, torque and mechanical centrifugal force analysis using 2-D finite element method (FEM) are established to obtain the objectives and constraints. Secondly, metamodeling techniques including different kind of design of experiment (DOE) and meta-models are used to construct the design space, objective functions and constraint functions. Comparison of metamodeling techniques is carried out and engineering guidelines on selection for electric machine design optimization are given. Finally, some different optimization methods are applied on the meta-models to determine the optimal design.



Hole Sensitivity Analysis for Topology Optimization in Non-Linear Magnetic System

Seung Geon Hong, Il Han Park

Sungkyunkwan University, Korea, Republic of (South Korea)

This study proposes a hole sensitivity analysis for topology optimization in a non-linear magnetic system. In a non-linear magnetic system, because a hole sensitivity formula cannot be derived as a closed form, we employ a numerical technique to express the hole sensitivity as an interpolation function of magnetic field and adjoint field. This interpolation function is obtained in advance of design optimization and used in optimization process to search the location where the air-hole is created. The proposed method widens the design space and lowers the possibility of reaching a local optimum for non-linear magnetic system. To demonstrate the usefulness of this method, a design problem of synchronous reluctance motor is tested in a non-linear magnetic system.