Hayaho Sato 교수

This article presents a fast population-based multi-objective optimization of electromagnetic devices using an adaptive neural network (NN) surrogate model. The proposed method does not require any training data or construction of a surrogate model before the optimization phase. Instead, the NN surrogate model is built from the initial population in the optimization process, and then it is sequentially updated with high-ranking individuals. All individuals were evaluated using the surrogate mode

Purpose This paper aims to present a deep learning–based surrogate model for fast multi-material topology optimization of an interior permanent magnet (IPM) motor. The multi-material topology optimization based on genetic algorithm needs large computational burden because of execution of finite element (FE) analysis for many times. To overcome this difficulty, a convolutional neural network (CNN) is adopted to predict the motor performance from the cross-sectional motor image and reduce the numb

A novel automatic design method for permanent magnet (PM) motors using a Monte Carlo tree search is presented. The optimal motor structures are determined through a tree search, in which the motors with different numbers of poles, current phase angles, PM configurations, and numbers of PMs are simultaneously considered. At the leaf nodes, parameter and topology optimizations are performed to obtain the optimal material shape and distribution. The proposed method was applied to the optimization o

This study proposes a novel topology optimization (TO) method for permanent magnet (PM) motors based on a variational autoencoder (VAE) and a neural network (NN). The VAE is trained to embed various shapes generated from the TO into the latent space. The NN is trained to predict the characteristics of the PM motor from its latent representation derived using the VAE. After training, TO is performed in the latent space based on the prediction using the NN. We adopt the Monte Carlo dropout to main

Self‐assembled alkyl phosphate layers have been formed on a flat, anodized aluminum substrate in dilute ethanol solution containing 2 wt% n‐tetradecylphosphonic acid (TDP) and examined by low‐voltage scanning electron microscopy as well as atomic force microscopy and X‐ray photoelectron spectroscopy. Locally, multi‐layered alkyl phosphate films have been formed on aluminum, being clearly observed by a low‐voltage scanning electron microscope operated at less than 1 kV. Atomic force microscopy ob

This article proposes a method based on 3-D discrete element and finite element methods to numerically evaluate the effects of the particle size distribution of soft magnetic composite (SMC) on the electromagnetic property. The three SMC models which obey Gaussian, homogeneous, and twin-peak distributions are compared using the proposed method. It is shown that the Gaussian model has the highest permeability and the lowest loss at high frequencies under the condition that the three models have t

This study proposes a novel multi-objective design method based on Monte Carlo tree search (MCTS) for the design of permanent magnet (PM) motors. The global configurations that define the entire structure are represented by nodes in a tree structure. After MCTS is performed to select a route extending from the root to leaf node, multi-objective topology optimization (TO) is performed at the leaf node to determine the detailed shape, considering a trade-off relationship among objective functions.

This paper presents the three-dimensional modeling of soft magnetic composite (SMC) based on the discrete element method. The proposed method makes it possible to take the possible contact among the magnetic particles in SMC as well as the distributed particle size into consideration. Based on this modeling, the macroscopic B-H characteristics is computed with the finite element method considering magnetic saturation. It is shown that the three-dimensional model of SMC can have larger initial pe

This study introduces a novel multi-material topology optimization method that can represent material distribution with an arbitrary adjacency relationship. Different state spaces for the representation of material distribution are applied to the optimization of a permanent magnet motor for comparison. It is shown that the proposed method yields a higher permanent magnet motor torque performance compared to that of conventional methods. The effect of topology optimization setting on the optimize

In this study, an approach was proposed to interpret the shapes of electric motors obtained by topology optimization (TO). A convolutional neural network (CNN) model was trained on motor images and used to predict the degree of influence of local structures of motors on their characteristics. This prediction was obtained as heatmaps for visualizing the importance of each local structure. The proposed method was applied to a permanent magnet motor model. The proposed method facilitated successful

This article proposes an effective method based on a magnetic circuit for the analysis of magnetic properties of soft magnetic composite (SMC). The present method constructs an imitation of SMC by the discrete element method (DEM), which analyzes the motion of the iron particles in SMC. Based on the resulting particle configuration, the magnetic circuit is generated, and the circuit equation is solved to evaluate the macroscopic permeability and the eddy current loss of the SMC assuming that the

This paper proposes a novel multi-objective design method of permanent magnet (PM) motor using Mote Carlo Tree Search (MCTS). In this method, the whole machine structure of a PM motor, such as the number of poles and magnet configuration, and detailed magnetic structure are simultaneously optimized. The proposed method is shown to give Pareto solutions that consist of PM motors with different number of poles and magnet configuration.

This paper presents a fast topology optimization of a permanent magnet (PM) motor using variational autoencoder (VAE) and neural network (NN) with dropout. PM Motors generated through topology optimization are embedded into low-dimensional latent space using VAE. For high reliability, the properties of the motor generated from the latent space are evaluated by NN with Monte Carlo dropout. It is shown that the proposed method can provide very fast approximations of the Pareto solutions at good ac