This presentation will focus on electromagnetic field analysis, artificial intelligence, and technologies that integrate the two, and discuss how they can be used in the design and development of electrical and electronic devices in the future. First, the presenter will report on the optimal design of motors for air mobility, wireless power transmission, microwave devices, and other examples where electromagnetic field analysis is used. He then discusses how artificial intelligence trained on the large amount of data generated by electromagnetic field analysis can be used for optimal design, reliability evaluation, equipment control, and fault diagnosis. He also discuss the possibilities and challenges that arise from applying artificial intelligence to design and development. Finally, based on the above discussion, he presents a possible picture of future engineers effectively using electromagnetic field analysis and artificial intelligence.

We will present our own unique techniques to motor development using topology and parametric optimizations to inspire unconventional and innovative ideas as well as drive motor design efficiency. Our case study runs a topology optimization to freely transform geometry for insight into innovative new ideas about the geometry before optimizing the dimensions. A multi-objective parametric optimization balances maximum torque and motor rotor rigidity while realizing lower computational costs using a deep learning-based surrogate model.

In this presentation, examples of applying data driven design to electric motor development are introduced.
The optimization problems of system specific power density and system efficiency with consideration of multi-physics are solved by MBD.
In motor topology optimization applied to SPM motor, Rotor shape and materials considering reduced magnet amounts for torque.

Magnetic fluxes from rotor magnets are available for detecting the angle of motors. Bearingless motors have motor windings to generate torque and suspension windings to generate force for magnetic levitation. These windings generate leakage flux, and may interfere with angle detection. Notably, PWM harmonics of current cause complicated problems when the motor is driven by an inverter.
This presentation will introduce an example of angle detection analysis that uses the PWM Current Vector Control macro component, which has been included from JMAG-Designer Ver. 22.0. A proposed method reduces angle error due to current, and the method is also effective for PWM harmonics.

We propose a 10-pole-12-slot switched reluctance motor (12/10 SRM) with a hex connection for a traction motor of EVs. A 12/10 SRM can be driven under hysteresis control or vector control, which enables a switching drive between them. In this paper, the performance of a 12/10 SRM under hysteresis control and vector control is computed by using the control circuit function of JMAG-Designer.

Considering the end users’ requirements of fast charging, and the maturity of SiC, more and more OEM selected high voltage battery system (800V) into vehicle application, but it’s a big challenging to Electrical motor, which is the key power component of Electrical drive unit. Based on high voltage application in EDS, the presentation will discuss the effect of high voltage system on drive motor performance, make a comparison of simulation and test results, finally, we would like to introduce the finding, which could bring some ideas for high voltage motor design.

The goal is to design a lightweight high-power density IWM, and use JMAG design and optimization to improve motor efficiency and performance, so that the motor can meet the weight requirement.
Then external software Ansys mechanical and CFD software are used to design and analyze the system of shell strength and heat dissipation to ensure that the motor meets the requirements.
And share the IWM production process and future prospects.

The widespread adoption of electric vehicles poses significant technical challenges for the automotive industry. Magna Powertrain addresses these challenges by supplying complete e-drive systems that meet the performance and efficiency requirements of traction motors. To meet these requirements in the short development cycles of vehicles, we employ numerical simulation tools from the initial phases of projects through to completion. These tools are crucial, for example, to design cooling concepts that effectively dissipate the waste heat generated by the high-power-density traction motors. A multiphysics tool like JMAG is invaluable for designing cooling concepts that meet the thermal requirements without compromising the machine’s electromagnetic performance. This presentation covers the thermal analysis of a Permanent Magnet Synchronous Motor (PMSM) using JMAG and explore future perspectives on using JMAG for traction motor design.

There is a great need for large-scale computation to simulate real phenomena with high accuracy and multi-case execution to search for optimal solutions from a large design space. However, HPCs for these purposes tend to be expensive, both on-premise and in the cloud, and this is one of the barriers to building a multi-core computing environment.
In this presentation, we will show an example of large-scale computation and multi-case execution using a one-node computer that can be installed on a low budget, and will discuss HPC selection points based on its performance.

The automobile industry trend is toward energy saving and environmental load reduction, and the electrification of motor vehicles is progressing. In addition to high efficiency, small size and light weight, the market’s demands for motors also include low cost, environmental response, and global production. Against this background, research and development of motors that do not use magnets is progressing. NIDEC has selected induction motor to develop next-generation traction motor systems.
In this presentation, we will introduce examples of large-scale analysis of induction motors using MPP and PSL, characteristic map analysis, and the results of comparative verification with actual machine.

Oil and gas plants as well as steel mills have broad demand for large capacity TMEIC synchronous motors for use in industrial compressors, pumps, and other such applications. In recent years, CO2 emission and running costs reductions are driving the need for even greater efficiency against a background of environmental problem and rising energy costs. In order to evaluate motor efficiency, prototypes of these motors is difficult because the production costs of large capacity machines over 10,000 kW are veryhigh. Finite element analysis provides a highly accurate solution to estimate the losses. This case study divides one electric motor into several models and runs three-dimensional analyses to obtain the losses produced in each part for a comparison against actual measurements. These analyses do take more time to evaluate the large-scale models necessary to simulate such a sizable machine, but this case study takes advantage of Massively Parallel Processing (MPP) to report on the effective reductions in the time required to run the analyses.

JMAG is focusing on enhancing its design automation capabilities with the aim of speeding up product development and increasing performance. This presentation will introduce the functions that have been realised in 2024 and those that will be further developed in 2025.
In particular, to find the optimum design that meets high requirements in a short time,
 – Model building and optimisation functions for rapid evaluation of various designs
 – Thermal analysis functions to enable designs that take thermal effects into account.
 – Fast massive-cases processing capabilities.

We will also report on the development status of JMAG-Express to make it easier for designers to utilise JMAG.

High-performance permanent magnet machines are gaining attention in many fields, including electric and hybrid vehicles, electric and hybrid airplane applications, and industrial and residential applications such as fans, pumps, and compressors. Some applications require high volumetric power density (kW/l), and others need specific power density (kW/kg). Some applications prefer efficiency to be a priority in the design. Not using heavy rare earth elements such as dysprosium is gaining attention for some applications. This presentation will provide design and simulation synchronous PM machine examples based on the research carried out at the Wisconsin Electric Machine and Power Electronic Consortium (WEMPEC) at the University of Wisconsin.

Since the resolver is placed very close to the motor and there is a risk that leakage magnetic flux will effect detection accuracy, it is necessary to confirm how the leakage magnetic flux from the motor effects the resolver.
On the other hand, actual measurement of leakage magnetic flux is difficult, and it is expected that the phenomenon will be understood through FEA, and a modeling method must be established.
In this presentation, we will analyze the leakage magnetic flux from the motor that is applied to the resolver. To evaluate the leakage magnetic flux, it is necessary to model the surrounding magnetic bodies, the analysis can be done by modeling magnetic bodies such as bolts. We will discuss that it is possible to perform effective evaluations, and that it is possible to calculate realistic times for large-scale models by performing parallel calculations.

DeepDrive, a young high-tech company based in Munich, has developed an innovative electric motor technology based on a dual-rotor radial flux design. With the target to become a key automotive supplier, DeepDrive focuses on reducing development times, cutting prototype costs, and minimizing time-to-market. A critical element in this process is the application of advanced simulation techniques. The motor technology is based on a novel counter-skewed wave winding, offering high torsional stiffness, excellent electromagnetic properties, and efficient thermal management. Due to the motor’s advanced and highly specific geometry, precise 3D simulations are essential to achieve optimal performance. For this, DeepDrive utilizes JMAG software to ensure accuracy in modeling and design.
In this presentation, the simulation strategies and optimization approaches employed in the development process will be detailed. Furthermore, simulation results will be validated against empirical measurement data, showcasing the effectiveness of the applied methods in advancing this cutting-edge motor technology.

Our planet, our people and bio-species, ALL need significant reduction of rare-earth magnets usage that are not sustainable for ALL. However, these rare-earth permanent magnets offer significant power, torque density and efficiency for electric machines, one of the main contributor to the Electrification that also reduces significantly global warming issues.

This dilemma can be solved with sound engineering and significant innovation. Through optimization tools within JMAG, YEESMA will present 2 case studies showing how to solve the trade-off between Sustainability, Cost and Packaging Constraints. These 2 applications are in the eMobility area, but the same solutions can be applied to Industrial or Renewable Energy sectors.

This presentation introduces design and analysis methods for synchronous reluctance motors and switched reluctance motors.
In the design of the synchronous reluctance motor, a method was used to define the shape of the flux barrier using Zhukovsky’s airfoil function in order to reduce the number of parameters. In addition, we used Python and JMAG-Designer to carry out a design that minimized torque ripple.
In the analysis of the characteristics of the switched reluctance motor, a 3D model was used to improve the analysis accuracy of the flux linkage characteristics. Furthermore, by taking into account the eddy currents in the wire, the measured values and analysis results for the load analysis were in good agreement.

The automotive industry embeds magnets deeper into the core of motors than previous products as the rapid shift to electric vehicles in recent years drives motor performance and miniaturization. This trend is heightening demand for magnetizing yokes that can produce more powerful magnetic fields to magnetize magnets embedded more deeply inside of these motors. The magnetizing current, coil windings, and yoke geometry are the three elements necessary to determine the magnetizing yoke specifications, considering the principles of the magnetic flux density produced by magnetizing yokes. This case study aims to combine two JMAG parametric optimizations to examine magnetizing yoke specifications.

Determining the ideal magnetizing yoke to sinusoidally control the magnetic flux density on the workpiece surface.
Before, we considered the best shape of magnetizing yoke for controlling the magnetic waveform in a workpiece when a magnetic field is applied, Now we can use JMAG functions to optimize the magnetization results and consider what shape will give us the ideal magnetization waveform.

The Soft Magnetic Composite (SMC) core, which is made by insulating and compacting iron powder, has a higher saturation magnetic flux density than ferrite cores and lower eddy current losses than electromagnetic steel sheets. Additionally, due to its three-dimensional isotropic magnetic properties, it allows for the design of magnetic circuits that were previously difficult to achieve. JFE Steel has developed an insulated coated pure iron powder (Denjiro^TM) for SMC cores by our unique water atomization technology. Although cores used in power electronics equipment operate under inverter excitation, the iron losses in SMC cores under inverter excitation hasn’t been investigated well. In this presentation, we will discuss the characteristics of iron losses in SMC cores under inverter excitation compared to electromagnetic steel sheets and an application example to an axial gap motor, which has recently attracted attention as a motor capable of miniaturization and high torque. Optimization calculations are conducted using genetic algorithm (GA), and the efficiency of the design which is optimized is compared between measurement and simulation.

Automotive electric motor designs must satisfy performance, heat, strength, and a wide variety of other requirements. Specific expertise about each of these requirements is vital to the design, but an approach without an evaluation considering every requirement often results in rework. A high-quality thermal design considering the cooling system is essential to large high-output drive motors built into vehicles. This case study examines the JMAG thermal analysis functions that should reduce rework and realize an effective thermal design process to verify whether the use of these simulations is efficient when designing cooling systems for automotive drive motors.

The problem with thermal evaluation of motors through actual testing is that it requires a large number of prototypes and a long development period.
To solve this problem, Magnetic field-thermal coupling analysis JMAG-Designer was investigated introducing.
In this presentation, comparison between analysis and actual testing using a mass production motor as the first step in verifying the analysis accuracy.

Hitachi Astemo, Ltd., are promoting analysis lead design and application for e-Axle in NV (noise vibration) field. E-Axle is a power transmission system in which parts such as a motor, gearbox, and inverter are integrated for eco-friendly electric vehicles. However, it is considered to be the one of noise sources, and it is important to evaluate the NV performance at the design stage. By solving the interaction between magnetic force using JMAG and gear meshing force using ANSYS Motion, the possibilities of NV evaluation in the transient actual operating condition for the e-Axle are verified. Moreover, the possibilities of analysis technology to apply for a motor with skew-typed rotor also verified. In the presentation, the details of the study will be introduced.

In recent years, with the miniaturization and higher efficiency of rotating machines, there is also an increasing demand for noise reduction. In particular, electromagnetic vibrations and noise caused by the resonance of the stator’s electromagnetic force and eigenmode can often be problems. To predict these electromagnetic vibrations and noise at the design stage through numerical simulations, it is necessary to reproduce the electromagnetic force and vibration characteristics. Traditionally, we have been identifying the vibration characteristics of the stator’s laminated steel sheets and windings from experiments, but now we are developing a method which can determine these characteristics analytically in order to realize virtual prototyping. In this presentation, we introduce a case study of electromagnetic-structure-acoustic analysis using JMAG and Actran, targeting the electromagnetic vibration and noise of three-phase induction motors.

Induction Hardening is one of the strengthening methods of steel, and it can be incorporated into production lines due to its short time process compared to other hardening methods. Setting the conditions for moving hardening of single diameter shafts or single-shot hardening by using dedicated coils for a specific part is relatively simple. However, in actual production lines, there are cases where stepped shafts must be heated using single diameter coil. In such cases, there are many design variables, making it difficult to derive optimal conditions manually. This presentation introduces the verification of the optimization calculation of hardening heating conditions for stepped shafts by using parallel execution of multiple cases on a cloud computing service with PSL.

In induction heating, the heating coil is an important factor in determining the temperature distribution and heat treatment quality of the treated product. However, heating coils and treated products are often individual drawing, and expertise is needed to design the optimal heating coil.
As one of the efforts to establish a new design method for heating coils, we performed topology optimization, designed and fabricated heating coils based on the results, and conducted quality evaluation, and will report on the results.

Lucid’s electric motors are optimized using a multi-physics and system level approach to deliver a compact powertrain that combines motor, gearbox, differential and thermal system into one cohesive package. Lucid’s powertrain sets multiple industry records with leading torque density of >20Nm/kg and power density of >15kW/kg while also delivering best in-class vehicle efficiency of up to 5miles/kWh. Miniaturization of the electric powertrain is key for creating spacious interior vehicle cabin and high efficiency has a compounding effect of reducing battery size and mass of the vehicle for the same range.
A power dense and ultra efficient motor presents unique challenges in electromagnetic design and thermal management. The presentation will cover examples of how Lucid uses 3D Finite Element Analysis tools to understand and overcome some of these challenges and extract the maximum performance out of the electric motor

Volvo cars have for some years been working with inhouse development of electric machines and electromagnetic simulations. FEA simulations are time consuming and therefore the simulation time is something that is beneficial to reduce to save cost. This results in that each simulation is of importance in terms of result but also development cost, and you want to reach a final design as quickly as possible with the best accuracy.
The use of Direct optimization using multi-objective genetic algorithm often needs to solve many cases to converge to designs fulfilling the optimization targets. Depending on the number of objectives it is common to run 5000 – 10000 cases. This can take several hundreds of hours. We have successfully run large optimizations for PMSM machines using PSL where the time saving is hundred-fold when comparing to a small number of solvers which is normally needed to ensure the usage of JMAG designer within the design team.
To save cost and development time reducing the number of test objects and test time is desired, a good prediction of the electric machine losses is enabling this. The most accurate way to calculate the electric machine losses, especially in magnets and winding is to simulate a 3D model of the design. 3D modelling is quite resource intense due to the large mesh size needed and thus a high simulation capacity is needed.
The above topics can be enabled or improved by utilizing the PSL solution provided by JSOL for JMAG. The usage of PSL is enabling a faster time to final design but also making sure that calculation of an efficiency map for a 3D model can be performed within a reasonable time frame.

In the development of electric vehicle motors, it is essential to meet a wide range of requirements, such as power performance, noise and vibration, demagnetization resistance, and compactness. Among these, rotor design demands significant development time. To rapidly create designs that align with current needs without being constrained by conventional thinking, the use of topology optimization in the electromagnetic field is on the rise. However, while topology optimization offers high shape flexibility, it is prone to instability in shape convergence under multi-material, multi-performance, and multi-constraint conditions, posing challenges for its application in product development. This paper introduces case studies on the integration of parametric and topology optimization (PO+TO) and the impact of objective functions in topology optimization.

In steel industry, continuous processing lines are used to manufacture most of steel strips, which are thin steel sheets used in the production of automobile bodies and containers. In continuous processing line, steel strips are continuously conveyed, and heating, cooling, and other processes are carried out continuously to efficiently produce products. In the context of carbon neutral movement, the electrification (applying induction heating) of the heating process in the line is an option to be considered. Until now, induction coils for running strips have often been designed empirically and it has been difficult to tell whether the design is suitable. To investigate the possibility of a data-driven design, the shape of the induction coil was studied by topology optimization using the temperature distribution in the running strip width direction as the objective function.

The analysis of induction heating coils is a coupled analysis of magnetic field and heat, and the calculation per case is long, and its optimization calculation is difficult to put into practical use in terms of time. Therefore, we will verify whether it is possible to shorten the total calculation time by setting up a generation that calculates with a surrogate model obtained by AI machine learning, rather than performing optimization calculations by finite element analysis (FEA) alone. Based on the optimal solution obtained, an induction heating coil is build using a 3D printer, and its validity is evaluated by comparing the actual measurement and analysis results.

■

As the demand for motor design increases in line with the acceleration of electrification, there is a growing shortage of designers and a demand for shorter development times. To address these issues, we are working on the development of a ‘Motor Design Agent’ using our proprietary generative AI technology.
In this presentation, we will present an automated motor design and development process with JMAG-Designer and AI Agents. The AI Agent works with JMAG-Designer to automatically perform tasks from idea generation, CAD model creation, analysis condition setting, design optimization and result evaluation.
The Motor Design AI Agent is not a tool that derives a single solution, but rather an agent that helps design engineers be more creative. This technology enables faster and longer-term verification than humans and increases the possibility of discovering innovative shapes. In addition, the complete automation of the design process is expected to significantly reduce design development time.

New solutions utilizing AI technology with MATLAB and Simulink will help us further improve the efficiency and reliability of motor control systems. Combining AI with conventional control methods opens new options to improve system utilization and competitiveness, and MATLAB and Simulink enable efficient simulation and verification, lowering the hurdles to technology adoption.

In this session, the construction flow of virtual sensor using neural networks and fault detection using machine learning algorithms will be explained. We will also share the key points of using JMAG-RT to build AI models with large amounts of data.

In motor design, it is common to reuse previous shapes and characteristics to shorten development times. In addition, with the expansion of machine learning, development processes that automate and optimize design are being adopted. However, rework due to problems with motor performance and vibration is also an issue. In this presentation, NWC will introduce the following:
・A method for considering power and thermal performance using 1D tools at the concept design stage.
・System analysis including motor excitation forces due to electromagnetic forces and torque ripple.
・System analysis example applying a motor model to an electric vehicle.