JMAG User Conference 2026 | Detroit

This talk presents a systematic framework for electric powertrain design, spanning from physics-based component loss modeling to full system optimization. We outline a structured design methodology that encompasses application definition, design space generation, and multi-objective optimization. Along the way, we’ll show real examples of how a single motor design can serve multiple platforms — from a truck to a sports car — by tuning the gear ratio and drive unit configuration to match each vehicle’s needs.

JMAG is being developed with a focus on data-driven design (DDD).
DDD is only possible with massive, reliable datasets, making the accuracy and speed of simulations more critical than ever before.

It is said that the introduction of machine learning-based surrogate models can significantly enhance the power of DDD. However, these surrogate models require extensive training using simulation results, which makes the speed of generating FEA results increasingly important—even for 2D analyses.

We have been continuously improving JMAG’s performance to meet the growing demands of DDD.

Additionally, the question of how surrogate models should be used in DDD remains an important topic of investigation in our JMAG development efforts.

The latest advancements will be presented in the upcoming presentation.

This presentation lays a framework for data-driven technology development in electric machines. Rather than approaching motor design as a single optimization problem for one application, the talk asks a broader question: For how many applications, and in what physical form, can a new motor technology move the needle meaningfully? This includes not only electromagnetic performance, but also material selection, mechanical arrangement, manufacturability, cooling, gearing, operating environment, and capital efficiency.

Designing electric machines for EVs requires a balance of electromagnetic, structural, and thermal performance to meet demanding specifications, including peak and continuous capabilities, packaging constraints, and efficiency targets. This presentation details an automated multiphysics optimization framework developed in JMAG to streamline the eMachine design process. By coupling electromagnetic, structural, and thermal analyses into a unified workflow, the approach systematically evaluates a broad design space while ensuring all critical constraints and vehicle-level requirements are met. This methodology enables faster convergence, more comprehensive trade-off analysis, and significantly reduced engineering effort.

High-fidelity NVH analysis and optimization of induction motors including: (i) 2D and 3D FEA to predict EM forces for induction motors; (ii) Rotor bar skew NVH optimization; (iii) Rotor bar number optimization; (iv) Analysis vs test correlation

Direct-drive applications, such as micro-mobility and actuators, require low-speed, high-torque capability to enable strong acceleration and efficient operation across a wide region. Transverse flux machines (TFMs) provide several benefits, including high torque density, modular structure enabling fault tolerance, and reduction in active materials through a compact design. Given that TFM stator and rotor parts are subjected to multidimensional flux paths, the electromagnetic modeling and design becomes rather challenging that require setting up complex 3D FEA simulations to analyze the electric machine’s behavior. In this presentation, a case study for the multiphysics design optimization of an outer-rotor TFM will be showcased using an integrated workflow within JMAG-Designer in specific operating conditions.

This presentation outlines how advanced electromagnetic FEA workflows using JMAG accelerate electric machine and drive unit development. The presentation first shows how large-scale FEA can be embedded directly into online optimization loops, enabling high fidelity results to guide real-time design updates. Then, a demonstration of how streamlined 3D FEA supports rapid motor and drive unit decisions early in the design cycle, providing quick insight into torque, losses and packaging tradeoffs is presented. This presentation also highlights the use of large-batch FEA to perform Monte Carlo variation analysis, quantifying sensitivity to materials, tolerances, and operating conditions while strengthening virtual model validation. Together, these capabilities position JMAG’s electromagnetic FEA as a fast, predictive engine for optimization, architecture selection, and robust verification.

Bearingless motors combine torque and radial force production in a single electromagnetic device. This is used to levitate their rotor, removing conventional contact bearings in favor of friction-free, oil-free, non-contact electromagnetic bearing shaft support, which increases the reliability of the rotating machinery. However, levitation control requires rotor displacement sensors, which add cost, bulk, and complexity to the motor system. This presentation explores preliminary results of using the motor coils as the displacement sensors for active levitation control, leveraging JMAG simulation and optimization for co-design of the electric motor inductance characteristics to suit sensorless levitation control algorithms.

This study focuses on the design methodology of heavy-duty traction motors targeting high continuous power capability. Various slot–pole combinations are systematically evaluated to understand their impact on electromagnetic performance, losses, and thermal behavior. Based on this comparative analysis, an optimal topology is selected for further detailed investigation and design optimization.