The powertrain is a key enabler of modern transportation and mobility. Since the Industrial Revolution, technological advancements have fueled the proliferation of manufactured powertrain systems with increasing power, speed and versatility. In particular, electric drive systems play a crucial role in converting electrical energy into mechanical motion. They are central to the efficient and clean utilization of renewable energy, including the recovery of excess energy that would otherwise be wasted.
In electric drive systems, the integration of tightly coupled mechanical, electrical and electromagnetic components into a highly compact system, required to operate under harsh conditions, presents a complex, constrained multi-physics design optimization problem. Despite the widespread adoption of computer-aided engineering (CAE), design decisions are typically made through nested, sequential iterations and often rely on heterogeneous data with varying levels of fidelity. This can significantly limit the potential for downstream performance optimization. A virtual twin, serving as a consistent reference throughout the entire design process, can be integrated with multi-physics simulation technologies to help address the challenges inherent in electric drive system development.
Electric Machine Initial Design
Electric machines, responsible for converting electrical to mechanical power—and vice versa during braking or deceleration—are critical for delivering torque efficiently and reliably across a wide speed range, minimizing undesirable effects such as heat, noise and vibration. Permanent magnet electric machines, in particular, are widely used for their high-power density, despite the high cost of rare earth materials, and exemplify the associated design challenges. Their efficiency (or thermal loss), torque magnitude and quality and operational speed range are highly sensitive to the configuration of magnets, windings and the iron cores of the stator and rotor.
The design of electric machines for electric drives can be performed efficiently and accurately using CAE tools that are tightly integrated with computer aided design (CAD) systems. High-fidelity CAE tools employ finite element methods for electromagnetic, structural and thermal analyses, and they allow for the easy incorporation of diverse conditions and scenarios.
As illustrated in Figure 1, by switching from the V-shaped magnet layer—commonly used in traction applications due to its higher efficiency across a wider speed range—to the I-shaped one, known for delivering higher torque with reduced torque ripple, results in notable performance differences with only a few parameter changes. This is achieved by leveraging data associativity between the parametric sketch, 2D CAD model and simulation model of the electric machine. In practice, the magnet and magnet hole parameters can be optimized to achieve the best overall performance, whether using a single layer or multiple layers.
Electric Machine Thermal
Effective cooling strategies for electric machines is critical not only for achieving peak and continuous performance but also for ensuring long-term reliability. These strategies may include various combinations of stator and rotor core cooling, slot cooling and end-winding cooling. The end windings are particularly vulnerable to thermal issues because, despite being a major source of heat, they are not surrounded by thermally conductive materials— such as the iron core—that facilitate heat dissipation. To address this, low-temperature oil jets can be used to cool these components directly.
As illustrated in Figure 2, when oil from a top inlet is distributed through multiple injectors, variations in the number of nozzles and their angular positioning affect oil surface coverage and subsequent dripping patterns. The pipe’s angular length, along with the number, placement and orientation of the coolant nozzles, is parametrized. Compared to the configuration on the left—with 5 nozzles spaced over 125 degrees—the configuration on the right, featuring 9 nozzles over 180 degrees, requires lower pressure to drive the coolant but results in a substantially higher stator temperature. The situation is further complicated by complex interactions involving the high-speed rotating rotor, internal airflow, and external factors such as vehicle acceleration and deceleration. In this scenario, efficiently exploring different configurations through simulation is critical to achieving optimal cooling performance without risking system failure.
Multiphysics Electric Drive
The housing of an electric drive system must be lightweight while ensuring sufficient structural rigidity, vibration resistance, and durability under harsh operating conditions. In many cases, its reliable performance assessment becomes feasible only after significant design decisions have already been made, at which point the design space is considerably constrained. However, a multi-physics, multi-fidelity design framework enabled by virtual twin technology can greatly improve the effectiveness of design optimization during the early stages of development.
For example, the housing shell thickness and internal spacing of gear components influence multiple performance targets, including gear lubrication efficiency, transmission losses, and structural vibration behavior. As shown in Figure 3, Design 2 features a significant increase in intermediate-speed shaft spacing compared to Design 1, resulting in a better-lubricated (less dry) high-speed pinion and reduced torque loss. In contrast, Design 3 combines moderate intermediate-speed shaft spacing with a substantial increase in high-speed shaft spacing, improving lubrication of the high-speed pinion while also suppressing vibrations at the high-speed shaft bearing—albeit with a slight increase in weight. This kind of concurrent multi-physics performance evaluation using a concept shell model enables efficient multi-objective design trade-offs.
On the other hand, additional housing features—such as stiffeners and shell splits with fasteners—can be incorporated into the parametric shell model, substantially extending the design space. This adaptability helps maintain consistency in design and engineering knowledge throughout the process, as illustrated in Figure 4. In the analysis on the left, various excitation sources—such as electromagnetic forces and those generated by gears and bearings—are considered within a simplified model that uses a basic housing shell and rigid-body representations of the gear reducer components, allowing for faster turnaround. In contrast, the analysis on the right incorporates a more detailed housing model, using feature templates and flexible body representations that account for structural modes, resulting in improved accuracy.
Conclusion
Virtual twin technology is transforming the development of electric drive systems by unifying modeling and simulation into a seamless, powerful engineering approach. As system-level parameterization and standardized multi-physics simulation practices continue to evolve, electric drive engineering is reaching new heights in speed, accuracy and collaboration. At the same time, the global push toward renewable energy is accelerating the synergy between electric drive innovation and virtual twin technology, delivering smarter designs, faster iterations and more sustainable outcomes than ever before.
Interested in the latest in simulation? Looking for advice and best practices? Want to discuss simulation with fellow users and Dassault Systèmes experts? The SIMULIA Community is the place to find the latest resources for SIMULIA software and to collaborate with other users. The key that unlocks the door of innovative thinking and knowledge building, the SIMULIA Community provides you with the tools you need to expand your knowledge, whenever and wherever.
Young-Chang is a SIMULIA Industry Process Expert Senior Specialist.