A review of structural design for in-wheel motors in new energy vehicles

1 Introduction

The in-wheel motor (IWM) drive system, also referred to as the hub motor, integrates the traction machine directly into the wheel hub and represents one of the most promising architectures for new energy vehicles (NEVs). By eliminating the half-shafts, differential, and mechanical transmission, the IWM architecture shortens the power path, improves drivetrain efficiency, frees chassis space for battery packaging, and enables independent torque control at each wheel for enhanced vehicle dynamics and regenerative braking (Murata, 2012; Chau et al., 2008). Against this background, the structural design of the IWM—spanning its electromagnetic topology, mechanical integration within the wheel, thermal management, and fault-tolerant configuration—has become a decisive factor determining whether this technology can move from niche high-performance applications to mainstream passenger vehicles.

Several reviews have surveyed in-wheel motor technology from particular angles: Chau et al. (2008) reviewed permanent-magnet brushless drives for electric and hybrid vehicles, Murata (2012) analyzed the innovation potential of the in-wheel-motor drive unit from a vehicle-dynamics perspective, and He et al. (2022) reviewed torque-distribution schemes for four-wheel in-wheel-motor electric vehicles. In contrast to these works, the present mini-review focuses specifically on the structural design dimension: it connects electromagnetic topology selection to mechanical integration, thermal management, and sealing reliability within the confined and hostile wheel environment, and it consolidates the most recent 2022–2026 literature on unsprung-mass mitigation, cooling architectures, fault tolerance, and corner-module integration. This cross-domain, structural-design-centred perspective constitutes the distinct contribution of the present review.

Review methodology: the literature was identified through systematic searches of IEEE Xplore, Scopus, Web of Science, and Google Scholar using keyword combinations including “in-wheel motor”, “hub motor”, “wheel-hub permanent magnet motor”, “axial-flux machine”, “unsprung mass”, and “in-wheel motor thermal management”. Articles were screened by title and abstract and selected when they addressed the electromagnetic, mechanical, thermal, or fault-tolerant structural design of in-wheel motors; peer-reviewed journal articles were prioritized over conference proceedings and non-scholarly web sources. The retained literature was categorized into five domains—electromagnetic topology, axial-flux and alternative topologies, mechanical integration and unsprung mass, thermal management and sealing, and fault-tolerant design\u2014which define the structure of Sections 2\u20136. A total of 34 peer-reviewed references are cited.

The remainder of this review is organized as follows: Section 2 examines electromagnetic topologies and design optimization; Section 3 discusses axial-flux and alternative topologies; Section 4 addresses mechanical integration and unsprung-mass challenges; Section 5 covers thermal management and sealing reliability; Section 6 reviews fault-tolerant design; and Section 7 provides a critical synthesis, identifies research gaps, and outlines a structured future research roadmap.

2 Electromagnetic topologies and design optimization

The electromagnetic topology of an in-wheel motor fundamentally determines its torque density, efficiency, and packaging compatibility within the confined wheel cavity. Among the various configurations, the permanent magnet synchronous motor (PMSM) has emerged as the dominant choice for NEV applications due to its high power density, superior efficiency, and wide constant-power speed range (Chau et al., 2008). The outer-rotor PMSM topology is particularly favored for direct-drive in-wheel applications because the rotor can be directly coupled to the wheel rim, eliminating the need for mechanical reduction gears and thereby improving overall drivetrain efficiency while reducing maintenance requirements (Yuan et al., 2019). Analytical design methodologies based on exact magnetic circuit models have demonstrated satisfactory accuracy when validated against finite element method (FEM) simulations, enabling rapid preliminary sizing before detailed electromagnetic optimization (Jing et al., 2023). Multi-objective optimization algorithms, notably the non-dominated sorting genetic algorithm II (NSGA-II), have been extensively applied to simultaneously minimize torque ripple, reduce permanent magnet mass, and maximize average electromagnetic torque and efficiency (Jing et al., 2023). These optimization frameworks typically treat structural parameters such as pole-arc coefficient, magnet thickness, air-gap length, and slot dimensions as design variables, subject to geometric constraints imposed by the wheel rim and brake system envelope (Wang, Q., 2025). Comparative studies between surface-mounted permanent magnet (SPM) and interior permanent magnet (IPM) configurations reveal that while SPM motors offer simpler manufacturing and higher initial torque density, IPM variants provide superior field-weakening capability and mechanical robustness of the magnets, making them preferable for high-speed highway operation (Guo et al., 2025). Furthermore, novel rotor structures incorporating V-shaped magnet arrangements with auxiliary air magnetic insulation grooves have been shown to reduce torque ripple from 19.14% to 4.84% while simultaneously suppressing harmonic iron losses (Guo et al., 2025). The adoption of modular stator designs and non-uniform air-gap profiles represents another promising avenue for iron loss reduction without compromising electromagnetic torque output (Mao et al., 2017). The selection of appropriate slot-pole combinations significantly influences cogging torque and winding factor, with fractional-slot concentrated windings gaining popularity for their short end-turns and high copper fill factor, despite introducing additional space harmonics that require careful rotor design to mitigate (Zhu and Howe, 2007). Hairpin winding technology, borrowed from conventional automotive traction motors, is increasingly being adapted for in-wheel applications to improve slot fill factor beyond 70% and enhance heat transfer from conductors to the stator core (Zhu and Howe, 2007). These advances collectively demonstrate that sophisticated electromagnetic design, coupled with computational optimization, can substantially enhance the performance metrics of in-wheel PMSMs within the severe geometric constraints of automotive wheels. Figure 1 systematically presents a panoramic framework of in-wheel motor technology, spanning from electromagnetic topologies to system integration, thermal management, and future evolution. At the electromagnetic design level, the radial-flux outer-rotor PMSM has become the dominant market choice owing to its manufacturing maturity and cooling accessibility; the axial-flux motor achieves higher torque density through its pancake geometry, particularly suited to the axial space constraints of wheel hubs; while the transverse-flux and YASA topologies, despite manufacturing complexity, offer differentiated solutions for specific application scenarios through their unique magnetic circuit designs. Nevertheless, electromagnetic performance optimization represents only one dimension of in-wheel motor design—when the motor is embedded within the wheel, an extreme operating environment, issues of mechanical integration, thermal coupling, and sealing reliability constitute far more severe system-level challenges. Transient thermal radiation from brake discs exceeding 300 °C directly threatens the magnetic stability of permanent magnets, while the increase in unsprung mass significantly degrades vehicle ride comfort and tire road-holding capability. Meanwhile, the limited heat dissipation area within the wheel cavity and the demanding IP67/IP68 sealing requirements create a sharp contradiction, compelling researchers to seek optimal balances among liquid-cooling channel topologies, phase-change materials, and high-temperature magnet grades. Looking forward, multi-phase winding and modular stator architectures provide a hardware foundation for fault-tolerant operation, while the introduction of SiC power devices, additive manufacturing, and the “corner module” integration concept signals that in-wheel motor technology is evolving from a single drive component toward a wheel-end comprehensive system encompassing braking, suspension, steering, and propulsion. Although core challenges including unsprung mass, cost competitiveness, and durability in harsh environments remain to be addressed, the gradual clarification of technical roadmaps has laid a solid foundation for the penetration of in-wheel motor technology from high-performance niche markets into mainstream passenger vehicle platforms.

3 Axial-flux and alternative topologies

While radial-flux PMSMs constitute the mainstream in-wheel motor technology, axial-flux permanent magnet (AFPM) motors have attracted considerable research attention due to their inherently pancake-shaped geometry, which aligns favorably with the axial space limitations of wheel hubs (Pippuri et al., 2013). The AFPM topology offers a significantly higher torque density than its radial-flux counterpart because torque production scales with the cube of the rotor diameter rather than the square, enabling substantial weight and volume reductions in the powertrain (Kim and Baek, 2025). Comparative design studies confirm that AFPM motors can be more compactly integrated into classical automotive wheel rims while meeting the torque and speed requirements of urban electric vehicles (Zhang, J., et al., 2023). The shorter magnetic flux path in axial-flux configurations contributes to reduced core losses and improved efficiency, particularly at low-speed high-torque operating points characteristic of urban driving cycles (Kim and Baek, 2025). However, the disc-like structure introduces manufacturing complexities related to precise air-gap control, stator-rotor alignment, and the structural integrity of the stator core under magnetic attraction forces (Ebrahimi et al., 2023). Dual-rotor single-stator arrangements are frequently employed to balance axial forces and enhance torque output, though this configuration complicates thermal management by sandwiching the stator windings between two heat-generating rotor discs (Zhang, J., et al., 2023). Transverse-flux machines (TFMs) represent another alternative topology that has been explored for scooter in-wheel traction applications, offering high torque density at low speeds but suffering from complex three-dimensional magnetic circuits and manufacturing difficulties associated with U-shaped iron cores and I-bridges (Pippuri et al., 2013). Recent analytical design methods for two-phase axial-gap TFMs aim to mitigate these manufacturing challenges through additive manufacturing of non-magnetic stator housings and simplified magnetic circuits suitable for low-cost production (Shi and Jian, 2018). Halbach magnetized arrays have been investigated as a means to enhance air-gap flux density and reduce back-iron requirements in axial-flux machines, though their implementation increases magnet assembly complexity and cost (Liu, J., et al., 2025). Despite these innovations, radial-flux motors retain advantages in manufacturing maturity, cooling accessibility, and supply chain availability, ensuring their continued dominance in mass-market NEV applications while axial-flux and transverse-flux designs occupy niche high-performance segments (Murata, 2012). The yokeless and segmented armature (YASA) topology represents a particularly advanced axial-flux variant that eliminates the stator yoke entirely, further reducing iron losses and weight while maintaining high torque density through careful magnetic circuit design (Wang, X., et al., 2017). Multi-disc axial-flux configurations, where multiple rotor-stator pairs are stacked along the shaft, offer further torque multiplication for heavy-duty commercial vehicle applications without increasing motor diameter beyond wheel packaging limits (Ebrahimi et al., 2023). Table 1 consolidates the comparative insights derived from Sections 1 and 2 into a unified multidisciplinary framework, juxtaposing radial-flux, axial-flux, transverse-flux, and YASA topologies across electromagnetic, mechanical, thermal, and fault-tolerant design domains. The axial-flux motor’s inherently higher torque density—enabled by torque scaling with the cube of rotor diameter—has been substantiated as a compelling advantage for wheel-hub packaging, yet this benefit is counterbalanced by manufacturing complexities in air-gap control, stator-rotor alignment, and the thermal management challenges inherent to sandwiching the stator between two heat-generating rotor discs. The transverse-flux machine, while offering exceptional low-speed torque density for scooter applications, remains encumbered by three-dimensional magnetic circuit fabrication difficulties that limit its scalability beyond niche markets. The YASA topology emerges as the most electromagnetically efficient variant by eliminating the stator yoke entirely, though its segmented armature structure introduces assembly and control complexities that currently constrain it to premium performance segments.

TABLE 1. Comparative overview of in-wheel motor topologies across design domains.

Torque-density and efficiency figures are indicative ranges synthesized from the cited literature (Pippuri et al., 2013; Kim and Baek, 2025; Wang, X., et al., 2017; Jing et al., 2023; Guo et al., 2025); cooling and sealing assessments draw on Guo, C., et al. (2022), Wang, Q., et al. (2023), Ušakovs et al. (2023), and Lim and Kim (2014); fault-tolerance entries follow Zheng et al. (2013) and Wang, Q., et al. (2026b).

4 Mechanical integration and unsprung mass challenges

The mechanical integration of the motor within the wheel hub introduces a critical trade-off between propulsion system compactness and vehicle dynamics performance, primarily manifested through the increase in unsprung mass. The unsprung mass comprises all components that move with the wheel assembly, including the tire, rim, brake, hub bearing, and the in-wheel motor itself; augmenting this mass degrades road-holding capability by increasing tire dynamic load fluctuations and worsens ride comfort through elevated body vertical acceleration (Feng et al., 2026). Quantitative assessments indicate that the additional unsprung mass from IWMs can increase body vertical acceleration by approximately 20% and tire dynamic load by roughly 9% under typical road excitation profiles, necessitating careful suspension system recalibration or active compensation strategies (Chen, H., et al., 2022). The proximity of the motor to the brake system further exacerbates thermal coupling, as brake disc temperatures can exceed 300 °C during aggressive deceleration, radiating heat toward the motor housing and potentially degrading permanent magnet performance if the Curie temperature is approached (Li, S., et al., 2015). Magnetic gap deformation under mechanical shock and cornering loads (Dorrell et al., 2010) represents another mechanical integration challenge, requiring robust bearing arrangements and stiff motor housings to maintain air-gap uniformity within acceptable tolerances; advanced suspension-motor co-design approaches have demonstrated the capability to reduce magnetic gap deformation by over 98% through optimized knuckle and bearing configurations (Murata, 2012). The elimination of drivetrain components such as half-shafts, differentials, and transmissions simplifies the vehicle architecture and frees interior volume for battery packaging or passenger space, but the concentrated mass at each corner demands reinforced suspension arms and potentially active or semi-active damping systems to maintain dynamic performance parity with conventional drivetrain vehicles (Zhang, J., et al., 2024). Furthermore, the direct-drive configuration eliminates gear mesh losses and backlash, enabling precise torque vectoring for enhanced vehicle stability and energy recovery during regenerative braking, though this benefit must be weighed against the dynamic penalties of increased unsprung weight (He, S., et al., 2022; Hou et al., 2018). The integration of magnetic suspension systems with in-wheel motors has been proposed as a means to simultaneously address unsprung mass concerns and improve ride quality through active vibration isolation, though such systems introduce additional complexity and power consumption (Zhang, J., et al., 2024). Wheel-end packaging constraints also dictate motor outer diameter limits, typically ranging from 280 mm for compact passenger cars to over 400 mm for commercial vehicles, fundamentally constraining the electromagnetic design space and necessitating high fill-factor winding techniques to achieve target torque within available volume (Wang, Q., 2025). The bearing system in IWMs must simultaneously support vehicle weight, transmit driving torque, and accommodate cornering-induced lateral loads, leading to the adoption of large-diameter double-row angular contact bearings or custom hub-bearing units integrated directly into the motor housing (Zheng et al., 2013).

5 Thermal management and sealing reliability

In-wheel motors operate in one of the most thermally hostile and environmentally exposed locations on the vehicle, creating severe constraints on thermal management and sealing reliability. Unlike conventional traction motors mounted within the protected chassis environment, IWMs are encased inside rotating aluminum rims, surrounded by tire sidewalls and brake components, and directly exposed to water jets, mud, salt spray, and road debris at all speeds and weather conditions (Ušakovs et al., 2023). The thermal rejection challenge is compounded by the limited surface area available for heat dissipation and the restricted airflow within the wheel cavity, which rotates at road speed and disrupts natural convection patterns. Continuous operation under peak load conditions can drive winding temperatures toward insulation class limits, while transient brake heating introduces additional thermal spikes that must be managed through the motor housing (Guo, C., et al., 2022). Liquid cooling circuits integrated into the motor housing or stator core have become the preferred thermal management strategy, with cooling channel geometries optimized through computational fluid dynamics to ensure uniform temperature distribution and avoid localized hot spots that could trigger demagnetization of NdFeB magnets (Wang, Q., et al., 2023; Wang, Q., et al., 2026a). The sealing requirement for IWMs typically demands IP67 or IP68 ingress protection ratings, necessitating sophisticated rotary seals, labyrinth designs, and potting compounds for electronic components; however, the thermal expansion mismatch between dissimilar materials and the fatigue loading from wheel impacts complicate long-term sealing integrity (Ušakovs et al., 2023). Axial-flux motors offer inherently shorter thermal paths from windings to ambient, potentially improving heat dissipation efficiency, though their flat geometry can complicate the routing of cooling manifolds within the wheel envelope (Li, T., et al., 2025). Advanced thermal management strategies under investigation include phase-change materials for transient heat absorption, direct oil cooling of end windings, and thermally conductive encapsulation resins that simultaneously provide electrical insulation and environmental protection (Lim and Kim, 2014). The coupled electromagnetic-thermal behavior of in-wheel motors necessitates multi-physics simulation approaches that account for temperature-dependent material properties, including magnet remanence and winding resistance, to accurately predict performance under continuous and peak duty cycles (Jiang and Jahns, 2015). The convergence of thermal and sealing design remains a critical determinant of IWM service life and reliability, particularly in regions with extreme ambient temperatures or aggressive road salting practices. Conventional radial-flux motors benefit from mature cooling technologies including water-jacket housings and forced air ventilation, while their cylindrical geometry facilitates straightforward integration of cooling channels along the stator outer diameter (Sarlioglu et al., 2017). The development of high-temperature magnet grades, including samarium-cobalt alternatives and dysprosium-enhanced NdFeB compositions, provides additional thermal margin for IWM applications where cooling capacity is fundamentally constrained by wheel-end packaging (Li, S., et al., 2015). Real-time thermal monitoring through embedded temperature sensors and thermal observer algorithms enables adaptive current derating strategies that protect insulation and magnet integrity during prolonged hill-climbing or high-speed cruising scenarios (Winkler et al., 2025).

6 Fault-tolerant design

The safety-critical nature of automotive propulsion systems demands fault-tolerant capabilities in in-wheel motor designs, ensuring continued operation or graceful degradation in the event of partial failures. Fault-tolerant PMSM topologies typically employ multi-phase winding configurations, modular stator segments, or redundant power electronic channels to isolate faults and maintain torque production capability even when one or more phases are compromised (Zheng et al., 2013). The distributed nature of IWM architectures inherently provides propulsion redundancy, as the failure of a single wheel motor does not result in total vehicle immobilization, unlike single-motor drivetrains; however, asymmetric torque production following a fault necessitates sophisticated vehicle dynamics control algorithms to maintain directional stability (Wang, Q., et al., 2026b). From a manufacturing perspective, the adoption of group technology and modular assembly processes enables economical small-batch production of diverse IWM variants, reducing tooling costs and lead times for customized motor designs tailored to specific vehicle platforms (Jing et al., 2023). Looking forward, the integration of wide-bandgap semiconductor devices such as silicon carbide (SiC) inverters promises to improve motor drive efficiency and power density while reducing thermal management burdens, though electromagnetic compatibility challenges in the wheel environment require careful shielding design (Morya et al., 2019). Advanced materials including soft magnetic composites for stator cores and high-temperature rare-earth magnets may further enhance performance margins, while additive manufacturing techniques enable topologically optimized housing structures that minimize weight while maintaining stiffness and thermal conductivity (Wang, X., et al., 2017). The trajectory of IWM development points toward increasingly integrated “corner module” concepts, where braking, suspension, steering, and propulsion functions are consolidated into a single replaceable wheel-end assembly, potentially revolutionizing vehicle architecture and maintenance paradigms. However, the persistent challenges of unsprung mass, cost competitiveness against conventional drivetrains, and long-term durability in harsh automotive environments continue to temper the pace of mass-market adoption, suggesting that IWM technology will likely penetrate first in premium performance vehicles, urban delivery fleets, and autonomous mobility pods where its packaging and control advantages outweigh incremental costs. The standardization of IWM interfaces and the development of scalable manufacturing processes will be essential for transitioning this technology from specialized applications to mainstream automotive platforms, while continued research into active suspension compensation and advanced thermal management will address the fundamental physics constraints that currently limit widespread deployment (He, S., et al., 2022). Modular stator concepts with independent phase windings on segmented cores offer particular promise for fault-tolerant operation, as electrical isolation between segments prevents fault propagation and enables continued operation at reduced power levels (Zheng et al., 2013). The emergence of digital twin frameworks for IWM systems, combining real-time sensor data with physics-based models, enables predictive maintenance strategies that anticipate bearing wear, insulation aging, and magnet degradation before catastrophic failure occurs, thereby improving fleet reliability and reducing total cost of ownership in commercial applications.

7 Critical synthesis, research gaps, and future roadmap

Critical synthesis. Across the topologies reviewed, the outer-rotor radial-flux PMSM remains the most manufacturing-mature and thermally accessible option and is therefore the pragmatic default for mass-market NEVs (Yuan et al., 2019; Wang, Q., 2025). Axial-flux machines deliver higher torque density in a pancake geometry ideally matched to wheel-hub packaging, but this benefit is purchased with demanding air-gap control, stator structural integrity under axial magnetic attraction, and thermal management of a sandwiched stator (Kim and Baek, 2025; Ebrahimi et al., 2023). Transverse-flux machines retain exceptional low-speed torque density for niche applications, yet their three-dimensional magnetic circuits and manufacturing complexity prevent broad adoption (Pippuri et al., 2013; Shi and Jian, 2018). The yokeless and segmented armature (YASA) variant is the most electromagnetically efficient topology but is constrained by segmented-core assembly and control complexity (Wang, X., et al., 2017; Li, T., et al., 2025). The central trade-off, therefore, is not electromagnetic performance alone but the interaction of topology with the mechanical, thermal, and sealing constraints of the wheel environment.

Principal research gaps. (i) Unsprung mass: systematic quantitative links between IWM mass, suspension tuning, and ride/handling metrics across vehicle classes remain sparse, and co-design frameworks that jointly optimize motor mass and suspension are only emerging (Feng et al., 2026; Zhang, J., et al., 2024). (ii) Thermal management: experimental data on cooling effectiveness in real wheel environments, particularly under combined braking and driving heat loads, are still limited compared with simulation studies (Guo, C., et al., 2022; Ušakovs et al., 2023). (iii) Sealing and durability: long-term sealing integrity under thermal cycling, salt exposure, and wheel-impact fatigue lacks standardized accelerated-test data. (iv) Fault tolerance: most fault-tolerant IWM studies remain simulation-based, and validated redundancy architectures with quantified post-fault torque capability are needed (Zheng et al., 2013; Wang, Q., et al., 2026b). (v) Cost and manufacturability: quantitative cost models comparing IWM systems with conventional drivetrains across production volumes are scarce, and scalable manufacturing processes are not yet established.

Future research roadmap. The following structured priorities are suggested: (1) unsprung-mass reduction through lightweight housing and rotor structures, integrated with active or semi-active suspension compensation; (2) thermal management combining liquid cooling, phase-change materials, and direct oil cooling, validated under combined braking–driving duty cycles; (3) sealing and durability engineering targeting IP67/IP68 compliance with accelerated life testing; (4) fault-tolerant architectures with modular multi-phase windings and redundant power channels, demonstrated experimentally; (5) cost reduction through topology simplification and high-volume manufacturing; (6) manufacturability improvements via additive manufacturing and modular assembly; (7) standardization of IWM interfaces, safety requirements, and test protocols; and (8) integrated corner-module design consolidating propulsion, braking, suspension, and steering, with control-oriented structural co-design. Progress along these eight axes will determine whether in-wheel motor technology transitions from high-performance niches to mainstream NEV platforms.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Quynh Van Le, Thai Nguyen University of Technology, Vietnam

Reviewed by: Prasanna Ramanan, Chennai Institute of Technology, India