.png){kind=icon}
.png){kind=icon}
Return
System-dynamics-oriented design of electromagnetic active suspension via physics-informed modeling
.png){kind=icon}
PRE
AI
J
H
L
Z
S
X
DOI：10.1016/j.ymssp.2026.114834.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Electromagnetic active suspension (EAS) performance is fundamentally governed by actuator dynamics, where the thrust saturation and ripple of permanent magnet linear synchronous motors (PMLSMs) directly affect the ride comfort under random road excitations. In conventional FEA-driven actuator design, the pathway linking electromagnetic design variables to system-level vibration responses is not explicitly revealed, which renders performance-oriented optimization computationally prohibitive. This study proposes a physics-informed modeling and optimization framework that establishes an explicit knowledge pathway from suspension dynamics to electromagnetic actuator design. An analytical Optimal Equivalent Radius-Maxwell Stress Tensor (OMST) formulation is developed to characterize the electromagnetic thrust, and embedded together with Maxwell equations into a physics-informed neural operator (PIMNO) built upon an enhanced Fourier Neural Operator. This enables accurate magnetic field and thrust prediction with substantially reduced data requirements. A MORIME scheme is further introduced to synergize with the PIMNO-OMST model for efficient design exploration. More importantly, device performance requirements are derived directly from system-level vibration simulations using a dual-condition robust weighting strategy, allowing actuator optimization to be guided by vibration and signal-domain indices rather than purely electromagnetic criteria. Experimental results demonstrate that the designed actuator yields suspension vibration responses, power spectral density characteristics, and ride comfort performance nearly identical to those obtained using FEA-driven optimization, while reducing computational cost by over 99%. The proposed framework provides a computationally efficient and physically consistent method for system-dynamics-oriented force performance design in EAS, bridging electromagnetic field-to-force modeling and mechanical system vibration analysis.
Keywords:
Electromagnetic active suspension
System-dynamics-oriented design
System dynamics
Maxwell stress tensor
Physics-informed neural operator
Multi-objective optimization
Journal
IF:
8.9
Papers:
1.2W
Citations:
6.6W
Organization
Cited Papers