Scholarlink

<p>Reluctance actuators (RAs) have been widely adopted in high-acceleration precision electromechanical systems such as lithography reticle stages due to their high thrust density and compact configuration. For system-level RA design, a multi-physics methodology balancing computational flexibility and predictive accuracy is essential to address bidirectional actuator-structure interactions. Current RA design methodologies predominantly focus only on electromagnetic performance optimization while neglecting thermal constraints in precision systems, particularly the thermal impact on external structures during continuous operation. To address this challenge, this paper proposes a bidirectional co-optimization framework integrating data-driven and physics-based approaches. Firstly, a high-fidelity structure-thermal surrogate model, neural network-LPTN hybrid network (NNLN), that synergizes data-driven and physics-based modeling approaches is proposed. This hybrid architecture achieves high-precision scalable thermal modeling while preserving physical interpretability, enabling dynamic coupling with external thermal models and effectively resolving the inherent trade-off between accuracy and scalability in conventional approaches, demonstrating over 50% higher accuracy in thermal prediction compared to conventional LPTN. Secondly, an observer-based electromagnetic-thermal co-optimization algorithm is developed, integrating electromagnetic performance and thermal behavior analysis within a parameter traversal framework to achieve efficient global optimization under multi-objective constraints, thereby supporting actuator-system co-iterative design. Validation through multiple design cases including lithography reticle stages demonstrates that the proposed methodology significantly enhances optimization efficiency and model compatibility under demanding requirements of high thrust (≥ 5000&#xa0;N), high acceleration (≥ 100&#xa0;g), millisecond-level dynamic response, and multi-component temperature/dimension constraints. This work establishes a collaborative design paradigm with enhanced precision, flexibility, and scalability for high-acceleration precision motion systems employing RAs.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A coupled structure-electromagnetism-thermal optimal design method of reluctance actuator in ultra-high acceleration lithography reticle stage

  • 期刊论文

  • 发表日期：

    2025年11月26日

查看全文
  • Yufei Zhao,
  • Weinan Ye,
  • Leijie Wang,
  • Xin Li,
  • Yu Zhu,
  • Ming Zhang

摘要

Reluctance actuators (RAs) have been widely adopted in high-acceleration precision electromechanical systems such as lithography reticle stages due to their high thrust density and compact configuration. For system-level RA design, a multi-physics methodology balancing computational flexibility and predictive accuracy is essential to address bidirectional actuator-structure interactions. Current RA design methodologies predominantly focus only on electromagnetic performance optimization while neglecting thermal constraints in precision systems, particularly the thermal impact on external structures during continuous operation. To address this challenge, this paper proposes a bidirectional co-optimization framework integrating data-driven and physics-based approaches. Firstly, a high-fidelity structure-thermal surrogate model, neural network-LPTN hybrid network (NNLN), that synergizes data-driven and physics-based modeling approaches is proposed. This hybrid architecture achieves high-precision scalable thermal modeling while preserving physical interpretability, enabling dynamic coupling with external thermal models and effectively resolving the inherent trade-off between accuracy and scalability in conventional approaches, demonstrating over 50% higher accuracy in thermal prediction compared to conventional LPTN. Secondly, an observer-based electromagnetic-thermal co-optimization algorithm is developed, integrating electromagnetic performance and thermal behavior analysis within a parameter traversal framework to achieve efficient global optimization under multi-objective constraints, thereby supporting actuator-system co-iterative design. Validation through multiple design cases including lithography reticle stages demonstrates that the proposed methodology significantly enhances optimization efficiency and model compatibility under demanding requirements of high thrust (≥ 5000 N), high acceleration (≥ 100 g), millisecond-level dynamic response, and multi-component temperature/dimension constraints. This work establishes a collaborative design paradigm with enhanced precision, flexibility, and scalability for high-acceleration precision motion systems employing RAs.