航空学报 > 2025, Vol. 46 Issue (18): 231817-231817   doi: 10.7527/S1000-6893.2025.31817

面向引力波探测航天器多物理场噪声抑制的组件布局优化

方子若1,2, 汤宁标1,2, 刘野1,2, 蔡志鸣1, 陈雯1,2, 朱振才1,2, 侍行剑1,2()   

  1. 1.中国科学院微小卫星创新研究院,上海 201304
    2.中国科学院大学,北京 100049
  • 收稿日期:2025-01-16 修回日期:2025-02-09 接受日期:2025-03-12 出版日期:2025-09-25 发布日期:2025-03-28
  • 通讯作者: 侍行剑 E-mail:shixj@microsate.com
  • 基金资助:
    国家重点研发计划(2020YFC2200901)

Component layout design optimization for multi-physical field noise suppression in gravitational wave detection spacecraft

Ziruo FANG1,2, Ningbiao TANG1,2, Ye LIU1,2, Zhiming CAI1, Wen CHEN1,2, Zhencai ZHU1,2, Xingjian SHI1,2()   

  1. 1.Innovation Academy for Microsatellites of Chinese Academy of Sciences,Shanghai 201304,China
    2.University of Chinese Academy of Sciences,Beijing 100049,China
  • Received:2025-01-16 Revised:2025-02-09 Accepted:2025-03-12 Online:2025-09-25 Published:2025-03-28
  • Contact: Xingjian SHI E-mail:shixj@microsate.com
  • Supported by:
    National Key Research and Development Program of China(2020YFC2200901)

摘要:

空间引力波探测任务对航天器核心区域的环境洁净度提出了极高的要求,为此,提出一种双层序列优化方法(BSOA),解决航天器组件布局设计(SCLD)问题以实现电磁力和自引力噪声的有效抑制。SCLD是一个典型的混合整数规划问题,BSOA方法将其进一步建模为双层优化问题进行求解,上层优化定义为整数非线性规划问题,确定组件的方向和区域;下层优化定义为实数非线性规划问题,优化组件在选定区域内的具体位置。通过引入反馈迭代机制,下层优化的结果能够反作用于上层决策,实现布局方案的渐进优化。在双层序列优化框架内,采用精英遗传算法实现上层问题的全局优化,并结合差分进化算法完成下层问题的局部搜索。针对优化过程中的多种技术挑战,提出混合编码策略以满足进化算法的编码需求,区域划分策略以实现安装位置的离散化处理,以及碰撞检测方法以识别组件几何约束违反情况。实验结果表明,该方法在复杂多约束条件下可高效求解布局设计问题,生成符合科学任务要求的布局方案,并在均值和标准差等性能指标上显著优于传统单阶段优化方法和双阶段优化方法,具有重要的应用潜力和拓展价值,为未来的引力波探测任务奠定了技术基础。

关键词: 引力波探测, 布局设计, 混合整数规划, 非线性规划, 双层优化

Abstract:

The space-based gravitational wave detection mission places extremely high demands on the cleanliness of the core environment within spacecraft. To meet these demands, a Bilevel Sequential Optimization Approach (BSOA) is proposed to solve the Spacecraft Component Layout Design (SCLD) problem, aiming to effectively suppress electromagnetic forces and self-gravity noise. SCLD is a typical mixed-integer programming problem, and the BSOA method is further formulated as a bilevel optimization problem for solution. The upper-level optimization is defined as an integer nonlinear programming problem to determine the orientation and region of components, while the lower-level optimization is defined as a real-valued nonlinear programming problem to optimize the placement of components within the selected region. By introducing a feedback iterative mechanism, the results of the lower-level optimization influence upper-level decisions, enabling progressive optimization of the layout scheme. Within the bilevel sequential optimization framework, an elite genetic algorithm is employed for global optimization of the upper-level problem using, while the lower-level problem is locally searched using a differential evolution algorithm. To address various technical challenges in the optimization process, a hybrid encoding strategy is proposed to meet the coding requirements of evolutionary algorithms, a regional division strategy is introduced to discretize installation positions, and a collision detection approach is implemented to identify violations of geometric constraints among components. Experimental results demonstrate that the proposed approach efficiently solves layout design problems under complex multi-constraint conditions, generates layout schemes that meet scientific mission requirements, and significantly outperforms traditional single-stage and two-stage optimization methods in terms of performance indicators such as mean and standard deviation. This approach exhibits significant application potential and extensibility, laying a technical foundation for future gravitational wave detection missions.

Key words: gravitational wave detection, layout design, mixed-integer programming, nonlinear programming, bilevel optimization

中图分类号: