EI、Scopus 收录
中文核心期刊

基于子域分解的集中力扩散舱段并行协同优化方法

SUBDOMAIN DECOMPOSITION-BASED DISTRIBUTED COOPERATIVE COEVOLUTION METHOD FOR OPTIMIZATION DESIGN OF THE CONCENTRATED-FORCE DIFFUSION COMPONENT

  • 摘要: 针对大尺寸和大载荷给重型运载火箭集中力扩散舱段带来的结构承载效率低、载荷扩散能力差、轻质优化难度大等难题, 提出了基于子域分解的集中力扩散舱段并行协同进化设计方法. 设计了集中力扩散舱段多层级变刚度结构方案, 实现主扩散区和非主扩散区结构刚度的合理调控. 采用线性静力分析结构应力分布, 并结合欧拉公式快速评估构件临界失稳应力, 进而建立了集中力扩散舱段的高维多约束轻质优化模型. 为高效求解该大规模优化问题, 借鉴“分而治之, 协同合作”思想, 通过对集中力扩散舱段结构子域合理分解, 将大规模全局优化问题重构为多个小规模优化子问题, 提出了基于优化解合作的并行协同模拟退火算法框架, 据此并行协同求解重构后的多个子优化问题, 实现了优化效率和精度的同步提升. 在49维设计变量、11个性能约束的重型运载火箭集中力扩散舱段轻量化设计应用中, 本文方法获得了相比初始设计减重898.98 kg、集中载荷扩散能力提升54.6%的优化方案, 相比文献工作, 不仅实现了结构进一步减重, 而且迭代次数降低了60%以上, 充分证明了所提方法的先进性和工程适用性. 研究成果不仅可直接服务于重型运载火箭集中力扩散舱段研制, 也可应用于其他类似复杂结构设计, 具有一定的理论意义和工程应用价值.

     

    Abstract: A subdomain decomposition-based distributed cooperative coevolution method is proposed in this paper to address the challenges of low load-carrying efficiency, poor load-diffusion performance, and difficulties in lightweight optimization design, which are posed by large dimensions and heavy loads in the concentrated-force diffusion component of the heavy-lift launch vehicle. Specifically, a hierarchical variable-stiffness structural configuration is designed to tailor stiffness properties of the main and auxiliary load-diffusion zones in the concentrated-force diffusion component. Subsequently, linear static analysis is employed to examine stress distribution, while Euler's formula is utilized to evaluate the critical buckling stress of components rapidly. This, in turn, enables the establishment of a high-dimensional, multi-constraint lightweight optimization problem. To efficiently solve this large-scale global optimization problem, guided by the principle of "divide and conquer, cooperate and collaborate", the global problem is first reconstructed into multiple smaller-scale subproblems through rational subdomain decomposition. Thereafter, a distributed cooperative coevolution simulated annealing algorithm framework is proposed, which solves these subproblems concurrently, thereby achieving a simultaneous improvement in optimization efficiency and accuracy. In application to the lightweight design of the concentrated-force diffusion component in the heavy-lift launch vehicle, the proposed method yields a solution that reduces weight by 898.98 kg and improves concentrated-load diffusion capacity by 54.6% compared to the initial design. Furthermore, and significantly, it not only achieves further structural weight reduction but also reduces the iterations by over 60% when benchmarked against existing literature, thus fully demonstrating its advancement and engineering applicability. Ultimately, the research findings can directly support the development of such critical launch vehicle components and are also extendable to other similar complex structural designs, thereby offering considerable theoretical significance and practical engineering value.

     

/

返回文章
返回