考虑高耗时约束的全电推进卫星多学科优化

Multidisciplinary Design Optimization for All-Electric Propulsion Satellite Considering Computationally Expensive Constraints

为提高卫星系统整体性能与优化设计效率,本文采用多学科设计优化(MDO)方法进行全电推进卫星总体参数优化。主要考虑轨道转移、位置保持、空间环境、供配电、结构及质量六个学科,以整星质量最小为优化目标,考虑轨道转移时间等约束条件,建立了全电推进卫星MDO模型。提出一种基于增广拉格朗日乘子法的高效全局优化方法(ALM-EGO)以快速求解卫星MDO问题。标准数值算例对比研究表明,对于处理高耗时约束优化问题,ALM-EGO方法在全局收敛性与优化效率方面具有一定的优势。最后,采用ALM-EGO求解全电推进卫星MDO问题,优化后的全电推进卫星设计方案满足各类工程设计约束,实现整星减重161.09 kg,从而验证了本文所建立模型的合理性和ALM-EGO方法的有效性。

The multidisciplinary design optimization（MDO） method is adopted to optimize the main parameters of an all-electric propulsion satellite in order to improve the satellite＇s overall performance and design efficiency. A MDO model of an all-electric propulsion satellite is built by considering the parameters of several disciplines,i. e. orbit transfer,position-keeping,space environment,power,structure,and mass,to minimize the mass of the satellite subject to the specific constraints（e. g.,orbit transfer time）. An augmented Lagrange multiplier based efficient global optimization（ALM-EGO） method is proposed to solve the satellite MDO problems efficiently. Several numerical benchmark problems are used to test the performance of the proposed method. The comparison results show that the ALM-EGO outperforms the competitive methods in efficiency and convergence when solving the problems with computationally expensive constraints.Finally,the ALM-EGO is used to solve the all-electric propulsion satellite MDO problem. After the optimization,the mass of the satellite is reduced by 161. 09 kg and all constraints are satisfied,demonstrating the effectiveness of the ALM-EGO and models built in this paper.