考虑不确定性的多学科设计优化方法研究综述

A Review of Multidisciplinary Design Optimization Methods Considering Uncertainties

目的研究多学科不确定性设计优化中多学科设计优化方法、不确定性建模与传递、不确定性设计优化的相关理论。方法通过研究并分析国内外相关文献,总结归纳考虑不确定性的多学科设计优化中的耦合系统解耦方法、参数和代理模型不确定性的建模方法,以及高效的不确定性传递和设计优化方法。结论系统探讨了在面对复杂多变的外界环境时,多学科设计优化对不确定性量化与传递的需求,提出多学科设计优化不仅要考虑确定性的系统,而且需要考虑由于外界环境变化导致的系统响应的不确定性。针对现有的多学科不确定性设计优化方法的理论研究,提出提高计算效率的关键在于将传统的三层嵌套循环计算框架解耦成单层循环。研究结果表明,考虑不确定性的多学科设计优化将成为复杂多学科系统设计的有力支撑,能显著提高系统的可靠性和稳健性,提高使用寿命,同时能够加快产品的更新换代设计。

The work aims to study theories on multidisciplinary design optimization,uncertainty modeling and propagation,and uncertainty design optimization in the uncertainty multidisciplinary design optimization.By studying and analyzing relevant literatures at home and abroad,the work summarized the decoupling method of coupling systems,the methods of quantifying parameter and surrogate modeling uncertainties,and efficient uncertainty propagation and design optimization methods in the uncertainty multidisciplinary design optimization.The work systematically discusses the demand of multidisciplinary design optimization for uncertainty quantification and propagation when facing complex and changeable external environment,and it proposes that multidisciplinary design optimization should not only consider the deterministic system,but also the uncertainty of system response caused by the change of external environment.After studying the existing researches of uncertainty multidisciplinary design optimization methods,it proposes that the key to improve the computational efficiency lies in decoupling the traditional three-layer nested loop calculation framework into a single-layer loop.The result shows that multidisciplinary design optimization considering uncertainty will become a strong tool for the design of complex multidisciplinary systems,which can significantly improve the reliability and robustness of the system,improve the service life and speed up the replacement design of products.