复杂系统尺寸公差的多目标多学科协同优化决策研究

In spite of substantial previously reported works in the literature on manufacturing tolerance assignments for a single component or multiple (system of) components or a tolerance chain, the effect of tolerances at the system level cannot be directly and easily obtained or estimated. This is because the complexity of the tolerance assignment problem grows rapidly and in essence the problem becomes intractable as the number of components (or subsystems) and corresponding couplings increases. Meanwhile, improvements in the precision of current mechatronic systems make the system performance more vulnerable to the uncertainty in manufacturing tolerances. In this context, the manufacturing tolerance assignments cannot be performed based on ad-hoc methods or according to the designers' experience or traditional methods anymore. The proposed research will develop a new multi-objective multi-disciplinary collaborative optimization method to address the problem of manufacturing tolerance assignments for complex engineering systems. With a typical application in automobile engines, the objective of this project is to improve the system performance, while limiting the increase in the manufacturing cost, by judiciously applying optimization methods to effectively assign and control manufacturing tolerances. The major tasks of the proposed project will include: 1) to establish a structured analysis model between manufacturing tolerances and output performance for each subsystem in the system; 2) to clearly define key manufacturing tolerances, structure variables, and uncertain parameters, and formulate a structured multi-disciplinary optimization problem; 3) to propose and implement a hybrid collaborative optimization method based on multi-objective sensitivity analysis and optimization algorithms; 4) to verify and validate the proposed methods and models and corresponding optimal manufacturing tolerance assignment solutions.

虽然对于单个零部件或公差环的公差设计已得到广泛研究，但是由于零部件和子系统数量的增加以及多个零部件的互相耦合，尺寸公差对复杂机电系统性能的影响已经无法直观获得。同时，现代机电系统的精密程度的提高导致尺寸公差能够极大地影响系统设计性能。因此，复杂系统的尺寸公差设计难以再依靠设计者的经验或传统的公差设计方法解决。本课题主要通过研究和应用多目标多学科协同优化方法来解决复杂非线性系统的尺寸公差优化与决策问题。以车用内燃机为主要应用背景，本课题的目标是运用有效的优化方法实现对尺寸公差的差异化控制，在不提高系统制造成本的前提下提高系统性能，内容包括：1）建立对应于各个子系统的尺寸公差和系统性能之间的多学科结构化分析模型；2）确定系统关键尺寸和结构变量、以及系统不确定性因素，形成结构化的多学科优化问题；3）提出并实现以多目标敏感度分析和优化方法为基础的协同优化方法；4）通过实验验证优化方法和结果。

虽然对于单个零部件或公差环的公差设计已得到广泛研究，但是由于零部件和子系统数量的增加以及多个零部件的互相耦合，尺寸公差对复杂机电系统性能的影响已经无法直观获得。同时，现代机电系统的精密程度的提高导致尺寸公差能够极大地影响系统设计性能。因此，复杂系统的尺寸公差设计难以再依靠设计者的经验或传统的公差设计方法解决。本课题主要通过研究和应用多目标多学科协同优化方法来解决复杂非线性系统的尺寸公差优化与决策问题。以车用内燃机为主要应用背景，运用有效的优化方法实现对尺寸公差的差异化控制，在不提高系统制造成本的前提下提高系统性能，内容包括：1）建立对应于各个子系统的尺寸公差和系统性能之间的多学科结构化分析模型；2）确定系统关键尺寸和结构变量、以及系统不确定性因素，形成结构化的多学科优化问题；3）提出并实现以多目标敏感度分析和优化方法为基础的协同优化方法；4）通过实验验证优化方法和结果。