基于蒙特卡洛-自适应差分进化算法的飞机容差分配多目标优化方法  

Multi-objective optimization method for aircraft tolerance allocation based on Monte Carlo-adaptive differential evolution algorithm

作　　者：荆涛 田锡天[1] JING Tao;TIAN Xitian(School of Mechanical Engineering,Northwestern Polytechnical University,Xi'an 710072,China)

机构地区：[1]西北工业大学机电学院,西安710072

出　　处：《航空学报》2022年第3期569-580,共12页Acta Aeronautica et Astronautica Sinica

摘　　要：针对飞机复杂组件装配过程中,由于装配偏差的非线性叠加导致通用容差分配模型优化方法存在精度误差和效率低下的问题,提出一种容差分配模型多目标优化方法。首先,分别面向制造成本、装配性能、质量损失构建基于替代工艺的容差分配模型,并采用最优化理论建立多目标优化模型。然后,提出基于蒙特卡洛-自适应差分进化算法的多目标优化策略,其中基于装配偏差传递的图树模型提取偏差的非线性关系模型,结合提取的非线性关系模型采用蒙特卡洛方法对初始样本进行前处理,以提高初始种群的多样性;自适应差分进化算法中的突变阶段采用Lévy flight概率分布进行改进提高全局搜索效率和鲁棒性。通过飞机登机门组件验证容差分配方法,结果表明所提方法能够更加精确、高效的优化容差分配。相比于初始容差分配,优化容差分配结果使得制造成本降低了21.78%,质量损失降低了11.12%,装配性能提高了12.28%。The optimization method of the general tolerance allocation model has precision errors and low efficiency due to nonlinear stack-up of assembly deviations in the assembly process of complicated aircraft components.In this paper,a novel multi-objective optimization method of tolerance allocation model is proposed.First,a tolerance allocation model for the manufacturing cost,assembly performance,and quality loss of the aircraft is constructed based on alternative processes.A multi-objective optimization model is established using the optimization theory.Then,a multi-objective optimization strategy is proposed based on the Monte Carlo-adaptive differential evolution algorithm,in which the non-linear relationship model of deviation is extracted based on the graph tree model of assembly deviation propagation.Based on the non-linear relationship model extracted,the initial samples are pre-processed using the Monte Carle method to improve the diversity of the initial population.In the mutation stage in the adaptive differential evolution algorithm,the Lévy flight probability distribution is used to improve the global search efficiency and robustness.The tolerance allocation method is verified by assembly of the aircraft boarding gate components.The results demonstrate that the proposed method can optimize tolerance allocation more accurately and efficiently.Compared with the initial tolerance allocation,the optimized tolerance allocation results in reduction of the manufacturing cost by 21.78%and the quality loss by 11.12%,and increase of assembly performance by 12.28%.

关 键 词：飞机装配 容差分配 多目标优化 非线性提取 蒙特卡洛-自适应差分进化算法 

分 类 号：V262.41[航空宇航科学与技术—航空宇航制造工程]