车门的高维多目标优化

High dimensional multi-objective optimization of car door

为改善某MPV车门力学性能,降低车门总成质量,针对传统优化算法在高维多目标优化时易出现解收敛性差的问题,采用基于非支配排序及参考点的NSGA-Ⅲ算法,结合试验设计及响应面模型等方法对车门进行多目标优化。研究结果表明:初始车门模型各目标的仿真分析结果与实验值误差均小于5%,一致性好;构建的响应面精度高;基于NSGA-Ⅲ算法优化获得的Pareto解分布均匀、收敛性好;优化后的车门内板模态频率及门框刚度未有明显提高,但仍满足企业要求,车门质量降低13.63%,轻量化效果较为明显。说明应用NSGA-Ⅲ算法能有效地解决传统优化高维目标收敛性差的问题,获得较为理想的结果。

In order to improve the mechanical properties of an MPV door and reduce the quality of the door assembly,aiming at the problem that the traditional optimization algorithm is prone to poor solution convergence of high-dimensional multi-objective optimization,NSGA-Ⅲalgorithm based on non-dominated ranking and reference points was used to optimize the door multi-objective by combining experimental design and response surface model.The results showed that the error between the simulation results and the experimental values of each target of the initial door model were less than 5%,and the consistency was good.The constructed response surface had high accuracy;The Pareto solutions obtained by NSGA-Ⅲalgorithm had uniform distribution and good convergence.After optimization,the first order frequency of the door and the stiffness of the door framed was not significantly improved,but still met the requirements of the enterprise.The weight of the door was reduced by 13.63%,and the lightweight effect were obvious.It showed that the application of NSGA-Ⅲalgorithm can effectively solve the problem of poor convergence of traditional optimization high-dimensional objectives,and obtain relatively ideal results.