EI、Scopus 收录
中文核心期刊

拉延成形多目标序列响应面法优化设计方法

Multi-objective optimization for sheet metal formnig of drawing with successive response surface method

  • 摘要: 为了克服常规响应面法在整个设计空间进行逼近导致精度低和传统的单目标优化设计,只能针对其中的一个目标进行优化的缺陷,提出了一种多目标序列响应面法,优化设计拉延筋几何参数的方法. 该方法通过移动、缩放等方式在设计空间中不断更新兴趣域,在不同的兴趣域中将实验设计、响应面法和多目标粒子群优化算法相结合,获得了一组最小化起皱、拉裂缺陷的等效拉延阻力非劣解. 利用最小距离选解法从非劣解集中挑选出一组成形效果最好的解, 并以此解作为下一迭代步兴趣域的中心,直到收敛至一组最优的等效拉延约束阻力. 以最优等效拉延阻力为约束条件,利用等效拉延筋阻力模型结合遗传算法对拉延筋的真实几何参数进行设计. 优化设计的拉延筋几何参数提高了板料的成形性能. 算例表明,该方法具有较高的精度和较强的工程实用性.

     

    Abstract: This paper describes a multi-objective optimization ofdrawbead geometrical parameters of automotive panel using successiveresponse surface method. The existing response surface method has a lowprecision in the entire design space, and conventional single-objectiveoptimization design does not provide for multi-objective conditions and hasbeen used to optimize only a single objective. The present method updatesthe region of interest (ROI) in the design space by panning and zooming. Ineach ROI, a set of pareto optimal solutions of equivalent drawbead restraintforce,are obtained to minimize the defects of crack and wrinkle by combiningdesign of experiments, successive response surface method andmulti-objective particle swarm optimization. Through the minimum distancemethod, a solution with best forming approach is obtained from the paretooptimal solutions, and it is used as the center point of the next region ofinterest. The optimal drawbead restraint force will be obtained by theiterative procedure. The optimal design geometric parameters of drawbead canbe obtained using the optimal drawbead restraint force and geneticalgorithm. These optimal parameters can be efficiently used to improve theforming properties of sheet. Numerical examples indicate that the presentmethod has higher precision and practicability compared with the existingtechniques.

     

/

返回文章
返回