基于Kriging模型的翻边成形参数优化

Research on parameter optimization in flanging process based on Kriging metamodel

基于数值模拟的参数优化是一个反复迭代的过程,往往要调用有限元模型进行反复迭代,计算周期较长。同时,在翻边成形数值模拟这样的非线性动态大变形分析中,成形优化问题应归为非线性规划问题,其目标函数和约束函数都不能显式的表达,目标函数和约束函数的导数绝大多数为严重的不连续。引用地质学中Kriging模型理论,建立工程问题中的Kriging模型,并介绍了Kriging模型的建立方法;利用对非线性函数的拟合,验证Kriging模型能很好的适应高度非线性的函数;将Kriging模型应用到翻边成形中,进行工艺参数的优化。结果表明,基于Kriging模型的优化,对翻边成形能取得满意的结果。

Parameter optimization in flanging based on numerical simulation is an iterative procedure usually adopting finite element model and the simulation time is long.At the same time,it is a nonlinear problem with the objective and constrain functions as implicit expression,and the derivatives being discontinuous.Making use of the Kriging theory in geology,the Kriging metamodel in engineering problems was established to save time.In order to prove accuracy and efficiency of Kriging method,the nonlinear functions as test functions were implemented.At the same time,the practical nonlinear engineering problems such as flanging were also optimized successfully by proposed method.The results prove that the Kriging model is an effective method for nonlinear engineering problem in practice.