车身用高强钢的回弹实验与分析

目的研究和分析身用高强钢在冲压过程中的回弹行为与规律。方法基于Wagoner的拉延弯曲实验模型,采用拉延弯曲(Draw-Bend)实验机,针对590 MPa级冷轧双相钢和420 MPa级低合金高强钢两个常用的典型车身用高强钢材料开展回弹实验研究,通过不同弯曲半径模拟不同冲压模具圆角半径,同时以不同张紧力模拟冲压压边力,讨论了两种材料的冲压成形及回弹性能与控制。结果两类钢的回弹量均随着冲压模具圆角半径的增大而减小,随着压边力的增大而减小,双相钢相比低合金能产生更大的回弹,厚板随冲压条件变化的回弹波动一般低于薄板。结论实验研究结果对高强钢在实际冲压过程中的回弹控制有着较强的指导意义。     

Method of closed loop springback compensation for incremental sheet forming process

The closed loop control model was built up for compensating the springback and enhancing the work piece precision. A coupled closed loop algorithm and a finite element method were developed to simulate and correct the springback of incremental sheet forming. A three-dimensional finite element model was established for simulation of springback in incremental sheet forming process. The closed loop algorithm of trajectory profile for the incremental sheet forming based on the wavelet transform combined with fast Fourier transform was constructed. The profile of processing tool path of shallow dishing with spherical surface was designed on the basis of the profile correction algorithm. The result shows that the algorithm can predict an ideal profile of processing track, and the springback error of incremental sheet forming is eliminated effectively. It has good convergence efficiency, and can improve the workpiece dimensional accuracy greatly.     

基于神经网络的拉弯回弹预测技术研究

以型材拉弯为对象,研究了利用神经网络预测拉弯回弹的关键技术,建立了拉弯回弹预测的人工神经网络模型,有效地解决拉弯控制参数和回弹量之间的非线性映射问题;并利用Matlab神经网络工具箱加以实现,取得了理想的预测精度,验证了预测模型的有效性。该研究推动了人工智能技术在钣金精密成形中的应用。