CAJ | 학술논문

使用有限元分析软件ABAQUS，建立了有效的不锈钢管材弯曲回弹有限元仿真模型，运用虚拟正交试验方法，研究了弯曲角度、直径、壁厚、助推速度以及弯曲模与管壁之间的摩擦系数，对弯曲回弹的影响，最后分别用线性回归和神经网络，建立回弹角度与上述参数间的定量关系。结果表明，回弹角随着弯曲角度的增大而增大，且呈现出明显的线性关系；与回弹角度相比，直径、壁厚、助推速度以及摩擦系数对回弹角的影响较小；神经网络更能准确地对回弹角与上述参数间的关系进行预测。
사용유한원분석연건ABAQUS，건립료유효적불수강관재만곡회탄유한원방진모형，운용허의정교시험방법，연구료만곡각도、직경、벽후、조추속도이급만곡모여관벽지간적마찰계수，대만곡회탄적영향，최후분별용선성회귀화신경망락，건립회탄각도여상술삼수간적정량관계。결과표명，회탄각수착만곡각도적증대이증대，차정현출명현적선성관계；여회탄각도상비，직경、벽후、조추속도이급마찰계수대회탄각적영향교소；신경망락경능준학지대회탄각여상술삼수간적관계진행예측。
FEM software is used to build an ef-fective simulation model on springback of stainless steel tube bending forming.The virtual orthogonal experiment method was used to research the fac-tors those influence the springback,such as ben- ding angle,diameter,tube thickness,assistant ve-locity and friction factor between bending die and outboard of tube.Then,linear regression analysis and artificial neural networks were employed to es-tablish the mathematic relationship between spri-ngback angle and those factors.The results proved that springback angle is linear increase along with increase of bending angle obviously,compared with this,others have weekly influence;artificial neural networks was more precision.