摘要
在分析传统单向拉伸实验的缺点的基础上 ,提出了一种应用神经网络、有限元模拟和实验相结合的板材塑性参数的识别模型。采用了具有双向应力状态的板料拉伸试件 ,利用 BP神经网络实现试件拉伸变形过程中载荷、位移曲线与材料性能参数之间的映射关系 ,从而可以得到双向应力状态和大变形条件下的材料参数。研究中对 0 8Al、LY1 2
Existing sophisticated methods for determination of constitutive parameters are still based on uni axial tension tests and so are somewhat deficient in the accuracy needed in simulating sheet metal forming operation. We present a new parameter identification method for determining the material parameters of sheet metals under bi axial stress state and large deformation using FEM (finite element method) and tests. We design a BP (Back Propagation) neural network for mapping between measured responses (load, displacement) and material parameters. Aluminum and steel sheets were tested to evaluate our new approach. Figs.3 through 6 clearly show that simulation results calculated by our new approach are closer to test results than those calculated by using material parameters by uni axial tension test.
出处
《西北工业大学学报》
EI
CAS
CSCD
北大核心
2001年第2期190-194,共5页
Journal of Northwestern Polytechnical University
基金
国防科技项目! (96 J18)
国家自然科学基金资助 !(5 980 5 0 15 )
关键词
神经网络
参数识别
有限元
材料参数
塑性成形
板料拉伸
neural network, parameter identification, FEM(finite element method), material parameter
