脉冲电流条件下Ni箔流动应力及其尺寸效应的神经网络模型  

A Neural Network Model for Flow Stress and Size Effects of Ni Foils under Pulsed Current

作　　者：李超[1] 槐宝 王继伟[1] 赵闪[1] 

机构地区：[1]哈尔滨理工大学材料科学与工程学院,哈尔滨150040

出　　处：《精密成形工程》2017年第5期134-138,共5页Journal of Netshape Forming Engineering

基　　金：国家自然科学基金(51305110)

摘　　要：目的揭示脉冲电流对金属箔材室温流动应力及其尺寸效应的影响规律,构建脉冲条件下Ni箔室温塑性流动的本构模型。方法采用单向拉伸试验,测试不同试样尺寸的Ni箔在不同电流密度条件下的单向拉伸性能。根据本构方程的输入、输出参数,设计BP神经网络结构,并利用试验数据对其进行训练及预测精度检验。结果单向拉伸试验结果表明,箔材厚度与晶粒尺寸之比及变形过程中引入的脉冲电流均会对材料的流动应力产生显著影响,且其影响规律高度复杂。脉冲电流条件下,构建的金属箔材流动应力尺寸效应神经网络模型,其预测相对误差控制在6%以内,相关系数R达到0.99。结论该神经网络模型准确描述了脉冲电流条件下金属Ni箔流动应力尺寸效应,为该材料塑性变形过程分析及合理制定工艺参数奠定了理论基础。The paper aims to reveal influences of pulse current on the flow stress and its size-effect of metal foil at room temperature, and establish a constitutive model of plastic flow of Ni foil at room temperature under pulsed condition. Uniaxial tensile test was adopted to test the uniaxial tensile properties of Ni foil with different sample sizes under different current density. According to the input and output parameters of the constitutive equation, the BP neural network structure was designed, and its training and prediction accuracy verification were carried out with the experimental data. The ratio of foil thickness to grain size and the pulse current introduced during the deformation process would have a significant effect on the flow stress of the material, and its influence law was highly complicated. The prediction of the relative error was less than 6% and the correlation coefficient R up to 0.99, based on the neural network model of the metal foil under the pulse current condition. The neural network model accurately describes the size effect of metal Ni foil flow stress under pulsed current condition and lays a theoretical foundation for the analysis of plastic deformation process and rationally setting of process parameters.

关 键 词：金属箔材 微塑性成形 尺寸效应 BP神经网络 

分 类 号：TG146.15[一般工业技术—材料科学与工程]