A comparative study on constitutive equations and artificial neural network model to predict high-temperature deformation behavior in Nitinol 60 shape memory alloy
A comparative study on constitutive equations and artificial neural network model to predict high-temperature deformation behavior in Nitinol 60 shape memory alloy
复制标题
镍钛诺60形状记忆合金高温变形行为本构方程与人工神经网络模型的比较研究
DOI：
10.1557/jmr.2015.144
复制
发表时间：
2015-06-28
影响因子：
2.7
通讯作者：
Li, Guifa
中科院分区：
文献类型：
--
作者：
Shu, Xiaoyong;Lu, Shiqiang;Li, Guifa
本文采用Arrhenius模型、多元线性模型和人工神经网络模型对锻造态镍钛诺60形状记忆合金的热变形行为进行了预测。利用650-850 ℃、应变速率0.01 ~ 1 s（-1）的等温热压缩试验获得的流变应力数据，计算了材料常数，建立了相应的本构方程。此外，通过比较预测相对误差、平均绝对相对误差和相关系数，对上述模型预测高温变形行为的能力进行了比较研究。结果表明，多重线性模型比Arrhenius型模型能更准确地预测流动特性。人工神经网络模型是更有效的，并有一个更好的预测能力，为锻造镍钛诺60合金比Arrhenius型和多线性模型。
The present study was conducted to predict the hot deformation behavior of the as-forged Nitinol 60 shape memory alloy by using the Arrhenius type, multiple-linear, and artificial neural network (ANN) models. The acquired flow stress data from isothermal hot compression tests in a temperature range of 650-850 degrees C under strain rate range of 0.01-1 s(-1) were used to calculate the material constants for establishing the corresponding constitutive equations. Furthermore, a comparative study has been made on the capability of the aforementioned models to predict the high-temperature deformation behavior by comparing the prediction relative errors, average absolute relative error, and correlation coefficient. The results show that multiple-linear model predicts the flow behavior more accurately than the Arrhenius type model. The ANN model is much more efficient and has a better prediction power for the as-forged Nitinol 60 alloy than both the Arrhenius type and multiple-linear models.