论文部分内容阅读
利用人工神经网络(ANN)模型来建立钛合金本构关系以及TB8钛合金热压缩试验数据,采用误差反向传播(Error Back-Propagation Networks)算法模拟了流变应力。结果表明:TB8钛合金在较宽泛温度948~1073K,应变速率在0.001~10s-1含有两个节点数为18的隐含层BP神经网络模型,这为研究TB8钛合金高温塑性变形行为提供了依据。对不同相区不同变形机制的TB8钛合金应力应变行为进行精确表征,训练阶段,最大绝对相对误差3.78%。在验证阶段,最大绝对相对误差4.06%,且大部分相对误差点分布在±2%的范围内,实现了较高的精度。
The artificial neural network (ANN) model was used to establish the constitutive relation of titanium alloy and the thermal compression test data of TB8 titanium alloy. The error back-propagation algorithm was used to simulate the flow stress. The results show that the TB8 titanium alloy contains two hidden BP neural network models with a node number of 18 at a wide range of 948 ~ 1073K and a strain rate of 0.001 ~ 10s-1, which provides a basis for studying the high temperature plastic deformation behavior of TB8 titanium alloy in accordance with. The stress-strain behavior of TB8 titanium alloy with different deformation mechanisms in different phase regions was accurately characterized. During the training phase, the maximum absolute relative error was 3.78%. During the verification phase, the maximum absolute relative error was 4.06%, and most of the relative error points were distributed within the range of ± 2%, achieving a higher accuracy.