[PDF全文下载] [全文在线阅读]
文章编号：1004-0609(2010)S1-s0468-05
TC17钛合金片层组织动态球化的神经网络预测模型
王凯旋1, 2，曾卫东1，赵永庆2，徐 斌3，田 飞1，朱艳春1，张尧武1，周义刚1
(1. 西北工业大学 材料学院，西安 710072；2. 西北有色金属研究院，西安 710016；
3. 宝钢公司特殊钢分公司，上海 200940)
摘 要: 在Gleeble−1500热模拟试验机上通过热压缩试验研究具有初始片层组织的TC17钛合金在变形温度为780~860 ℃、应变速率为0.001~10 s−1、变形量为15%~75%范围内的组织演变，定量分析热变形参数对片层组织动态球化过程的影响。采用结合贝叶斯归一化算法的BP人工神经网络，建立TC17钛合金片层组织动态球化演变的预测模型，误差分析表明模型精度较好。
ANN model for prediction of dynamic globularization in
TC17 titanium alloy
WANG Kai-xuan1, 2, ZENG Wei-dong1, ZHAO Yong-qing2, XU Bin3, TIAN Fei1,
ZHU Yan-chun1, ZHANG Yao-wu1, ZHOU Yi-gang1
（1. School of Materials Science and Engineering, Northwestern Polytechnical University, Xi’an 710072, China;
2. Northwest Institute for Nonferrous Metal Research, Xi’an 710016, China;
3. Special Steel Branch, Baoshan Iron & Steel Co., Ltd., Shanghai 200940, China）
Abstract: The isothermal hot compression on Gleeble−1500 system in the temperature range of 780−860 ℃ and strain rate range of 0.001−10 s−1, was used to study the kinetics of microstructural evolution of TC17 titanium alloy with initial lamellar microstructure. The quantitative results show that dynamic globularization kinetics and kinetics rate are sensitive to deformation conditions. A model for predicting volume fraction of dynamic globularization was built based on the above quantitative results and BP artificial neural network (ANN) combined with Bayesian regularization. The well coincidence of the predicted results with measured ones shows the feasibility of the model.
Key words: TC17 titanium alloy; lamellar microstructure; dynamic globularization; kinetics; artificial neural network
