基于模糊神经网络的Ti40合金高温本构关系模型  

High temperature constitutive relationship model of Ti40 alloy based on fuzzy-neural network

作　　者：韩远飞[1] 曾卫东[1] 赵永庆[2] 舒滢[2] 张学敏[1] 周义刚[1] 

机构地区：[1]西北工业大学凝固技术国家重点实验室,西安710072 [2]西北有色金属研究院,西安710016

出　　处：《中国有色金属学报》2010年第10期1971-1976,共6页The Chinese Journal of Nonferrous Metals

基　　金：国家重点基础研究发展计划资助项目(2007CB613807);新世纪优秀人才支持计划资助项目(NCET-07-0696);凝固技术国家重点实验室开放课题资助项目(35-TP-2009)

摘　　要：采用Gleeble1500热模拟实验机对Ti40合金在变形温度900～1100℃,应变速率0.01～10s-1,最大变形程度约60%的条件下的高温流动应力变化规律进行研究。针对该合金高温变形过程中复杂的流变行为,以实验所得数据为基础,基于模糊神经网络方法建立该合金的高温本构关系模型,并与实验结果进行对比。结果表明:基于模糊神经网络建立Ti40合金的高温本构关系模型是切实可行的,模型的精度较高,最大误差为8.14%,不超过10%,可以很好地描述Ti40合金在高温变形时各热力学参数之间高度非线性的复杂关系,弥补传统回归模型不能反映变形全过程的局限性,是一种便捷、有效的具有广泛应用前景的表征工程材料本构关系的方法。The hot deformation characteristics of Ti40 alloy were studied in the temperature range of 900-1 100 ℃,the maximum deformation degree of 60% and strain rate range of 0.01-10 s-1 with Gleeble 1500 thermal simulator.Based on the results obtained from experiment,a constitutive relationship model of Ti40 alloy was established by the fuzzy-neural network and compared with the experimental result.The results indicate that the method is feasible to construct a constitutive relationship model for this alloy and high in the degree of fitting,Simultaneously,the complicated nonlinear relationship of thermodynamic parameters can be well described by the network model,making up for the limitations which traditional regression cannot reflect the whole deformation process.This method is not only a wide prospect of application but also an effective way to establish the constitutive relationship model of titanium alloys.

关 键 词：Ti40钛合金 本构关系 模糊神经网络 变形行为 

分 类 号：TG146.4[一般工业技术—材料科学与工程]