基于BP神经网络的Ti_3Al基合金本构关系模型的建立  

Modeling of Constitutive Relationship of Ti_3Al-based Alloy based on BP Neural Network

作　　者：张小波[1] 王克鲁[1] 鲁世强[1] 宋昕[1] 

机构地区：[1]南昌航空大学航空制造与工程学院

出　　处：《特种铸造及有色合金》2013年第3期224-226,共3页Special Casting & Nonferrous Alloys

基　　金：国家自然科学基金资助项目(51164030)

摘　　要：应用THERMECMASTER-Z型热模拟试验机对Ti3Al基合金进行等温恒应变速率压缩试验,在变形温度为950～1350℃、应变速率为0.001～10s-1、最大真应变为1.2下获得流动应力数据。采用流动应力数据,并基于BP神经网络方法,建立了该合金的高温本构关系模型。结果表明,BP神经网络建立的高温本构关系模型具有很高的预测精度,可用于指导Ti3Al基合金热变形过程的有限元模拟和热加工工艺的制定,为本构关系模型的建立提供了一种准确有效的方法。The hot deformation tests were conducted for Ti3Al-based alloy by THERMECMASTER-Z thermal simuIator. The flow stress data was obtained at temperatures of 950～1 350 ℃, strain rates of 0. 001～10 s^-1 and maximum true strain of 1.2. According to the characteristic of empirical data and the theory of BP neural network, modeling of the high temperature constitutive relationship of Ti3 Al-based alloy was built up. The results show that the constitutive relationship model of BP neural network exhibits a higher prediction accuracy, which can he used to FEM simulation of Tis A1 alloy during thermal deformation process＇and to guide the hot working process.

关 键 词：TI3AL基合金 本构关系 BP神经网络 

分 类 号：TG146.23[一般工业技术—材料科学与工程]