一个基于BPNN的TC11钛合金蠕变本构关系模型  

A creep constitutive model of TC11 titanium alloy based on BPNN

作　　者：彭相华[1] 罗迎社[2] 余敏[2] 

机构地区：[1]中南林业科技大学涉外学院,湖南长沙410004 [2]中南林业科技大学流变力学与材料工程研究所,湖南长沙410004

出　　处：《中南林业科技大学学报》2011年第12期183-186,共4页Journal of Central South University of Forestry & Technology

基　　金：湖南省教育厅一般项目(09C1016);中南林业科技大学青年科学研究基金项目(2009043B)

摘　　要：TC11钛合金在金属成形过程受载荷、温度、时间等因素的作用,其蠕变本构关系极其重要且复杂。针对这一情况,利用BP神经网络高精度的非线性拟合能力、记忆能力和很强的泛化能力,提出了一个基于BPNN(误差反向传播算法)的TC11钛合金蠕变本构关系模型。对关系模型进行训练、学习和仿真,并将其仿真结果与推广的Burgers模型推导的理论值和蠕变实验值进行误差分析,结果表明该模型具有很高拟合精度和一定的实际应用价值。TC11 titanium alloy is influenced by load,temperature and time during the rheological forming process and its creep constitutive relationship is important and complex.Aimed to this case,a constitutive relationship model for TC11 titanium alloy based on BPNN（Back-Propagation Neuron Network） was proposed according to the high precision of nonlinear fitting,memory capacity and strong generalization capacity of BP neuron network.The model was trained,studied and simulated,and then the simulation results were compared to theoretical values obtained by Burgers model and the creep testing values.The error analysis shows that the BPNN model owns high fitting precision and certain practical value.

关 键 词：BP神经网络 TC11钛合金 蠕变 本构关系 

分 类 号：TG146.23[一般工业技术—材料科学与工程]