2519铝合金本构关系的人工神经网络模型

A Constitutive Relationship Model of the 2519 Aluminium Alloy Based on Artificial Neural Network

2519铝合金是一种新型的装甲材料.变形时,各热力学参数之间存在着非常复杂的非线性关系.采用Gleeble1500热模拟机上圆柱体轴对称高温压缩试验数据建立了该合金本构关系神经网络模型,利用所建立的网络模型对其它一些热力学状态下材料的流变应力进行了预测.实验结果表明:在目标函数为0.2、隐层节点数为5、学习率为0.1时,预测数据与实验数据吻合良好,系统误差较小(拟合度为2.6%),表明已形成了一个知识基的本构关系模型.

The aluminum alloy of 2519 is a new material for armory use. There are very complex nonlinear relations among the thermal-dynamical parameters in the process of deforming. An artificial neural network model for constitutive relationship was constructed with the compressing experimental data of cylinder specimens at elevated temperatures on the Gleeble 1500 thermal simulator.Flow stress of the material under various thermal dynamics conditions have been predicted by the network model .The results show that the systematical error is small (fitness 2.6%)with the objective function of 0.2, the number of nodes of 5 in the hidden layer and a learn-rate of 0.1. A knowledge-based constitutive relation was constructed in this study.