摘要
利用体积不变原理,通过拉伸试验测量真实应力应变关系的试验数据,再利用人工神经网络(BP网络)进行逼近,从而得到用BP网络表示的应力应变本构关系。并以梁的纯弯曲和圆轴的扭转弹塑性问题为例,提出了基于人工神经网络的求解新方法,计算结果验证了其方法的合理性。
In this paper test data of real stress-strain relationship are acquired by tensioning test according to the law of volume invariable. A back propagation network is applied to build up a real stress-strain relationship by training the network with the test data. A novel method related to the BP network is presented and it can be solved some elastic-plastic problems It has been seen that the method is available according to the examples such as pure bending of a beam and torsion of a shaft.
出处
《试验技术与试验机》
2006年第1期24-27,共4页
Test Technology and Testing Machine
基金
安徽省自然科学基金项目(03042306)
关键词
应力应变
人工神经网络
弹塑性
塑性极限载荷
stress strain
artificial neural networks
elastic plastic
limit load of plasticity
