基于BP神经网络的7050铝合金本构关系模型及加工图

Model of Constitutive Relationship and Processing Map for 7050 Aluminum Alloy Based on BP Neural Network

利用热模拟试验机对7050铝合金进行等温压缩试验,获得了不同变形温度、不同应变速率和不同真应变下的流动应力数据。以试验数据为基础,建立了7050铝合金的BP神经网络本构关系模型。分析表明,该神经网络本构关系模型具有较高的精度,并得到了相关性和平均相对误差的验证。利用BP神经网络修正的数据,根据动态材料模型（DMM）建立功率耗散图和失稳图,通过叠加得到7050合金的热加工图,并利用热加工图确定了该合金的加工安全区和流变失稳区。分析得出了最佳变形工艺参数：变形温度为420～450℃,应变速率为0.01～0.10s-1,该区域的峰值功率耗散系数η为0.40。

Isothermal compression testing on 7050 aluminum alloy was conducted by using thermal simu- lator, and the flow stress data was presented at different strain rates and different temperatures for vari- ous true strains. Based on the experimental data, a BP neural network constitutive model for 7050 aluminum alloy was established. The analysis shows that the artificial neural network model for constitu- tive relationship exhibits a higher predicted precision, and it is verified from relativity and average rela- tive error. Using the data modified by applying BP neural network, the processing map of 7050 alumi- num alloy was drawn through the power dissipation map and instability map which were established based on the dynamic material modeling （DMM）. Then the processing zone and flow instability region based on processing map were determined. The results show that the optimal deformation parameters are presented as follows： deformation temperature of 420～450 ℃, strain rate of 0. 01～0. 10 s-1 , and the maximum power dissipation coefficient n is 0. 40.