论文部分内容阅读
研究不同时效温度下(165℃、200℃、250℃)时效工艺对Al-Cu-Mg-Ag合金力学性能的影响,在此基础上,采用Levenberg-Marquardt算法训练神经网络对样本进行学习,在溶质原子在两种强化相中的定量关系尚不存在的前提下,建立了以时效温度与时间为输入参数和抗拉强度、屈服强度与伸长率为目标函数之间的函数关系。发现在目标函数为0.0005,隐层节点数为11,学习率为0.1时,系统误差较小。利用所建立的网络模型预测不同时效状态下材料的力学性能,发现预测数据与实验数据吻合良好,证明了网络的可靠性,为进一步研究工艺参数对力学性能的影响规律和工艺的优化设计提供了理论依据。
The effect of aging process at different aging temperature (165 ℃, 200 ℃, 250 ℃) on the mechanical properties of Al-Cu-Mg-Ag alloy was studied. On the basis of this, Levenberg-Marquardt algorithm was used to train the neural network to study the samples. The quantitative relationship between the solute atoms in the two kinds of strengthening phase is not existed yet. The relationship between the aging temperature and time as input parameters and the tensile strength, yield strength and elongation as the objective function is established. It is found that the system error is smaller when the objective function is 0.0005, the number of hidden layer nodes is 11 and the learning rate is 0.1. Using the established network model to predict the mechanical properties of materials under different aging conditions, the predicted data are in good agreement with the experimental data, and the reliability of the network is proved. This provides a theoretical basis for further studying the influence of process parameters on the mechanical properties and the optimal design of the process Theoretical basis.