(中南大学 材料科学与工程学院，长沙 410083)
摘 要: 在变形温度为200~400 ℃、应变速率为0.001~1 s−1条件下，对ZK60镁合金进行热压缩实验，建立一个单隐层前馈误差反向传播人工神经网络模型，研究该镁合金的流变行为。模型的输入参数分别为变形温度、应变速率和应变，输出为流变应力，中间隐含层包含23个神经元，并采用Levenberg-Marquardt算法对此网络模型进行训练。结果表明：ZK60镁合金的流变应力随变形温度升高和应变速率降低而减小；其高温压缩流变应力曲线可描述为加工硬化、过渡、软化和稳态流变4个阶段，但在较高温度和较低应变速率时，过渡阶段不很明显；所建神经网络模型可以很好地描述ZK60镁合金的流变应力，其预测值与实验值吻合很好；利用该模型预测的变形温度和应变速率对流变应力的影响结果与一般热加工理论所得结果一致。
关键字: ZK60镁合金；人工神经网络；流变应力；热压缩变形
（School of Materials Science and Engineering, Central South University, Changsha 410083, China）
Abstract:Compression tests for ZK60 magnesium alloy were carried out in the temperature range of 200−400 ℃ and strain rate range of 0.001−1 s−1. A feed-forward back-propagation artificial neural network with single hidden layer was established to investigate the flow behavior of this magnesium alloy. The input parameters of the model were temperature, strain rate and strain while flow stress was the output. A network contains 23 neurons in the hidden layer, and Levenberg-Marquardt training algorithm was employed. The results show that the flow stress of the ZK60 magnesium alloy decreases with increasing deformation temperature and decreasing strain rate. The flow stress curves obtained from the experiments are composed of four different stages, such as work hardening stage, transition stage, softening stage and steady stage. While for the relatively high temperature and low strain rate, the transition stage is not very obvious. The proposed model can describe the flow behavior of the ZK60 magnesium alloy precisely, the predicted results agree with the experimental values. The predicted results of the effect of deformation temperature and strain rate on the flow behavior of the ZK60 alloy are consistent with what is expected from he fundamental theory of hot compression deformation.
Key words: ZK60 magnesium alloy; artificial neural network; flow stress; hot compression deformation
