基于人工神经网络的AZ31镁合金形变组织晶粒尺寸预测

Application of Artificial Neural Network in Predicting Grain Size of AZ31 Alloy

研究了不同挤压工艺下AZ31镁合金的微观组织,获得了不同变形条件下晶粒尺寸的试验数据,并结合人工神经网络建立了不同挤压工艺条件下晶粒度的BP模型。结果表明,用该网络模拟得到的结果与试验数据较吻合,建立的人工神经网络模型能够精确预测热挤压条件下变形参数与晶粒度的关系,对镁合金零件热变形工艺设计和产品质量控制具有特别重要的意义。

Microstructure of AZ31 magnesium alloy under different extrusion techniques is studied and the experimental data of different thermal mechanical processing are obtained; a BP algorithmic model of artificial neural network was established. It is found that the simulating results are better in agreement with the experimental results. The artificial neural network model can predict the grain sizes accurately under thermal extrusion of different thermal mechanical processing. It is significant for a successful forming process design and the production quality control to predict numerically microstructure evolution in magnesium alloys during the thermal mechanical processing.