基于正交试验-人工神经网络模型的镁合金薄板加热拉伸工艺优化
Optimization of Heated Tensile Process for Magnesium Alloy Sheet Based on Orthogonal Experiment-artificial Neural Network Model
张 华, 陈 丰, 夏显明
点击：8551次 下载：0次
作者单位：安徽科技学院 工学院, 安徽 凤阳 233100
中文关键字：正交试验； 极差分析； 方差分析； 在线加热拉伸； 人工神经网络模型
英文关键字：orthogonal experiment; range analysis; variance analysis; online heating tensile; artificial neural network model
中文摘要：设计了正交试验表，在万能试验机上对AZ31镁合金试样进行了在线加热拉伸试验，对正交试验结果进行了极差分析和方差分析，并且通过人工神经网络模型优化，得到了AZ31镁合金薄板的最佳拉伸工艺参数，即拉伸温度300 ℃，拉伸速度10 mm/min，变形量20%。该结论为指导AZ31镁合金薄壁拉伸件的生产提供了理论依据。
英文摘要：Orthogonal experiment table for AZ31 alloy was designed. Online heating tensile tests were carried out using universal testing machine. Range analysis and variance analysis of orthogonal experimental results were carried out. The process was optimized by artificial neural network model. The best tensile process parameters for AZ31 alloy are achieved, that is, tensile temperature is 300 ℃, tensile speed is 10 mm/min, deformation is 20%. The theoretical basis is provided for thin-walled tensile parts production of AZ31 alloy.