摘要
根据AZ31B镁合金挤压成形的加工特点,选择了挤压温度、挤压速度、挤压比、模具温度和挤压道次5个特征参数,利用离散型Hopfield神经网络的相关理论和方法,建立了试件力学性能的评价模型。结果表明,建立的神经网络模型可以很好地对AZ31B镁合金的抗拉强度等级进行分类,大大提高对试件力学性能评价和判断的效率。
According to the main processing characteristics of AZ31B magnesium alloy, selecting five characteristic parameters such as extrusion temperature, extrusion speed, extrusion ratio mole temperature and extrusion pass, the evaluation model of mechanical properties of the specimens was established by using the theory and method of discrete Hopfield neural network. The results show that the neural network model can classify the tensile strength of AZ31B magnesium alloy and improve the efficiency of evaluating and judging the mechanical properties of AZ31B magnesium alloy.
引文
[1]胡迎春,李尚平,廖廷华.基于Hopfield神经网络的结构优化算法研究[J].中国机械工程,2007,18(12):1456-1459.
[2]李永新.AZ31合金的挤压工艺数值模拟及参数优化[J].铸造技术,2017(3):705-710.
[3]黄光胜,李红成,张雷,等.工业态AZ31B镁合金薄板的拉伸性能与组织变化[J].重庆大学学报,2009,32(4):367-370.
[4]孟淑英,杜康云,李铁龙,等.连续铸轧AZ31B镁合金的组织及性能研究[J].热加工工艺,2017,46(13):167-168.