基于声发射信号的铝合金材料损伤表征识别  

Damage characterization and recognition of aluminum alloys based on acoustic emission signal

作　　者：张卫冬[1] 张习文[1] 杨斌[1] 丁贤飞[1] 艾轶博[1] 

机构地区：[1]北京科技大学国家材料服役安全科学中心,北京100083

出　　处：《北京科技大学学报》2013年第5期626-633,共8页Journal of University of Science and Technology Beijing

基　　金："十一五"国家科技支撑计划资助项目(2009BAG12A07-D07);国家自然科学基金资助项目(61273205;51005014);教育部中央高校基本科研业务专项(FRF-SD-12-028A)

摘　　要：随着高速铁路的不断提速,高铁轻量化设计中广泛采用高强铝合金材料,但高速列车齿轮箱体服役安全评价亟待完善.本文针对高速列车齿轮箱体使用的铝合金材料服役特性,搭建了声发射检测拉伸试验系统,运用BP神经网络算法对声发射信号进行训练与识别,实现对箱体材料拉伸损伤表征识别与材料服役状态的安全预警.本研究为材料损伤状态的无损实时识别提供了一种识别与预警方法.With the rapid development of high-speed rails, high-strength aluminum alloys are widely used in the lightweight design, but the service safety assessment of gear boxes in high-speed trains needs to be improved in China. An acoustic emission tensile test system was built for high-speed train gearbox shells made of aluminum alloys. After training and recognition by a BP neural network, acoustic emission signal was used for characterizing tensile damage in the materials and warning the materials service status. The research provides a method of nondestructive real-time characterization and warning for damage in aluminum alloys.

关 键 词：铝合金 声发射 损伤探测 神经网络 模式识别 

分 类 号：TG146.21[一般工业技术—材料科学与工程] TP18[金属学及工艺—金属材料]