摘要
通过对拉伸件成型状态的声发射测试,进行了拉伸过程AE特征参数信号的提取。对采集到的信号进行局域波分解后提取各IMF(Intrinsic Mode Function)的能量值作为初始特征参数,应用遗传算法对初始特征参数进行优化,生成最优特征参数。采用简单的马氏距离方法,将正常状态和微裂纹状态两种质量状态下的实验数据进行计算,比较两种状态下马氏距离的大小,取其中最小判别距离对应的状态为测试样本的状态类型。研究结果说明了该方法可以有效地识别出拉伸件的微裂纹AE信号,从而判断出拉伸件的初始裂纹状态,实现AE信号特征参数的优化及对金属拉伸件成型质量状态的识别。
The characteristic parameters of acoustic emission （AE） signals were extracted through acoustic emission tests for molded state metal drawing parts. The collected signals were decomposed in local fields to extract the energy value of each IMF （ intrinsic mode function） as the initial characteristic parameters, they were optimized with a genetic algorithm to become the optimal characteristic parameters. The test data both under the normal condition and the crack state were computed with the simple Mahalanobis distance method, the values of Mathalanobis distances under the two states were compared, and then the state corresponding to the minimum discriminated distance was chosen as the state type of the tested sample. The study results showed that this method can be used to recognize effectively the crack AE signals of drawing parts to judge their initial crack state, and to realize characteristic parametric optimization of AE signals and molding quality state recognition of metal drawing parts.
出处
《振动与冲击》
EI
CSCD
北大核心
2012年第17期154-158,共5页
Journal of Vibration and Shock
基金
科技型中小企业创新基金(09C26213201011)
关键词
拉伸件
声发射
遗传算法
特征参数
状态识别
drawing parts
acoustic emission （AE）
genetic algorithm
characteristic parameters
status identification
