摘要
金属裂纹声发射信号特征提取是根据其进行故障诊断的关键,提出了BP神经网络和模式识别相结合的提取金属材料疲劳声发射信号特征的新方法,并利用美国PAC公司SAMOS声发射检测系统采集到声发射的各种参数,应用该方法选择出一些对分类识别最有效的特征参数;并采用可分离性判据进一步验证其正确性。
The collection on characteristic parameters of emitted signals of metal cracking sound is the key to carrying out accordingly the fault diagnosis. With the combination of BP neural network and mode recognition a new method for collecting the characteristic parameters of emitted signals of metal materials fatigue sound has been put forward, and using the SAMOS sound emission detection system of American PAC Company to collect various kinds of parameters of sound emission. Some most effective characteristic parameters for sort recognition were selected by the use of that method, and its correctness was further verified by adopting the separability criterion.
出处
《机械设计》
CSCD
北大核心
2010年第2期84-87,共4页
Journal of Machine Design
基金
国家自然科学基金资助项目(50465002)
广西自然科学基金资助项目(桂科基0448014)
关键词
声发射
特征提取
BP神经网络
模式识别
acoustic emission
characteristics collection
BP neural network
mode recognition
