摘要
本论文利用美国PAC公司SAMOS声发射检测系统采集到各种声发射信号,通过软件滤波和硬件滤波,以及独立分量分析融合多个传感器采集到的信号,然后分别对其进行特征提取和模式识别,最后通过D-S证据理论进一步融合其识别结果,提高对疲劳裂纹识别的准确度,为是否报警作理论依据。
In this paper, the acoustic emission signals are gathered by American PAC corporation SAMOS acoustic emission testing system. The signals gatheed from sensors are fused with the software filter and the hardware filter, as well as the independent component analysis, their features are extracted and their patterns are recognised respectively. The recognised results are fused further based on the D - S evidence theory to improve the accuracy of the results and as a theoretical basis for alarm.
出处
《热处理技术与装备》
2008年第3期66-70,共5页
Heat Treatment Technology and Equipment
基金
国家自然科学基金项目(50465002)
广西自然科学基金项目(桂科基0448014)
关键词
声发射
特征提取
独立分量分析
D—S证据理论
acoustic emission
feature extraction
independent component analysis
dempster- shafer（ D - S） evidence theory
