首页> 外文会议>Annual Symposium on Quantitative Nondestructive Evaluation; 19980719-24; Snowbird,UT(US) >AN ADAPTIVE CLASSIFIER FOR DETECTING HELICOPTER DRIVETRAIN DAMAGE USING ACOUSTIC EMISSION
【24h】

AN ADAPTIVE CLASSIFIER FOR DETECTING HELICOPTER DRIVETRAIN DAMAGE USING ACOUSTIC EMISSION

机译:一种利用声发射检测直升机传动系统损伤的自适应分类器

获取原文
获取原文并翻译 | 示例

摘要

This paper describes recent developments in a program to detect damage to helicopter drivetrains using acoustic emission (AE). Data obtained from an SH-60 drivetrain on an NAWC test stand was correlated with seeded fault damage in order to identify acoustic emission characteristics unique to the various sources. The objective is to extend prior work in applications of pattern recognition techniques and advanced machine intelligence to AE by designing and implementing an autonomous adaptive procedure to recognize and classify drivetrain damage from AE data. Our approach was evaluate time series and frequency features of AE waveforms for evidence of damage-related effects, bound the feature values associated with baseline data, and finally take advantage of an unexpected, but not surprising, characteristic associated with damage of increasing severity. In essence we detect damage by identifying signals whose feature values exceed the baseline limits, then calculating a 'distance' to the feature. This distance increases with damage severity so that increasing distance, i.e. having a positive 'velocity', differentiates between damage which is becoming increasingly severe and damage that is stabilized. An example of the former is a growing crack, while stable damage may consist of a spall or broken gear tooth.
机译:本文介绍了使用声发射(AE)检测直升机传动系统损坏的程序的最新进展。从NAWC试验台上的SH-60动力传动系统获得的数据与种子故障损坏相关联,以便确定各种来源特有的声发射特性。目的是通过设计和实施自动自适应程序来识别和分类来自AE数据的传动系统损坏,从而将模式识别技术和高级机器智能在AE中的应用扩展到以前的工作。我们的方法是评估AE波形的时间序列和频率特征,以寻找与损害相关的影响的证据,限制与基线数据相关的特征值,并最终利用与严重性增加的损害相关的意外但不令人惊讶的特征。本质上,我们通过识别特征值超过基线极限的信号来检测损坏,然后计算到特征的“距离”。该距离随着损伤的严重性而增加,使得增加的距离(即具有正的“速度”)区分变得越来越严重的损伤和稳定的损伤。前者的一个例子是裂纹不断扩展,而稳定的损坏可能包括剥落或齿轮齿断裂。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号