• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Technology solutions for affordable sustainment >DYNAMIC CONDITION BASED FEATURE EXTRACTION STRATEGY FOR MACHINE HEALTH MONITORING APPLICATIONS
【24h】

DYNAMIC CONDITION BASED FEATURE EXTRACTION STRATEGY FOR MACHINE HEALTH MONITORING APPLICATIONS

机译：基于动态条件的机器健康监测中的特征提取策略

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Machine health condition monitoring and fault diagnosis are two important tasks for the optimization of factory up time, diagnosis of component failure modes, and eventually support predictive maintenance decisions. That is why research to prognostics and health management (PHM) is being conducted world wide in different industrial applications. Within these applications, data acquisition is a key to collect machining data and understand machine conditions. The traditional data acquisition devices deal with controller data, sensors and other manufacturing systems at the same time. The feature extraction strategy is static once the PHM system being designed and established. However, which features to extract and when to extract them are highly related to realtime machine/component conditions. Therefore, this paper presents a framework for dynamic condition based feature extraction strategy for machine health monitoring applications. A knowledge base for the relationship of features and corresponding failure modes is developed first using quality function deployment (QFD) format. Then, a decision tree for prioritizing feature extraction order will be built based on the calculation of Bayesian probabilities. Because of the hierarchy and priority, the DAQ system can dynamically extract critical features based on the real-time condition of monitored machine systems.
机译：机器健康状况监视和故障诊断是优化工厂正常运行时间，诊断组件故障模式并最终支持预测性维护决策的两项重要任务。这就是为什么在全球范围内针对不同的工业应用进行预后和健康管理（PHM）研究的原因。在这些应用中，数据采集是收集加工数据和了解机床状况的关键。传统的数据采集设备同时处理控制器数据，传感器和其他制造系统。设计和建立PHM系统后，特征提取策略是静态的。但是，要提取哪些功能以及何时提取它们与实时机器/组件条件高度相关。因此，本文提出了一种基于动态条件的特征提取策略框架，以用于机器健康监控应用。首先使用质量功能展开（QFD）格式开发了功能与相应故障模式之间关系的知识库。然后，将基于贝叶斯概率的计算来建立用于优先处理特征提取顺序的决策树。由于层次结构和优先级，DAQ系统可以根据受监视机器系统的实时状况动态提取关键特征。

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