报告详情
  1. 活动首页
  2. 报告清单
  3. 报告详情
Fault Diagnosis of External Gear Pumps via Open-Set Class-Incremental Continual Learning
编号：116 访问权限：仅限参会人 更新：2026-10-09 14:07:55 浏览：30次 张贴报告

报告开始：暂无开始时间（Asia/Shanghai）

报告时间：暂无持续时间

所在会场：[暂无会议] [暂无会议段]

暂无文件

摘要
Abstract—External gear pumps in long-term service usually face continuously evolving fault modes and the persistent emergence of unknown fault categories. This renders traditional fault diagnosis methods built on the closed-set assumption inadequate for practical operation and maintenance. For this reason, an open-set class-incremental continual learning (OSCIL) framework is proposed in this work for external gear pump fault diagnosis. It integrates unknown fault detection, new fault learning and historical knowledge retention in a unified framework. A fault feature representation space with a residual network is first built to provide the feature basis for continual learning. An experience replay mechanism is then introduced to alleviate catastrophic forgetting during incremental learning. Furthermore, a prototype-distance-based open-set decision mechanism is constructed to reject unknown faults according to the degree to which a sample deviates from the known fault distribution. Experiments carried out on an external gear pump fault diagnosis setup show that the proposed OSCIL maintains stable classification performance across multi-stage incremental learning tasks. It effectively avoids the forgetting of known classes, and simultaneously preserves the ability to recognize unknown fault modes.
 
关键词
Open set, class-incremental learning, continual learning, external gear pump, fault diagnosis
报告人
川 李
教授 重庆工商大学

稿件作者
anshun jiang Chongqing Technology And Business University
川 李 重庆工商大学
自强 蒲 重庆工商大学
发表评论
验证码 看不清楚，更换一张
全部评论
重要日期

  • 会议日期

    11月06日

    2026

    至

    11月08日

    2026

  • 10月15日 2026

    初稿截稿日期

  • 11月10日 2026

    注册截止日期

主办单位
IEEE Instrumentation and Measurement Society
承办单位
Sichuan University
联系方式
登录查看完整联系方式
历届会议
移动端
在手机上打开
小程序
打开微信小程序
客服
扫码或点此咨询