摘要详情
  1. 活动首页
  2. 摘要清单
  3. 摘要详情
108 / 2025-04-20 16:08:40
Weighted Mixture of Probabilistic PCA with Shared-Private Subspaces for Process Monitoring under New Operating Conditions
new modal process monitoring, probabilistic model, shared projection matrix
T02 数据驱动的故障诊断
全文待审
袁 鑫龙 / 中南大学
王 凯 / 中南大学
袁 小锋 / 中南大学
王 雅琳 / 中南大学
In modern process industries, frequent changes in


operating conditions often result in process data exhibiting


complex multi-modal distributions. Traditional fault detection


methods struggle to generalize to newly emerging operating


modes due to limited data samples. Additionally, real-world


industrial data are prone to anomalies caused by sensor faults or


environmental disturbances. To address these issues, this paper


introduces a shared-private subspace decomposition mechanism


and a sample-level weighting strategy based on the traditional


Mixture of Probabilistic Principal Component Analysis (MPPCA)


model. A fault detection model based on this method is designed,


and industrial case studies demonstrate its ability to effectively


extract global structural features from multi-modal process data,


showing significant advantages in fault detection tasks under


zero-shot and noisy scenarios in new operating conditions.
重要日期

  • 会议日期

    08月22日

    2025

    至

    08月24日

    2025

  • 04月25日 2025

    初稿截稿日期

主办单位
中国自动化学会技术过程的故障诊断与安全性专业委员会
承办单位
新疆大学
新疆自动化学会
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