Statistical Process Monitoring with Missing Data

Project description

Kernel Based Statistical Process Monitoring and Fault Detection in the Presence of Missing Data.

IEEE Transactions on Industrial Informatics. 2021. DOI:10.1109/TII.2021.3119377.

Importance/impact, challenges/pain points

（1）Nonlinear processing monitoring is sensitive to missing data；

（2）Online monitoring requires efficient method.

Solution description

FINLMC: Fast Incremental Nonlinear Matrix Completion

Key contribution/commercial implication

（1）Online missing data imputation;

（2）Stable method for process monitoring;

（3）Low false alarm rate;

（4）High fault detection rate.

Team/contributors

Jicong Fan, Tommy W.S. Chow, S. Joe Qin.

•Numerical results