Note: under developing stage
Now the offline modelling, fault detection and isolation have been completed. But the isolation result result slightly differs from the paper.
See simulation.ipynb for example
Python implementation of structured joint sparse principal component analysis. This implementation is based on the Transformer
type in scikit-learn
.
SJSPCA can be used both for fault detection and fault isolation. For fault detection, the T2 and SPE statistic is calculated.
For details, refer to the paper published on IEEE TII: Structured Joint Sparse Principal Component Analysis for Fault Detection and Isolation.