Fig 1.
Classification of fault detection and diagnosis method [1].
Fig 2.
Flowchart of static PCA [30].
Fig 3.
Moving window PCA flowchart [30].
Fig 4.
Moving window dynamic PCA flowchart.
Fig 5.
A diagram of the Tennessee Eastman process simulator [13].
Table 1.
Type and description of process faults [33].
Fig 6.
Trends of TEP simulated data for normal process.
Table 2.
Comparison of FAR for MWDPCA with other methods.
Table 3.
Comparison of MDR for MWDPCA with other methods.
Table 4.
Comparison of DTD for MWDPCA with other methods (sample).
Fig 7.
Monitoring performance of MWDPCA for normal process.
Fig 8.
Monitoring performance of MWDPCA for fault 2.
Fig 9.
Monitoring performance of MWDPCA for fault 7.
Fig 10.
Monitoring performance of MWDPCA for fault 6.
Fig 11.
Monitoring performance of MWDPCA for fault 8.
Fig 12.
Monitoring performance of MWDPCA for fault 12.
Fig 13.
Monitoring performance of MWDPCA for fault 13.
Fig 14.
Monitoring performance of MWDPCA for fault 14.
Fig 15.
Monitoring performance of MWDPCA for fault 17.
Fig 16.
Monitoring performance of MWDPCA for fault 18.
Fig 17.
Trends of TET data for normal process.
Fig 18.
Monitoring performance of MWDPCA for TET data normal process.
Fig 19.
Monitoring performance of MWDPCA for (A) step change (B) slow ramp.
Table 5.
DTD and MDR for MWDPCA applied on TET data.