Anomaly Detection of Continuous Wet Granulation using Multivariate Statistical Process Control (MSPC)

The multivariate statistical process control (MSPC) was applied as an anomaly detection method in the continuous manufacturing of drugs. It was confirmed whether abnormal changes in process parameters which were intentionally introduced during operation of a continuous wet granulator could be detected by MSPC. As abnormalities, the water addition rate, center blade rotation speed, and particle sizes of raw materials were varied, whereby all of the abnormalities could be detected with high precision by monitoring of statistics using MSPC. Moreover, it was possible to correctly confirm the normal operating state as being the normal operating.