Operating Condition Recognition of Pre-denitrification Bioprocess Using Robust EMPCA and FCM

Poor-quality data has become a serous problem to the model, control and optimization. An improved robust EMPCA integrated with fuzzy c-means (FCM) clustering is used to classify the operational state in the activated sludge process. The method is demonstrated by IWA simulation benchmark. The experimental results show the proposed the method can accurately classify the operational state of the activated sludge process in the PC-space after outliers and missing data are effectively detect, rectified.