Application of principal component analysis to the anaerobic digestion process

Abstract

The anaerobic digestion of food wastewater was performed by operating a pilot-scale thermophilic (55 oC) two-stage continuous stirred tank reactor (CSTR). Through the reactor operating experiments, the variables of the anaerobic digestion such as pH, nitrogen, solids, organics, and biogas were organized into a time-indexed dataset. Principal component analysis (PCA) was applied to extract relevant information from the experimental dataset. Its purpose was to investigate the group of variables that mostly contribute to the early detection of operating problems. In addition, the adaptation of PCA for the dataset could be improved by introducing data smoothing and outlier detection.