Integration of NIRS and PCA techniques for the process monitoring of a sewage sludge anaerobic digester.

This study investigates the use of Hotelling's T(2) control charts as the basis of a process monitor for sewage sludge anaerobic digestion. Fourier transform near infrared spectroscopy was used to produce partial least squares regression models of volatile fatty acids, bicarbonate alkalinity and volatile solids. These were utilised in a series of principle component analysis models along with spectral data from digestate and feedstock samples to produce a pseudo steady state model, which was then used with an independent test set to evaluate the system. The system was able to identify disturbances to the digester due to a temporary alteration of the type of feedstock to the digester and separately, halving of the hydraulic retention time of the digester. It could also provide advance warning of disturbances to the digester. This technique could be used to improve the performance of sewage sludge anaerobic digesters by enabling optimisation of the process.