Determination of the adequate minimum model complexity required in anaerobic bioprocesses using experimental data

BACKGROUND: Modelling of anaerobic bioprocesses requires adequate conciliation between model accuracy and complexity. Anaerobic digestion is a highly complex biocatalysed process in which organic (waste) materials are converted into biogas. Complex mathematical models of the process can be difficult to manage, while too simplistic models may be insufficient for specific applications. RESULTS: A technique based on principal component analysis (PCA) is presented to assess the most adequate complexity in anaerobic bioprocess models using preliminary data. The technique is applied to experimental data from a continuously operated pilot-scale anaerobic digestion process treating diluted wine and to process simulation data generated by a highly complex model (ADM1). The results obtained suggest that models incorporating only four reactions can describe most of the experimental data variability. Similarly, a simplification of the highly complex modified ADM1 used to a four-reaction model could attain equivalent simulation accuracy. A set of possible four-reaction networks was evaluated using the PCA information and a best candidate network of reactions is obtained. CONCLUSION: The developed PCA-based tool allows for the assessment of the adequate complexity in mathematical models of anaerobic digestion of ethanol, a process that appears describable with a high level of accuracy by considering only four biological reactions. Copyright © 2008 Society of Chemical Industry