Quantitative analyses of anaerobic wastewater treatment processes: identifiability and parameter estimation.

We investigated the problem of identifying the parameters of a nonlinear fifth order model describing the population dynamics of two main bacterial groups in an anaerobic wastewater treatment process. In addition to addressing problems concerning structural and practical identifiability, we also analyzed how mathematical descriptions of bacterial population dynamics can model real data. Using three data sets recorded under different experimental conditions, we estimated important biochemical parameters and demonstrated that our model could describe the data successfully. Parameters, which are simultaneously determined using information from all three experiments, have more reliable estimates. We conclude that, after appropriate estimation, this model can be used for optimization and the control of continuous processes. Copyright 2002 Wiley Periodicals, Inc. Biotechnol Bioeng 78: 89-103, 2002; DOI 10.1002/bit.10179