Modeling and parameter estimation of a biogas plant using maize silage in a two step model

The increasing demand for efficient systems for methane production from biomass has led to the deeper investigation of model predictive control strategies considering the yield and process stability under flexible operation. Most published work uses the complex ADM1 model as basis for the mechanistic description of the anaerobic digestion process. Unfortunately, the ADM1 has shown to be too complex for its application with respect to the usual monitoring capacities at biogas plants. Therefore, the present research proposes the application of theAMOCO two-step model which has a comparably simple structure requiring less experimental data. Results from a pilot-scale biogas reactor show the proper modeling of the pH value, methane and biogas production in two 50 liter reactors for 80, and 92 days, respectively. In total, thirteen parameters were optimized using Matlab 2013B and a non-linear regression toolbox that uses the Gauss-Newton method.