Modelling anaerobic digestion of agricultural waste: From lab to full scale
Tatiana Segura, Paul Zanoni, Ulysse Brémond, Constance Lucet-Bérille, Antoine Pradel, Renaud Escudié, Jean‐Philippe Steyer
Abstract
• The ADM1 model effectively represented four reactors fed with different substrates. • Co-digestion of maize silage and cow slurry showed the synergy of both substrates. • Maximum operating ranges were determined through simulations. • A scale-up case was studied with the ADM1 model. Biogas production through anaerobic digestion offers a promising alternative to address climate change. In this study, the ADM1 model was used to simulate the digestion of four different substrates: a mixture of rye and maize silage, a mixture of cow slurry and maize silage, cow slurry alone, and food waste. Furthermore, the determination of total solids (TS) content was integrated into the model. Based on experimental data from 5 L Continuous Stirred Tank Reactors (CSTR), ADM1 model parameters were calibrated for each substrate, primarily differing in hydrolysis and inhibition constants. These parameters, along with two additional sets of parameters from the literature, were subsequently applied in simulations to assess methane productivity, yield, and TS under increasing organic loading rates (OLR) for each substrate. Among the substrates, food waste showed the highest productivity, yield, and solids removal, while rye and maize silage substrate was the most unstable, with system failure at the lowest OLR (7 kgVS . m - 3 . d - 1 ) compared to the other substrates. In addition, co-digestion of cow slurry and maize silage showed synergies between maize silage and cow slurry, achieving a productivity of 2.62 N m 3 . m - 3 . d - 1 . Moreover, the parameters determined for rye and maize silage mixture were further used to simulate a full-scale anaerobic digestion unit fed with rye and maize silage as substrate. A difference in volatile fatty acid accumulation was found between the lab- and full-scale systems, suggesting a possible better microbial adaptation to inhibitory factors in the full-scale system. Further investigation into inhibition effects is recommended to improve the predictive accuracy of the ADM1.