Development and bench-scale optimisation of a reactor system for rumen-based anaerobic digestion

Abstract

Biomass is becoming an important feedstock to be used in anaerobic digesters throughout the world due to its abundance and fast growth. Most common anaerobic digesters do not have the hydrolytic organisms to digest fibrous biomass, which leads to a high retention time required for biogas production from these feedstocks. This study aimed to evaluate the use of a rumen-based anaerobic digester to decrease the retention time required for the anaerobic digestion of lignocellulosic biomass. The digester was designed and constructed to run similarly to the rumen simulation technique (RUSITEC) which was used in other studies to study the rumen dynamics in animals. The designed digester was operated for 15 days to determine the digestion characteristics of the feedstock and the amount of biogas produced. Rumen fluid was used as an inoculum, with barley straw being used as the feedstock. The biogas production was high in the designed semi-continuous rumen reactor at 12.69 mL/gVS/day, with a methane content of 41.5% with over 50% of the feedstock undigested. The solid feedstock was removed after digestion and the protein content was measured in the digested grass to determine if it would be suitable as a feedstock for animals. The protein in the solid digestate increased from 4.75% of total solids to 7.5% in the early stages of the AD process but later decreased to 1.5% of total solids due to a lack of nitrogen in the feedstock for microbial growth. Batch digesters were used to test the effect of different organic loadings of barley straw on the biogas production in mesophilic digesters using rumen fluid as inoculum. The biogas produced from the different organic loadings, indicated that the highest biogas production of 269 mL/gVS was obtained at an organic loading of 16.24 gVS/L, which is extremely high loading rate for anaerobic digesters. The lowest biogas production was obtained from an organic loading of 2.04 gVS/L and 24.41 gVS/L, which amounted to 185 mL/gVS and 205 mL/gVS added, of biogas respectively. The organic loading of 24.41 gVS/L led to an increase in the total VFA and drop in pH below 6, which had detrimental effects on the amount of biogas produced having 25% less than the organic loading of 16.24 gVS/L. The organic loading of 16.24gVS/L was used to determine the microbial capability of rumen fluid to degrade different type of lignocellulose biomass. The biogas production was measured for three different grass feedstocks namely: Napier grass, barley straw and kikuyu grass. The biogas potential from different lignocellulosic biomasses did not differ significantly between one another and the biogas production were 275 mL/gVS for barley straw, 282 for napier grass, and 289 mL/gVS for kikuyu grass. The experimental data obtained by digesting different biomass feedstocks were fitted to various kinetic models (modified Gompertz, two-fraction first order, Monod type, and first order) efficiently simulated the biogas production from different organic loadings and ISRs with coefficient of determination values above 91%, with the modified Gompertz model having the best coefficient of determination value above 98%. The kinetic modelling revealed a good fit from all models with a coefficient of determination (R2) above 95%. The use of rumen fluid for the mono-digestion of lignocellulosic biomass has proved to be an effective tool to decrease the retention time required for biogas production and increase the rate of biogas produced. The decrease in retention time can potentially lead to smaller reactor systems to be built or increase the organic loading for more biogas production.