Exploration of potential methodological bias on feeding mode and gas sampling during semi-continuous anaerobic...

Laboratory-scale semi-continuous experiments are widely used to simulate industrial anaerobic digestion. However, common feeding and gas sampling practices may not accurately reflect full-scale operations. This study aimed to compare two feeding modes: manual (once daily, five days per week) and automated (four times daily, seven days per week), along with two gas sampling methods: daily sampling using a gas bag or an online sampling every two hours. Two 20 L continuously stirred tank reactors were operated at 37 °C and fed with a mixture of cattle slurry, maize silage, and biowaste. Manual feeding caused fluctuations in biogas flowrate and methane content, while the automated system ensured stable results. Nevertheless, both feeding strategies yielded comparable weekly methane production (311 ± 3 vs. 306 ± 5 NL.kg^-1 VS). Bag sampling prior to manual feeding overestimated methane content and production (+4 %) compared to online sampling, highlighting that frequency and timing of gas sampling can introduce substantial bias.