Representing full-scale CSTR biogas systems at laboratory scale: Start-up and microbial stabilisation

Laboratory-scale continuous stirred-tank reactors (CSTRs) are widely used to study and optimise anaerobic digestion (AD). However, it remains unclear to what extent the microbiology of such systems accurately represents full-scale processes and how much time they require to reach representative and stable conditions. In this study, start-up and stabilisation of process performance and corresponding community composition was investigated in five parallel 10-L CSTRs operated as laboratory-scale models of a full-scale agricultural biogas plant. The reactors were inoculated from the full-scale plant and operated under corresponding process conditions using the same substrate. Replicate reactors were operated and monitored for 213 days, while one reactor was run for an additional 330 days. All replicate reactors exhibited highly reproducible behaviour, including an initial transient disturbance characterised by volatile fatty acid (VFA) accumulation, followed by process recovery and stable operation. In the period following the disturbance, the laboratory-scale reactors represented the full-scale process well in terms of gas production efficiency and overall process stability. However, the microbial community composition diverged markedly from the original inoculum and was separated from the full-scale process during the whole experimental period. These findings indicate that stable process performance does not necessarily coincide with a stable microbial community, as substantial shifts in community composition occurred despite stable reactor operation. This highlights that extended stabilisation periods may be required even in direct scale-down reactors and that there are limitations to how well the microbiology in a full-scale system can be represented in the laboratory.