A volatile fatty acids (VFA) sensor based on headspace chromatography was tested for online monitoring and control of a pilot-scale manure digester. The sensor showed satisfying results in terms of sensitivity and reliability for monitoring of the digester. The online VFA and biogas production data were used for automatic control of the digester based on feed flow manipulation. The control approach was based on optimization of biogas production while using VFA concentration as the alarm threshold. A rule-based supervisory system with a cascade controller was used to optimize the biogas production from the digester. The alarm state was set at 40 mM total VFA and 10 mM propionate concentration. The control algorithms could successfully maximize the biogas production without overloading the process. However, as the algorithm was based on a fixed biogas yield parameter and only used the biogas parameter for optimization, it could not distinguish between the decreases of biogas production from inhibition and from lower organic content in the substrate, which resulted in undesired decreasing of the control gas setpoint when the substrate was diluted. It was necessary to adjust the yield parameter in order to get this control approach to function properly, which is not suitable for the full-scale biogas plant where the organic content of waste streams can vary. An alternative approach could be a modified rule-based algorithm that includes VFA parameters to help distinguish between different process scenarios.
Water Science and Technology, 2012, Vol 66, Issue 11, p. 2496-2503