Online analysis and optimisation of the anaerobic fermentation process

Abstract

As the contributions from renewable energy to the power grid increases, biogas plants could profit from on-demand production. Flexible feeding can supplement gas storage for further benefits. This change in operating methodology introduces new requirements for monitoring and process control. This thesis is focused on the development of measurement technologies to improve the methods of analysis of the anaerobic digestion process. Spectroscopic measurements were performed to characterise the key process parameters of the liquid phase that would be measured. A prototype MIR (Mid Infrared) MEMS (Microelectromechanical Systems) based spectrometer was combined with a diamond ATR (Attenuated Total Reflection) probe. These were tested and evaluated in a laboratory for measurement of selected process parameters in the liquid phase of the anaerobic digestion process.A second prototype MIR MEMS based spectrometer was tested for the measurement of the concentrations of biogas. This was combined with a raspberry pi to create a portable measurement system. Machine learning methods were tested and evaluated to find which technique gave the best performance. This approach showed high robustness and gave a performance comparable to classical electrochemical measurements. The relationship between feeding rate and gas production rate was investigated, in particular how the relationship changes as the digester conditions change at higher feeding levels. This work was performed considering the scenario of on demand production at laboratory and pilot scale. The testing found a difference in the response as the digester approaches inhibition, and this difference was exploited in order to make an assessment of the digester performance and inhibition levels.Finally, simulations using ADM1 (Anaerobic Digestion Model No. 1) were ran which replicated the pilot scale tests, and attempted to recognise the variation in the response of the gas production, and by doing so would enable automated control of the biogas plant. A PID (Proportional Integral Derivative) controller was implemented to control the simulation for a year long period. This work concluded that online measurements combined with modulated feeding for flexible power production can provide an improved understanding of the anaerobic digestion process and inhibition state.