Robust Nonlinear Control of a Pilot-Scale Anaerobic Digester

Abstract

Recently, anaerobic digestion (AD) has shown potential applicability to reduce organic matter from wastewater. However, AD processes are not widely applied at industrial scale. One of the main factors that explain this phenomenon is related to the lack of control schemes that guarantee a suitable operation of the process. As well, AD control is a non trivial task, because the control schemes must operate under an uncertain environment because of the difficult to monitoring AD processes (on-line measurements are limited and expensive). Particularly, this chapter is devoted to the development and the experimental validation of a robust nonlinear approach to regulate (i) the total amount of decomposable organic pollutant agents measured as total organic carbon (TOC) and, (ii) the concentration of volatile fatty acids (VFA); by using in both cases the input flow rate as manipulated variable. The proposed approach is a model-based controller obtained from geometric control tools and the definition of an uncertain but observable function that lumps the uncertain terms associated to the dynamics of the controlled variables TOC and VFA (e.g., feeding composition, kinetic growth functions and parameters). The robust scheme is composed of a feedback linearizing control law and a high-gain Luenberger observer which adapts the linearizing control law from estimates of the uncertain function. The robust approach is experimentally validated in a fixed-bed AD process used in the wastewater treatment of industrial distillery vinasses. Several experiments are performed to evaluate the controller performance and robustness under different set-points, feeding concentrations and sampling times (2 and 30min). Results show that both, the regulation of COT and VFA can be addressed in spite of the full ignorance of the kinetic growth functions, noisy measurements and unknown feeding composition.