Full-Scale Digesters: Model Predictive Control with Online Kinetic Parameter Identification Strategy

This work presents a nonlinear model predictive control scheme with novel structure of observers aiming to create methodology that allows feasible implementations in industrial anaerobic reactors. In this way, new step-by-step procedure has been proposed and tested by solving two specific drawbacks reported the literature responsible for inefficiencies those systems real environments. Firstly, implementation structures based on modeling depends microorganisms’ concentration measurements; technology achieves is not cost-effective nor viable. Secondly, reaction rates cannot be considered static because, extended digestion (EAM2), large fluctuation parameters unavoidable. To face these drawbacks, acidogens methanogens, values have estimated using data collected sensors. After 90 days operation, error convergence was lower than 5% both observers. Four controller (MPC) configurations are used test all previous information trying maximize volume methane demonstrate satisfactory operation wide range scenarios. The results an increase efficiency, ranging from 17.4% 24.4%, as reference open loop configuration. Finally, operational robustness MPC compared simulations performed traditional alternatives industry, proportional-integral-derivative (PID) controllers, where some simple scenarios manage longer sufficient disrupt normal PID controller. For controller, simulation shows close 100% value.