Empowered parameter identification procedure for anaerobic digestion lumped model, stability and reliability analysis

Mathematical modeling of Anaerobic Digestion is fundamental for the prediction of crucial quantities such as CH4 and CO2 content. Among many possibilities, the AMOCOHN model, an upgraded version of the AMOCO model, represents a simplified mathematical model and is qualified for control purposes, avoiding long calculation time while assuring precise enough results. Their differences rely on the parameter identification procedure. While AMOCO model calibration is linear-based, AMOCOHN is nonlinear-based. Because of its inherent sensitivity to fluctuations, this produces some deceptive results in terms of coefficients assessment when the regression is performed with a nonlinear algorithm. This work aims to simplify and improve the robustness and stability of the AMOCOHN identification procedure by integrating an adaptation of the same dual-step approach originally proposed in AMOCO. The results, illustrated by a comparison between literature and ADM1 simulation data, reveal a mean R2-adj of 0.98, demonstrating its effectiveness while preserving the model’s simplicity and flexibility