Development of long-term dynamic BioWin® model simulation for ANAMMOX UASB micro-granular process

Three different innovative mathematical models were established to assess the volumetric nitrogen conversion rates of a lab-scale ANAMMOX upflow anaerobic sludge blanket reactor. Despite the vast technological and economical advantages of ANAMMOX, major challenges in process implementation call for mathematic simulations of the process, optimization of operating conditions, and kinetic/statistical analysis of the entire process. In this study, all developed mathematical models implemented via BioWin (R), were calibrated and validated, with adequate representations of a bench-scale micro-granular ANAMMOX process, to understand the potential setbacks of ANAMMOX process start-up and stabilization. Fundamental calculations of the kinetic and stoichiometric constants were integrated in the BioWin (R) software, and the adjusted parameters based on experimental analysis were applied for the assessments. Based on the results from the statistical approach, one of the models (Model III) exhibited a precise prognosis of the effluent data for the entire operational phases with a mean relative error (MRE) of approximately 1.96, 4.36 and 2.54% for nitrogen removal efficiency, removal rate and loading rate, respectively. Evaluating alkalinity and pH during the operation, led to identifying an acceptable fit between the experiment and Model III results, with a MRE of-7.19 and-0.35%, correspondingly. This study confirms the reliability of ANAMMOX-based process modeling and high predictive ability with BioWin (R). The presented simulation constants and modeling outline, can be further employed in full-scale applications design and development.

Automated summary

Three innovative mathematical models were established to assess the volumetric nitrogen conversion rates of a lab-scale ANAMMOX upflow anaerobic sludge blanket reactor. The models were calibrated and validated using BioWin (R), with Model III showing precise prognosis of effluent data. This study confirms the reliability of ANAMMOX-based process modeling and high predictive ability using BioWin (R).