Full-Scale Digesters: An Online Model Parameter Identification Strategy

This work presents a new standard in the model, identification, and control of monitoring purposes over anaerobic reactors. One requirement that guarantees a normal controller operation is for the faculty to measure the data needed periodically. Due to its inability to easily obtain the concentrations of acidogenic bacteria and methanogenic archaea periodically using reliable and commercial sensors, this paper presents an algorithm composed of an asymptotic observer (considering the reaction rates are unknown), aiming to estimate these concentrations. This method represents a significant advantage because it is possible to perform a resource-saving strategy using standard measurements, such as pH or alkalinity, to calculate them analytically in natural environments. Additionally, two yield parameters were included in the original anaerobic model two (AM2) to unlock implementations for a wide range of organic substrates. The static parameter identification was improved using a new method called step-ahead optimization. It demonstrates significant improvements fitting the mathematical model to data until a 78.7% increase in efficiency (compared with the traditional optimization method genetic algorithm). After the period of convergence, the state observer evidences a small error with a maximum 2% deviation. Finally, numerical simulations demonstrate the structure's strengths, which constitutes a significant step in paving the way further to implement feasible, cost-effective controls and monitoring systems in the industry.