Dynamic model adaptation to an anaerobic digestion reactor of a water resource recovery facility

Abstract

This study deals with model adaptation of the AM2 model to an anaerobic digestion reactor of a water resource recovery facility, namely a 6000m3 reactor at VEAS WWRF, the largest of Norway. The model is based on the mass balance with six states including acidogens, methanoges, alkalinity, organic substrate, volatile fatty acid and inorganic carbon. The model adaptation is applied firstly to simulated reactor data for testing the algorithms, and then to experimental data. The experimental data are collected from laboratory analysis and online measurements from January to October 2017. The data of the first 100 days are used for model identification, and the remaining data for model validation. Identification analysis is based on the Fisher Information Matrix and the Hessian matrix. Also, a sensitivity analysis of the parameter estimates is accomplished.

Keywords

• Anaerobic digestion
• Fisher Information Matrix
• identifiability analysis
• mathematical model adaptation
• sensitivity analysis
• sloppiness
• water resource recovery facility