Early-warning process/control for anaerobic digestion and biological nitrogen transformation processes: Batch, semi-continuous, and/or chemostat experiments. Final report

Abstract

The objective of this project was to develop and test an early-warning/process control model for anaerobic sludge digestion (AD). The approach was to use batch and semi-continuously fed systems and to assemble system parameter data on a real-time basis. Specific goals were to produce a real-time early warning control model and computer code, tested for internal and external validity; to determine the minimum rate of data collection for maximum lag time to predict failure with a prescribed accuracy and confidence in the prediction; and to determine and characterize any trends in the real-time data collected in response to particular perturbations to feedstock quality. Trends in the response of trace gases carbon monoxide and hydrogen in batch experiments, were found to depend on toxicant type. For example, these trace gases respond differently for organic substances vs. heavy metals. In both batch and semi-continuously feed experiments, increased organic loading lead to proportionate increases in gas production rates as well as increases in CO and H{sub 2} concentration. An analysis of variance of gas parameters confirmed that CO was the most sensitive indicator variable by virtue of its relatively larger variance compared to the others. The other parameters evaluated including gas production, methane production, hydrogen, carbon monoxide, carbon dioxide and methane concentration. In addition, a relationship was hypothesized between gaseous CO concentration and acetate concentrations in the digester. The data from semicontinuous feed experiments were supportive.