Steady-state and transient behavior in microbial methanification: II. Mathematical modeling and verification

It has been shown that the Upflow Anaerobic Sludge Bed (UASB) system data reported earlier(1) cannot be explained by simple Monod-type substrate consumption patterns. An autoinhibition model was also ruled out because the substrate concentration range over which hysteresis was observed was much larger than such a model would predict. However, propionic and acetic acids were found to inhibit each other's conversion machineries. Since in the UASB system the biocatalyst is flocculated, it was found that a model additionally incorporating this facet of the reactor set-up could explain the steady-state data very well. Using the parameters generated from steady-state data and data from butyric acid step change,(1) i.e., the entire set of parameters (Table I), a very good agreement between predicted and observed data was found. International Mathematical and Statistical Libraries (IMSL) and Upjohn's NONLIN library combined with various root-finding and integrating subroutines were used for parameter estimation. The model thus described was used to predict the response of the UASB system when acetic acid and propionic acid influent concentrations were stepped-up/down. The agreement between the predicted and observed data was found to be excellent in each case during the step-up schedule. During the step-down the data seemed to indicate that the UASB system, like any other chemostat, responded faster than predicted. This could be due to the fact that when the culture has to "gear up" part of the lag time is the time required for the cell to produce the requisite amount of enzymes. In the case of "gearing down" this time is not required and the system responds faster.