Study of the use of ADM1 model as platform of modeling for different configurations of reactors used in the anaerobic treatment of vinasse

Abstract

Over the years it is increasingly growing the search for measures to mitigate and even eliminate the negative effects of waste from production processes of industries, looking viable alternatives to the environmental waste does not affect the environment and human health. The vinasse residue from ethanol manufacturing process, has a high potential pollutant because of its physico-chemical characteristics, such as high concentrations of organic matter, high suspended solids concentration, high demand biochemical of oxygen (BOD), low pH, high corrosiveness and high concentrations of potassium, nitrogen, phosphorus and sulfate. This wastewater may be treated using anaerobic biological reactors, which have the advantage of less energy expenditure when compared to aerobic systems, less sludge generation rates and the possibility of energy recovery due the use of biogas form (hydrogen, methane). However, the technology employed with the use of biogas depends of the treatment system, this is, a system composed for a acidogenic reactor followed by methanogenic reactor (two phase system), biogas will generate two streams, one consisting predominantly of hydrogen and another of methane. In the system made only by a methanogenic reactor (single phase system), as the name say, the gas stream is primarily composed of methane. From the experimental data collected for different conditions, it is important to design mathematical models based on momentum transfer processes, mass and energy, and microbial kinetics in order to allow predictions and simulations of treatment in different scenarios and assess which initial conditions are more relevant to the treatment process. For this, it will be used modeling platform ADM1 that presents many biochemical processes and physicochemical anaerobic digestion. The data modeling in the anaerobic treatment of vinasse with ADM1 platform will be estructured in Matlab, in which the hydrogen production (acidogenic reactor) will be verified first, after that will be verified the methane production (methanogenic reactor) and at the end the production of hydrogen and methane will be verified (acidogenic reactor followed by methanogenic reactor). Following a local and global sensitivity analysis will be performed to optimize the input parameters or despise those who have no influence on the final results, for each of the models. Then, with the final model in hand, this will undergo a statistical validation step using the Maximum Likelihood Method and Chi-square test in order to verify that it is reliably the actual data. Finally, the model predictions for different conditions will be evaluated and, with it, one can predict which scenarios lead to increased production of methane, hydrogen or both.