Aplicação do modelo ADM1 na biodigestão anaeróbia da vinhaça para a produção de hidrogênio utilizando a plataforma EMSO

Abstract

The ethanol production in Brazil overtake volumes of 20 billion of litters per year and for each litter of alcohol is produced about 13 litter of vinasse, waste from distillation for alcohol production. Due the high charge of organic matter in this waste, some proposals for the destination of vinasse is the reuse of this in the fertigation, although this activity over time provide to soil and water sources nearby undesirable characteristics with respect to quality. Taking account the volume of produced vinasse, it´s necessary that new technology bring solutions and better results for the treatment of this waste with high charge of organic matter. The vinasse anaerobic treatment has been showing as an option of technology very advantageous among the other possibilities of treatment. The treatment gets about 50% of organic matter removal and also can produce hydrogen and methane, two combustion gases that have the possibility of them use as energy in the process. In this context, new studies for the feasibility and structuring of the factor and variables of the vinasse treatment process bring great benefits and readiness for the real application. Understanding, therefore, this requirement, the research realize an evaluation of a kinetic personalized model, based on ADM1, through the simulation in EMSO platform, finding the parameters and calibrating the variables e coefficients that takes the kinetics of acidogenic stage for the hydrogen production. In a model built for batch reactor, was possible to verify the preview of reaction’s tendency, showing, for the substrate uptake data, comparing between the experimental data and the model data a r² of 0,97. The hydrogen production evaluated for the batch reactor had a convergence with the experimental results, getting a r² of 0,95. In this way, the batch model could preview the experimental data over time. For APBR, the modified model results was able to converge with the experimental data for pH, with a variation of 9% between the values. It has not been possible to reach a conclusion if the model can describe the behavior of the reactor for hydrogen production and consumption of substrate, whereas the experimental data did not show a trend, anyway, the variation between experimental and modeled values for both variables was about 7%.