Adaptive fuzzy inference system in modeling the alkaline pre-treatment (NaOH) of elephant grass

Abstract

Abstract In order to increase its efficiency in the bioethanol production process, the biomass needs to undergo a pre-treatment, which can greatly burden the biochemical process. For this reason, industries are using some mathematical models that optimize internal processes, such as fuzzylogic, as it uses linguistic values ​​instead of numerical ones, which provides the addition of uncertainty to the problem, better understanding by non-specialists, answers faster and lower computational costs. Given the above, the objective of this work was to optimize the independent variables in the alkaline pre-treatment with NaOH in elephant grass to quantify the percentage of cellulose and lignin using fuzzylogic. For the input values ​​of this system, the variables of NaOH concentration and temperature were used, and for the outputs, the percentages of cellulose and lignin. The interpretation of the use of fuzzy logic in this experiment using real laboratory data, managed to optimize the best concentration of NaOH and the best temperature in an alkaline pre-treatment with NaOH to quantify cellulose and lignin, concentrations between 1.15 to 2 .15%, with temperatures between 100.0 to 102.5°C for the highest exposure of cellulose and 0.0 to 1.15% for concentrations and temperatures between 85.5 to 91.5°C for delignification. It was observed that the concentration significantly affects the result of greater exposure of cellulose and biomass delignification in the evaluated parameters, so the mathematical model proved to be efficient for this evaluation.