Modelagem e simulação das etapas de liquefação, sacarificação e fermentação do processo de produção de etanol de milho: estudos para otimização do processo

Abstract

The use of corn for first-generation ethanol production has been increasing exponentially in Brazil, making the optimization of this process essential to enhance its competitiveness relative to sugarcane. This study models, simulates, and optimizes the liquefaction (L), saccharification (S), and fermentation (F) stages of corn ethanol production, evaluating four process configurations: (i) sequential L, S, and F (LSF), (ii) simultaneous L and S followed by F (SLSF), (iii) separate L with simultaneous S and F (LSSF), and (iv) simultaneous L, S, and F. Mathematical models based on kinetic parameters and literature data were implemented in Python using Google Colaboratory. Uncertainty propagation in the modeling was analyzed using the Monte Carlo Method (10,000 samples) for ethanol concentrations. A plant operating 330 days per year was simulated to estimate the OPEX, with revenues derived from ethanol, corn oil, and DDGS (Dried Distillers Grains with Solubles). The differential evolution method was employed to maximize the annual net profit. Using initial concentrations reported in the literature, the concentrations of the process components were obtained. The highest uncertainty in final ethanol concentration was 7.45%, observed in the SLSF scenario. In terms of optimization, the SSF scenario resulted in the lowest final ethanol concentration (115 g·L⁻¹), the lowest initial starch concentration (239 g·L⁻¹), and the lowest annual net profit (R$ 27 million). Corn accounted for the highest share of the total product cost averaging 73%. DDGS contributed approximately 31% of the total revenue across all scenarios. The sensitivity analysis indicated that variations in the prices of ethanol, corn, and DDGS were the factors that had the greatest impact on the economic indicators of the simulation, with particular emphasis on the SSF scenario, which showed an approximate 600% variation in its annual net balance. Finally, the LSF scenario presented better economic results.