Integrated modeling using fuzzy logic and response surface in predicting and optimizing bioethanol production conditions

Background and Objective: One of the most important aspects of bioethanol production is the optimization of fermentation process conditions. In this regard, modeling and accurate prediction of processes using modern methods are essential. Red mulberry juice, as a rich source of natural sugars, can be used as a raw material for bioethanol production. Saccharomyces cerevisiae is a key microorganism in the fermentation of sugars for bioethanol production. The present study, after reviewing related literature and background, includes: bioethanol production using diluted mulberry juice, optimization of operating conditions for high-purity bioethanol production, separation and purification of bioethanol to a high purity level, and a comparison between two experimental design methods—response surface methodology (RSM) and fuzzy logic—for predicting the output results.
Materials and Methods: The raw material used for bioethanol production was red mulberry juice with a Brix of 78, obtained from local markets in Sanandaj. Saccharomyces cerevisiae PTCC 5269 was obtained and activated from the Pasteur Institute of Iran. The mulberry juice was diluted with distilled water to a Brix of 10. The prepared solution was adjusted to an acidic pH of 4.75 (the optimal growth value for the yeast used). A yeast dosage of 0.375 g/L was added and the mixture was placed in a stirred incubator at 200 rpm and 30°C. Fixed values included an inoculum volume of 150 mL, and an initial feed volume of 350 mL with a Brix concentration of 10. An experimental design was conducted to optimize the pH, duration, and temperature of the bioethanol production process to maximize ethanol yield.
Results: To validate the independent variables in the experimental design, the p-value threshold of 0.05 was used as described in the methods. For the independent variables (temperature, process time, and pH), the first-order effects on the dependent variable (ethanol concentration) were significant, with p-values of 0.009, 0.001, and 0.009, respectively. The determination coefficient (R²) of the fitted model was 96.93%, indicating a good predictive capability. The optimum point identified from the experimental design corresponds to approximately 66 hours of process time at 30.65°C with a pH of 5.18, which would yield an ethanol production of about 14.86% when using red mulberry juice. The ethanol produced in this study exhibited relatively low purity, with an optimum yield of 14.62%; impurities are attributed to suspended and insoluble matter (acids, sugars, proteins) and other by-products that can form under specific fermentation conditions. Separation of ethanol from the optimum solution was performed via distillation. After 12 hours of distillation, the product was purified and analyzed by refractometry, achieving a purity of 90.39%. Subsequently, multiple membership functions were evaluated for designing the fuzzy model. After evaluation, the model with 4, 3, and 3 membership functions for the first, second, and third input parameters and 36 fuzzy rules was selected as optimal. This model demonstrated acceptable performance with an average relative error of 8.51%.
Conclusion: Red mulberry juice shows strong potential for bioethanol production. Response surface methodology (RSM) is an effective experimental design and statistical analysis tool for reducing the number of experiments. Saccharomyces cerevisiae demonstrated a robust ability to produce bioethanol, and analyses indicated that this factor significantly influenced ethanol yield. The final product characteristics showed good ethanol content with low turbidity and color. Distillation for separation and purification led to higher-purity ethanol with a lower Brix than the starting material. For future work, researchers are encouraged to repeat the study with other substrates capable of producing bioethanol and to compare results across different materials and conditions.