Modeling and control of nonlinear fermentation dynamics in brewing industry

This paper presents a mathematical modeling and advanced control strategy for the beer fermentation process, which is characterized by nonlinear biochemical kinetics and time-dependent dynamics. A biokinetic model was developed to describe the relationship between yeast growth, sugar consumption, and ethanol formation. The system was represented as a cascade of several continuous stirred-tank reactors (CSTRs), and experimental data confirmed a fermentation cycle of approximately 10 days. During this period, biomass concentration reached 6.8 g/L and ethanol levels exceeded 42 mmol/L. Substrate concentration (S) declined from 120 to 5 g/L, demonstrating effective conversion. The model was linearized around an operating point and reformulated into a 12-state space system with input variables: temperature (set at 20–22 °C) and pH (maintained within 4.2–4.5). These inputs were controlled using fuzzy logic control (FLC) and model predictive control (MPC). Simulation results indicated that the FLC controller reduced temperature deviation to ±0.3 °C and minimized pH fluctuation below ±0.05. The MPC strategy improved substrate consumption efficiency by 8.5% and decreased fermentation time by 12 hours under optimized input profiles. The combined FLC–MPC control scheme demonstrated superior robustness, smooth trajectory tracking, and adaptability to biological variability compared to traditional methods. The developed framework supports intelligent brewery automation and provides a scalable foundation for further integration of digital fermentation technologies.