Integration of Electrochemical Impedance and Neuro-Fuzzy Modeling for Acid Concentration Prediction in Fermentation

This study investigates the use of electrochemical impedance spectroscopy (EIS) and adaptive neuro-fuzzy inference system (ANFIS) modeling to predict the concentration of butyric and acetic acids in a fermentation broth. Initially, a simulation of the glycolysis process was conducted to produce butyric acid and acetic acid from glucose. Using insights from the simulation data, mixtures of acetic and butyric acids were prepared at known concentrations. EIS measurements were performed on these mixtures to extract impedance parameters, which were subsequently fitted to equivalent circuit models. The extracted parameters served as inputs for the ANFIS model, which was trained to establish a relationship between electrochemical parameters and acid concentrations. The developed ANFIS model acts as a soft sensor, enabling accurate prediction of acid concentrations in fermentation broth based on electrochemical data. This approach offers a novel method for real-time monitoring and control of fermentation processes, enhancing efficiency and product quality. The findings demonstrate the potential of combining EIS and ANFIS for advanced process analytics in biotechnological applications. Future work will focus on validating the model with diverse fermentation conditions and scaling up the methodology for industrial use.