A General Methodology for Optimized Takagi-Sugeno Fuzzy Modeling of Nonlinear Continuous Fermenter for Biofuels (Ethanol) Production Using Golden Section Search Approach

The overcoming inclusion of biotechnology in biofuels industry involves several challenges among which are found the variety of operational cycles, the highly nonlinear behavior of the processes and the need for measurement of intermediate variables. In order to reproduce biological conversion of biodiesel production discharge products into other biofuels, experimental data from ethanol production from glycerol/glucose mixture was analyzed implementing fuzzy techniques to investigate and model the nonlinear behavior of the process. This paper presents a general methodology for TS fuzzy modeling based on a novel approach on data structured regression which consists on combination of fuzzy c-regression model and clustering using a golden search algorithm approach to adjust the proper number of membership functions to fit the model and minimize the statistic difference among the experimental data, simulated data and the Fuzzy Inference System results.