Data-Driven Modeling of a Bioreactor Using a Takagi-Sugeno Type Approach

In this article, we explore an analytical approach to generate Takagi-Sugeno (TS) type models from ’experimental’ data using the Sparse Identification of Nonlinear Dynamics (SINDy) technique. Our method combines the approximation of reduced nonlinear models with the linearization of the same model at specific operating points, allowing us to capture the dynamics of complex systems under variable conditions for each point of interest. We then present a detailed implementation of this approach, highlighting its application in a real-world scenario of modeling a biotechnological process.