We focus on the natural, life, and engineering sciences in the fields of information, energy, and bioeconomy. In previous projects, a detailed dynamic model of the system was developed based on first-principle (white-box) approaches. While such models provide valuable insight, advanced model-based control methods such as Model Predictive Control (MPC) are often limited by discrepancies between the nominal and the actual system behavior. The objective of this thesis is to improve model accuracy through data-driven modeling and parameter estimation techniques. Potential approaches include data-driven parameter estimation for white-box models, the development of black-box models using machine learning methods (e. g., neural networks) and their combination (grey-box). * Literature review on dynamic modeling (system identification), especially data-driven approaches and comparison of ...
mehr