Adaptive Sampling of Dynamic Systems for Generation of Fast and Accurate Surrogate Models

For economic nonlinear model predictive control and dynamic real-time optimization fast accurate models are necessary. Consequently, the use of surrogate to mimic complex rigorous is increasingly coming into focus. systems, focus so far had been on identifying a system's behavior surrounding steady-state operation point. In this contribution, we propose novel methodology adaptively sample process generate dataset for building models. The goal developed algorithm cover an as large possible area feasible region original model. To demonstrate performance presented framework it applied chlor-alkali electrolysis.