Research theme: Modeling

The models for designing controllers are primarily developed based on rigorous first principles on the laws of nature. Such models and related numerical solvers play an increasingly dominant role in virtual prototyping of new cars, industrial plants in the process industry, etc. The scientific challenge is the appropriate complexity reduction to make the overall optimisation of the plant design feasible. Contributions towards this challenge include the modeling of processes by large-scale first principles models and the use of measurement data for quantifying the dynamically relevant properties of non-linear chemical/physical processes in view of process monitoring, detection, (re)design, control and optimisation, as well as the development of model reduction and model uncertainty qualification tools for large scale industrial processes, that make a trade off between performance and computational efficiency tractable.

• Model reduction of port-Hamiltonian fluid dynamics
• Nonlinear control systems analysis
• Relating Lagrangian and Hamiltonian descriptions of electrical circuits
• Modeling and analysis of hybrid systems
• Model based optimization of fed-batch bioprocesses