A digital twin with population balance and deep neural network models links image, process, and context data. This creates an explainable early-warning system that detects process deviations and provides well-founded hypotheses about their causes.

By coupling CFD, LBM, and 1D models, hydrodynamics and reaction processes in jet-loop reactors are modeled at various levels of detail. Experimental data are used for validation and provide a robust basis for design, optimization, and scale-up.

Using an absorber model, we simulate CO₂ separation in air contactors. This allows us to investigate factors that influence the absorption process and derive targeted approaches for more efficient direct air capture technology.

The dynamic behavior of repeated adsorption-desorption cycles is investigated using numerical models. This allows for the evaluation of different operating conditions and the targeted optimization of the process.

The adsorption-desorption behavior in fixed-bed columns is investigated using dynamic simulations. In conjunction with experimental data, process conditions can be optimized and the technical feasibility of the process can be evaluated.

Using fluid dynamics, reactor, and micromixing models, we describe the complex processes occurring within the reaction-mixing pump. The resulting simulation tool highlights the impact of local mixing on rapid reactions and supports design and scale-up.