Activity: Talk or presentation › Talk at conference or symposium › Science to science
The design and scale-up of industrial scale aerated stirred bioreactors is mainly based on empirical knowledge although they are used in many industrial sectors. Biological processes including the conversion of carbon substrates and dissolved oxygen to carbon dioxide can be simulated with knowledge about the mass transfer and transport processes inside the reactor. Computational fluid dynamics provides a basis to model the species transfer and transport and hence lay the foundation for new reactor designs. Graphical processing units are employed to achieve short simulation times. To approximate the Navies Stokes equations the lattice Boltzmann method is used to take advantage of the processor’s parallelism. The bubble movement is simulated by solving the Newton’s equation of motion. Stochastic models are used for coalescence and breakup.
To achieve the goal of simulating the biological activity in the reactor a species transport solver for e.g. the dissolved oxygen and carbon dioxide is essential. Current solutions which can be efficiently parallelized suffer from negative concentration values and artificial inertia. Our novel solver is based on the lattice Boltzmann method which results in fast and efficient execution on GPUs while avoiding the unphysical effects mentioned before. The speed and memory consumption of this new solver is evaluated by comparing it to a particle based solver.
All models were validated with literature data. The simulation program is now applied for comparing reactor designs and scale-up concepts.