Vertical Bubble Reactor

Responsible:

Jens Lang

Cooperation:

W. Ruppel, BASF Research Institute

Literature:

J. Lang, High-Resolution Selfadaptive Computations on Chemical Reaction-Diffusion Problems with Internal Boundaries, Chem. Engrg. Sci., 51, 1055-1070, 1996.

Gas-fluid systems give rise to propagating phase boundaries changing their shape and size in time. We consider a synthesis process of two gaseous chemicals A and B in a cylindrical bubble reactor filled with a catalytic fluid. The bubbles stream in at the lower end of the reactor and rise to the top while dissolving and reacting with each other. The right proportions of such reactors depend, among other things, on the rising behaviour of the bubbles and specific reaction velocities. Therefore, modelling and simulation of the underlying two-phase system can provide engineers with useful knowledge necessary to construct economical plants. A fully three-dimensional description of the synthesis process would become too complicated. We use a one-dimensional two-film model developed by W. Ruppel (1993). It is based mainly on the assumption that the interaction between the gas and the reactor fluid (bulk) takes place in very thin layers (films) with time-independent thickness. In the first film F₁ the chemical A dissolves into the bulk. From there it is transported very fast to the second film F₂ where reaction with chemical B takes place. As a result new chemicals C and D are produced causing further reactions.


Vertical Bubble Reactor in Section	Two-Film Model

As a consequence of the applied two-film model the dynamical synthesis process can be simulated with a fixed spatial domain involving the bulk and the film F₂. The spatial discretization needs some adaptation due to the presence of internal boundary conditions between bulk and film, see the literature cited above for a more thorough discussion. The grid evolution shows that at the beginning the reaction front is travelling very fast from the outer to the inner boundary of the film where the chemical B enters permanently. During the time period (here: reactor length) [0.1,0.5] the reaction zone does not change its position which allows larger time steps. After that with decreasing concentration of the chemical A at the outer boundary the front travels back, but now with moderate speed. Obviously, the adaptively controlled discretization is able to follow automatically the dynamics of the problem.

Bubbble-grid

Grid Evolution of the Bubble Reactor