381219 Large Bioreactor Simulation with Graphic Processors

Wednesday, November 19, 2014: 8:55 AM
313 (Hilton Atlanta)
Christian Witz1, Akshay Prakash1 and Johannes G. Khinast2, (1)Institute of Process and Particle Engineering, Graz University of Technology, Graz, Austria, (2)RCPE GmbH, Graz, Austria

So far, the engineering process of the reactor design was mostly driven by empirical knowledge. The effectiveness of the bioreactor and thus the biopharmaceutical manufacturing process could be enhanced by combining the state of the art empirical knowledge with a multiphase and multiscale simulation. Simulation results could then be used for the development and the design of new reactor generations. However, despite the recent improvement of computational capabilities, only a few seconds of real time operation of an industrial scale reactor takes months to simulate..

Thus, the goal of this study is to use graphic cards to speed up this simulation. The Compute Unified Device Architecture (CUDA) technology of nVidia has made the computational power of graphic processing units (GPUs) available for scientific calculations [1]. To take full advantage of the processor’s parallelism, localized calculation algorithms are needed. Therefore the lattice Boltzmann method is used for modelling the fluid flow field, which is agitated by a Rushton turbine. The method is based on the lattice gas automata [2] and uses a regular grid with evenly distributed nodes. The static and moving boundaries are modelled with the modified bounce back algorithm.

The calculation of the bubble movement is done by solving the Newton’s equation of motion. The sum of the forces acting on each bubble, i.e. the drag, the buoyancy, the lift force, the history force, the added mass effect and gravity, is used to determine the acceleration of the particle. The acceleration and the time step length give the velocity and the position change at the end of the time step. For coalescence and breakup stochastic models are used. The phases are coupled with a two way approach.

As the simulation results are produced on the graphic card, intermediate data can be visualized while the calculation is running. To simulate large reactors, the code has a multi GPU functionality, hence it can distribute the workload of the simulation on several graphic processors.

Validation is done with literature data as well as with results from holdup measurements of a 150l laboratory reactor.


[1]        http://www.nvidia.com/cuda

[2]        Sukop, M.: Lattice Boltzmann Modeling, Springer, Berlin, 2007.

Extended Abstract: File Not Uploaded