ENGINEERING AND CFD

Applications in the field of computational fluid dynamics (CFD) often demand high resolutions and thus computational resources, in particular regarding memory and arithmetic operations. The waLBerla framework is designed for massively parallel simulations of different applications from CFD. waLBerla is a lattice Boltzmann based numerical fluid flow software framework for the simulation of numerous physical applications, e.g., blood flow in the human heart or moving obstacles representing cells or bacteria in a fluid. The framework is developed carefully for outstanding single-core performance as well as excellent scalability. Especially on the large scales, this approach enables to utilize the scarce and expensive computing resources most efficiently and allows for domain sizes otherwise not achievable on current supercomputers.

(1) One project using waLBerla investigates collective swarming behavior of numerous self-propelled microorganisms at low Reynolds numbers (Stokes flow), e.g., a swarm of Escherichia coli (E. coli) bacteria. In this project, waLBerla is consistently coupled with the researachers's rigid body simulation tool pe (physics engine). This allows the scientists to simulate self-propelled devices consisting of fully resolved rigid bodies of arbitrary shape in 3D. An effortless exchange of the constituents of the considered micro devices, adapting the surrounding channel geometry according to the specific needs, or regarding regimes beyond low Reynolds numbers are only some of the potential benefits associated with the use of this coupled software framework.

(2) Another project implemented within waLBerla is the simulation of electron beam melting, an additive manufacturing method. Electron beam melting is used to produce successive layers of a part in a powder bed and offers the ability to produce components closest to their final dimensions with good surface finish. Currently, this process is faster than any other technique of comparable quality. However, the parts are not produced at sufficient rate to make them economically viable for any but very high value applications. One key output of the project is the knowledge surrounding the use of the high powder electron beam gun, including the process control, and a modelled and validated understanding of beam-powder bed interaction. The outcome of the simulations are compared with real experimental data and therefore the model parameters are adjusted in such a way that the resulting numerical melt pool sizes correspond to the experimental ones.

(3) A third project is motivated by the increasing importance of lab-on-a-chip (LoC) systems. The great interest in LoC systems is attributable to the fact that they can be used as portable biological analysis devices for point-of-care diagnostics. Electro-osmosis and electrophoresis are the mechanisms of choice for microfluidic manipulation and actuation in LoC devices. At the small scales of LoC systems, measurements of the flow are very difficult. Thus, simulations are required for the design and optimization of those systems. In order to capture the multiple physical effects accurately at small scale a very fine discretization and small time steps are necessary, resulting in the need for a large amount of computational resources. In the scope of this project, the separation of charged macromolecules in electrolyte solutions inside channels of dimensions relevant for LoC is simulated.

HPC system SuperMUC of the Leibniz Supercomputing Centre served as computing platform for all three projects.

References:

• Ammer, R., Markl, M., Ljungblad, U., Körner, C., & Rüde, U. (2014). Simulating fast electron beam melting with a parallel thermal free surface lattice Boltzmann method. Computers & Mathematics with Applications, 67(2), 318-330. doi:10.1016/j.camwa.2013.10.001
• Bartuschat, D., & Rüde, U. (2015). Parallel Multiphysics Simulations of Charged Particles in Microfluidic Flows. Journal of Computational Science, 8(0), 1-19. doi:10.1016/j.jocs.2015.02.006
• Godenschwager, C., Schornbaum, F., Bauer, M., Köstler, H., & Rüde, U. (2013). A Framework for Hybrid Parallel Flow Simulations with a Trillion Cells in Complex Geometries. Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis (pp. 35:1-35:12). Denver, Colorado: ACM. doi:10.1145/2503210.2503273
• Pickl, K., Hofmann, M., Preclik, T., Köstler, H., Smith, A.-S., & Rüde, U. (2014). Parallel Simulations of Self-propelled Microorganisms. In M. Bader, A. Bode, H.-J. Bungartz, M. Gerndt, G. R. Joubert, & F. Peters, Parallel Computing: Accelerating Computational Science and Engineering (CSE) (pp. 395-404). München: IOS Press. doi:10.3233/978-1-61499-381-0-395

Scientific Contact:

PD Dr.-Ing. habil. Harald Köstler
Universität Erlangen-Nürnberg
Lehrstuhl Informatik 10 (Systemsimulation)
Cauerstraße 11, D-91058 Erlangen (Germany)
e-Mail: harald.koestler@informatik.uni-erlangen.de