link to homepage

Institute for Advanced Simulation (IAS)

Navigation and service

Fire Dynamics Group

Plume Movie

The modelling of fire and smoke spread in buildings using computational fluid dynamics (CFD) methods became very popular in fire safety engineering and science. Increasingly accurate geometry and phenomena modeling are the benefits of those methods, however they are in general numerically and computationally very demanding. Three dimensional turbulent reactive flows with radiative, convective and conductive heat transfer have to be modeled on length scales ranging from less than one millimetre to several hundred metres: The use of high performance computing (HPC) suggests itself.

The fire dynamics group emerged 2012 within the division "Civil Safety and Traffic" of the Jülich Supercomputing Centre. The division is closely coupled to the Bergische Universität Wuppertal, where the group contributes to various teaching activities and supervision of Bachelor, Master and PhD students. The primary scientific goal of the group is the development and validation of models and methods in the field of fire dynamics.

All our work, i.e. methods, software and data sets, is freely available either directly via the webpages or on request.

Research Topics

The research topics of this group cover the application of numerical and statistical methods on complex infrastructures, like underground stations, within the ORPHEUS project. This includes also the development of tools to automate and ease the use of HPC facilities.

Besides the pure application of simulation software like the Fire Dynamics Simulator (FDS) or FireFOAM, we develop new numerical models in the context of smoke and fire dynamics. The two novel approaches are the usage of graphics processing units (GPU) and adaptive mesh refinement (AMR) for flow simulation.

Adaptive Mesh Refinement

Currently, we started experimental work to provide high accurate data for model validation. The focus lies on buoyancy driven flows in compartments and the application of non-invasive measurement techniques. Therefore we make use of the particle image velocimetry (PIV) and laser Doppler velocimetry (LDV) to measure spatio-temporal velocity fields.



We are recently involved in the BMBF funded project ORPHEUS and our computing time is granted by the "Kommission zur Vergabe von Supercomputing Ressourcen" (VSR) with the project names jjsc11 and jjsc27.