The evaluation of life safety in buildings in case of fire is often based on smoke spread calculations. However, recent simulation models – in general, based on computational fluid dynamics – often require long execution times or high-performance computers to achieve simulation results in or faster than real-time.
Therefore, the objective of this study is the development of a concept for the real-time and prognosis simulation of smoke propagation in compartments using a graphics processing unit (GPU). The developed concept is summarized in an expandable open source software basis, called JuROr (Jülich's Real-time simulation within ORPHEUS). JuROr simulates buoyancy-driven, turbulent smoke spread based on a reduced modeling approach using finite differences and a Large Eddy Simulation turbulence model to solve the incompressible Navier-Stokes and energy equations. This reduced model is fully adapted to match the target hardware of highly parallel computer architectures. Thereby, the code is written in the object-oriented programming language C++ and the pragma-based programming model OpenACC. This model ensures to maintain a single source code, which can be executed in serial and parallel on various architectures.
Further, the study provides a proof of JuROr's concept to balance sufficient accuracy and practicality. First, the code was successfully verified using unit and (semi-) analytical tests. Then, the underlying model was validated by comparing the numerical results to the experimental results of scenarios relevant for fire protection. Thereby, verification and validation showed acceptable accuracy for JuROr's application. Lastly, the performance criteria of JuROr – being real-time and prognosis capable with comparable performance across various architectures – was successfully evaluated. Here, JuROr also showed high speedup results on a GPU and faster time-to-solution compared to the established Fire Dynamics Simulator. These results show JuROr's practicality.