Currently, neither built-in nor third-party support exists for computational fluid dynamics (CFD) post-processing techniques or the handling of common visualization file formats. This thesis presents a solution for visualizing massive-scale polygonal CFD datasets in Unity. The viability of immersive massive-scale CFD post-processing in Unity was examined through a mixed-reality application developed around this solution utilizing near-state-of-the-art hardware and data from a rapid compression expansion machine simulation. The performance of the solution was measured through benchmarks and the functionality of the application was determined through hands-on testing.
Multiple implemented optimizations are explored, including some improvements on approaches from the literature. These include the parallelization of multiple steps throughout the data pipeline and the use of a low-level rendering interface of Unity.
The developed visualization solution achieves runtime data loading times ranging from 2.9 to 47.1 seconds resulting in pre-processed transient datasets ranging in size from 1.54 to 25.95 gigabytes. Basic post-processing techniques were successfully implemented in the developed application. In benchmarking, the application reached the headset’s optimal frame rate of 90 in static viewing of individual objects and averaged a frame rate of over 58 during transient playback.