78th Annual Meeting of the APS Division of Fluid Dynamics (Nov 23 — 25, 2025)

P029: Visualizing the Structure of Pressure and Vorticity in Turbulence Simulations on a 32768³ Grid

• View Larger
• Download Original

We present new visualizations from a direct numerical simulation of isotropic turbulence at an unprecedented resolution of 32768³ grid points. Pressure fluctuations in turbulence, while not as intermittent as enstrophy or energy dissipation, are known to exhibit strong negative skewness, suggesting a novel structure. We find that regions of low-pressure seem to form highly elongated structures across a broad range of scales. These structures are compared and contrasted with the small 'worms' of vorticity found in previous studies. Additionally, by filtering the vorticity at the Taylor scale in regions of the flow where it is largest in magnitude, we observe an apparent correspondence between the pressure and filtered vorticity fields. For further details on our most recent simulations, see Yeung et al., J. Fluid Mech. (2025, in press).The images in this poster were generated by volume rendering, in which planes of data are mapped to an appropriate color and opacity that best highlights the structures of interest, stacked on top of each other with an offset, and composited together using the 'over' operation. Performing the rendering this way is much easier to implement in parallel on highly distributed systems compared to other techniques, and yields a classic oblique projection that is scale preserving, free of perspective distortions, and well suited for technical diagrams in particular. The largest of the images has been generated at an image resolution of 436912 (32768 plus a final offset of 32768/3 along both horizontal and vertical axes), and then carefully down-sampled to 109232 in such a way that preserves the smallest details near the limits of human perception.Supported by INCITE 2025 award of resources on the Frontier exascale supercomputer at the US Department of Energy Oak Ridge Leadership Computing Facility, and NSF subcontract via The Johns Hopkins Univ. (Grant 2103874). We thank the JHTDB project team (led by Prof. C. Meneveau) for hosting a subset of our 32768³ DNS data and providing valuable input for this visualization effort.