Georgia Institute of Technology
Investigations of Lagrangian Intermittency and Inertial Particle Dynamics in Turbulence Using Extreme-Scale Computing
Abstract
dc:description.abstractFrom pollutant transport in the atmosphere or the ocean to fuel-air mixing in a combustion engine, the Lagrangian perspective of following individual particle trajectories is fundamental to studies involving turbulent dispersion and transport. This thesis develops a high-resolution direct numerical simulation (DNS) framework to investigate intermittency and inertial particle dynamics in the Lagrangian frame. A GPU-accelerated parallel algorithm is presented for tracking fluid and inertial particles in DNS of isotropic turbulence, scaling up to 35 trillion grid points and 1 billion particles on Frontier, the world's first exascale supercomputer. The key algorithmic elements are cubic spline interpolation, a local decomposition via particle migration for particles moving across sub-domains, and ghost layers to enable interpolation for particles close to sub-domain boundaries. The particle tracking algorithm's memory footprint is carefully optimized to be able to run the largest simulation size at nearly the full system capacity on Frontier. Detailed performance analysis shows that the algorithm achieves good strong and weak scaling at a cost nearly independent of particle count. The all-to-all communication required in distributed pseudo-spectral DNS algorithms is the dominant cost in the overall elapsed wall time per step. A peak performance model is developed to determine an upper bound for the achievable all-to-all performance on a given platform. Using this computational foundation, three research directions are pursued. The Lagrangian velocity increment and its statistical properties are crucial in characterizing small-scale intermittency in the Lagrangian frame. A fundamental relation in Lagrangian Kolmogorov theory is concerned with inertial range scaling of the second-order velocity structure function over intermediate time lags at sufficiently high Reynolds numbers. While significant theoretical support for asymptotic constancy of the scaling constant is known, unambiguous testing of the scaling is challenging. In this thesis, DNS of forced isotropic turbulence performed on the Frontera supercomputer at Taylor-scale Reynolds numbers 140 to 1300 is used to examine the inertial range scaling of the second-order Lagrangian velocity structure function. The velocity increment is decomposed into convective and local contributions, which are subject to a strong but incomplete mutual cancellation dependent on Reynolds number and time lag. Results show that a narrower range of temporal versus spatial scales, together with particle displacement effects, governs the observed truncated scaling in the Lagrangian inertial range and that the value of the scaling constant is likely higher than prior estimates in the literature. In addition to the velocity increments, the statistics of the velocity gradients play a central role in investigations of small-scale intermittency. Here, the conditional statistics of the pressure Hessian contracted with the velocity gradient tensor, given its second and third invariants, are studied at Taylor-scale Reynolds numbers 390 to 2500, to characterize the pressure Hessian's influence on the velocity gradient dynamics in the invariant space. Finally, particles with inertia are known to depart from fluid trajectories and accumulate in regions of either high strain rates or rotation rates in a phenomenon known as preferential concentration. In this thesis, inertial particles with Stokes drag and added-mass effects are tracked across a range of parameter values, at Taylor-scale Reynolds numbers 390 and 650 to study preferential sampling and clustering. It is shown that preferential sampling strength varies non-monotonically with Stokes number for light and heavy particles, and the peaks of spatial clustering and preferential sampling occur at different Stokes numbers.
Degree
thesis:*- Name thesis:degree_name
- Aerospace Engineering, PhD
- Department thesis:degree_department
- Aerospace Engineering
- Grantor
- Georgia Institute of Technology
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Vaideswaran, Rohini
- Advisor dc:contributor.advisor
-
- Yeung, Pui-Kuen
- Committee members dc:contributor.committeemember
-
- Meneveau, Charles
- Menon, Suresh
- Lai, Chung Kei Chris
- Tsikata, Sedina
Rights
- Language dc:language.iso
- English
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/82138