Shriram Jagannathan

Massively parallel direct numerical simulations of forced compressible turbulence: a hybrid MPI/OpenMP approach

A highly scalable simulation code for turbulent flows which solves the fully compressible Navier-Stokes equations is presented. The code, which supports one, two and three dimensional domain decompositions is shown to scale well on up to 262,144 cores. Introducing multiple levels of parallelism based on distributed message passing and shared-memory paradigms results in a reduction of up to 33% of communication time at large core counts. The code has been used to generate a large database of homogeneous isotropic turbulence in a stationary state created by forcing the largest scales in the flow. The scaling of spectra of velocity and density fluctuations are presented. While the former follow classical theories strictly valid for incompressible flows, the latter presents a more complicated behavior. Fluctuations in velocity gradients and derived quantities exhibit extreme though rare fluctuations, a phenomenon known as intermittency. The simulations presented provide data to disentangle Reynolds and Mach number effects.

Partially-Averaged Navier-Stokes (PANS) Simulations of Lid-Driven Cavity Flow—Part II: Flow Structures

The vortical flow structures in low and high Reynolds number lid-driven cavity flows are examined using Partially-averaged Navier-Stokes (PANS) and unsteady Reynolds-averaged Navier-Stokes (URANS) simulations. The spanwise aspect ratio (SAR) of the cavity is 3:1:1 and the Reynolds numbers based on cavity height and lid velocity are \(10^4\) and \(10^6\). It is demonstrated that, while the mean flow statistics are nearly the same for URANS and PANS, the complex vortex structures are captured much better by PANS. The difference between the URANS and PANS structures are even more distinct at higher Reynolds numbers. Furthermore, it is shown that the PANS small-scale statistics at different levels of resolution are self-similar and scale according to established turbulence theory.

Low Pressure Turbine Relaminarization Bubble Characterization using Massively-Parallel Large Eddy Simulations

The separation and reattachment of suction surface boundary layer in a low pressure turbine is characterized using large-eddy simulation at Ress = 69000 based on inlet velocity and suction surface length. Favorable comparisons are drawn with experiments using a high pass filtered Smagorinsky model for sub-grid scales. The onset of time mean separation is at s/so = 0.61 and reattachment at s/so = 0.81, extending over 20% of the suction surface. The boundary layer is convectively unstable with a maximum reverse flow velocity of about 13% of freestream. The breakdown to turbulence occurs over a very short distance of suction surface and is followed by reattachment. Turbulence near the bubble is further characterized using anisotropy invariant mapping and time orthogonal decomposition diagnostics. Particularly the vortex shedding and shear layer flapping phenomena are addressed. On the suction side, dominant hairpin structures near the transitional and turbulent flow regime are observed. The hairpin vortices are carried by the freestream even downstream of the trailing edge of the blade with a possibility of reaching the next stage. Longitudinal streaks that evolve from the breakdown of hairpin vortices formed near the leading edge are observed on the pressure surface.

Massively-Parallel DNS and LES of Turbine Vane Endwall Horseshoe Vortex Dynamics and Heat Transfer

Higher turbine inlet temperatures enable increased gas turbine efficiency but significantly reduce component lifetimes through melting of the blade and endwall surfaces. This melting is exacerbated by the horseshoe vortex that forms as the boundary layer stagnates in front of the blade, driving hot gasses to the surface. Furthermore, this vortex exhibits significant dynamical motions that increase the surface heat transfer above that of a stationary vortex. To further understand this heat transfer augmentation, the dynamics of the horseshoe vortex must be characterized in a 3D time-resolved fashion which is difficult to obtain experimentally. In this paper, a 1st stage high pressure stator passage is examined using a spectral element direct numerical simulation at a Reynolds number Re = U∞ C/v = 10,000 . Although the Re is lower than engine conditions, the vortex already exhibits similar strong aperiodic motions and any uncertainty due to sub-grid scale modeling is avoided. The vortex dynamics are analyzed and their impact on the surface heat transfer is characterized. Results from a baseline case with a smooth endwall are also compared to a passage with film cooling holes. Higher Reynolds number simulations require a Large Eddy Simulation turbulent viscosity model that can handle the high accelerations around the blade. A high-pass-filter sub-grid scale model is tested at the same low Reynolds number to test its effectiveness by direct comparisons to the DNS. This resulted in a significant drop in turbulence intensity due to the high strain rate in the freestream, resulting in different dynamics of the vortex than observed in the DNS. Appropriate upstream engine conditions of high freestream turbulence and large integral length scales for all cases are generated via a novel inflow turbulence development domain using a periodic solution of Taylor vortices that are convected over a square grid. The size of the vortices and grid spacing is used to control the integral length scale, and the intensity of the vortices and upstream distance is used to control the turbulence intensity. The baseline DNS exhibits a bi-modal horseshoe vortex, and the presence of cooling-holes qualitatively increases the number of vortex cores resulting in more complex interactions.Copyright