Wake Induced Boundary Layer Transition in Turbomachinery
Principal Investigator: Prof. Paul Durbin, Stanford University

The Center for Integrated Turbulence Simulations (CITS) is a multidisciplinary organization established in July 1997 at Stanford University to develop new turbulence models and associated numerical simulation methodologies that will enable a new paradigm for the design of advanced systems in which turbulence plays a controlling role. The initial focus of CITS is a comprehensive program on gas turbine engines, supported by ASCI Strategic Alliance Program.

The CITS is closely linked with the Stanford/NASA Center for Turbulence Research (CTR). CTR has pioneered the use of Direct Numerical Simulations (DNS) for study of turbulence physics and assessment of turbulence models, and has led in the development of Large Eddy Simulation (LES) for high Reynolds number turbulent flows. CITS will integrate developments from the CTR with advances in other disciplines to enable the complex simulations needed for advanced system design.

CITS research focuses initially on the compressor, combustor, and turbine components of aircraft gas turbine engines. Subsequently, the component simulations will be integrated to simulate situations in which component interactions are crucial. Unsteady phenomena of interest in these simulations include rotating stall in the compressor, aeroelastic blade vibrations, instabilities in the combustor due to heat release coupling with the acoustic modes, and the heat transfer from the hot combustion products to the first blade rows of the turbine.

In a paper recently submitted to the Journal of Fluid Mechanics, Prof. P. Durbin, Dr. X. Wu and their collaborators at CITS describe the work they have done at PSC so far. Airfoil interactions in the passages of axial turbomachines result in a complex and inherently unstable flow field. Specifically, the boundary layer over a rotor suction surface is subjected to substantial unsteadiness that stems from impinging wakes of the upstream stator or rotor. While it is very difficult to use physical experiments to generate benchmark data for studying the detailed mechanisms of this "wake induced transition problem," the authors report significant progress towards generating such datasets by DNS. Working closely with Ravi Subramanya, a researcher at the Pittsburgh Supercomputing Center, they were able to carry out spatially developing, 3D, time-accurate DNS of boundary layer transition induced by periodically passing wakes, on PSC's CRAY T3E. The 3D, time-dependent Navier-Stokes equations were solved using 52.4 million grid points, so that essential fluctuation scales inside the boundary layer were resolved. This is one of the largest grids ever reported in wall-bounded transition and turbulence simulations. Future modeling efforts for a variety of systems will benefit both from the data generated in this work, and from the computational methodologies that have been developed and tested.

The figures show a 3D volume rendering of wakes impinging on the boundary layer of a flat plate. The flat plate is the vertical surface of the box shown in the bottom of the figures. The transition from laminar flow on the left, to turbulent flow on the right side of the plate is clearly visible. The bottom figure highlights the turbulence by hiding the regions of laminar flow. The study seeks to determine the mechanism that triggers the transitioning of the boundary layer flow from laminar to turbulent. Ongoing work at PSC seeks to study the problem for periodic wakes spanning a range of frequencies.