Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Cartesian and contravariant formulations of the compressible Navier-Stokes equations applied to turbomachinery flows

Thesis/Dissertation ·
OSTI ID:6376260

The contravariant components of the full, two-dimensional, compressible Navier-Stokes (NS) equations are derived from Cartesian components and from tensor calculus. The Cartesian, full and thin-layer, components of the NS equations; in Chain Rule Conservation Law Form (CRCLF) are employed for solution of a variety of turbomachinery related problems. For numerical solution the Beam-Warming implicit scheme is employed. One-sided first and second-order accurate expressions are used for the streamwise convective derivatives. A modification of the FLARE approximation, consistent with the eigenvalues of the inviscid linearization Jacobians, is employed in regions of separated flows for the thin-layer equations. Results are presented for a variety of channel and cascade flows. Using the Cartesian CRCLF component, the contravariant decomposition is performed numerically within the code. The effect of Reynolds numbers, and different thickness/chord ratios for constricted channels and cascades are studied. Second-order accurate, one-sided differences show improved accuracy, compared with first-order accurate but require smaller time steps for stability. The total enthalpy variation, for Mach numbers ranging from 0.3 to 0.5, is small for the geometries considered. Results for circular cascades agree well with the results of other investigators. A comparison of the thin-layer approximation employing the Cartesian and contravariant components, shows lightly improved results when using the contravariant components. An analytical evaluation of the difference between Cartesian and contravariant thin-layer approximations is performed. Improved results using the body fitted NS equations are expected with closely spaced, thick cascade blades in low Reynolds number flows.

Research Organization:
Cincinnati Univ., OH (USA)
OSTI ID:
6376260
Country of Publication:
United States
Language:
English