Engineering topics
Publications and source records attributed to Salas, M. D..
It is shown that the inviscid equations governing steady axisymmetric flow with swirl, admit solutions with closed streamlines. Results are obtained using two different numerical algorithms. The first is based on a multigrid method for nonlinear eigenvalue problems, while the second is based on a least squares formulation.
Convergence acceleration of iterative solutions of potential and Euler equations, based on Wynn's epsilon-algorithm, is demonstrated. The extra computational work, to apply the technique, is negligible, while the storage requirement is definitely affordable with the present computers, at least, for two-dimensional inviscid flow problems.
The breakdown of the conservative potential approximation occurs for all the airfoils tested. It develops as soon as the shock waves appear in the flow field. Since shock waves are not properly represented by the potential approximation, it is conjectured that the breakdown is due to the isentropic shock jump condition of the potential approximation.
Far-field boundary conditions for the Euler equations are formulated and applied to transonic lifting flow over an airfoil in an unbounded domain. An expansion of the linearized small-disturbance equation in the far field is developed and the leading-order term, corresponding to a point vortex representation for the airfoil, is retained. A comprehensive evaluation across the Mach number range of the procedure's effectiveness in eliminating dependence of the numerical results on the boundary extent is presented. Extension of the method to three dimensions is also outlined.
Explore the source record for details and available documents.
The conservative full-potential equation is used to study transonic flow over five airfoil sections. The results of the study indicate that once shock are present in the flow, the qualitative approximation is different from that observed with the Euler equations. The difference in behavior of the potential eventually leads to multiple solutions.
The incompressible axisymmetric steady Navier-Stokes equations and the Euler equations are solved numerically to model the breakdown of a vortex. Although the solutions obtained for the Euler equations show a 'vortex breakdown-like' structure, their behavior is very different from that of the Navier-Stokes equations. The details of the numerical algorithmms used are presented, and the results obtained are compared to those in the literature.
Explore the source record for details and available documents.
The Chebyshev spectral collocation method for the Euler gasdynamic equations is described. It is used with shock fitting to compute several two-dimensional gasdynamic flows. Examples include a shock/acoustic wave interaction, a shock/vortex interaction, and the classical blunt-body problem. With shock fitting, the spectral method has a clear advantage over second-order finite differences in that equivalent accuracy can be obtained with far fewer grid points.
The nonuniqueness problem occurring at transonic speeds with the conservative potential equation is investigated numerically. The study indicates that the problem is not an inviscid phenomenon, but results from approximate treatment of shock waves inherent in the conservative potential model. A new bound on the limit of validity of the conservative potential model is proposed.
Spectral methods for compressible flows are introduced in relation to finite difference and finite element techniques within the framework of the method of weighted residuals. Current spectral collociation methods are put into historical context. The basic concepts of Fourier spectral collocation methods are provided. Filtering strategies for shock-capturing approaches are also presented. Fourier shock-capturing techniques are evaluated using a one-dimensional, periodic astrophysical 'nozzle' problem.
A finite volume formulation for the Euler equations using Cartesian grids is presented and used to study complex two-dimensional configurations. The formulation extends methods developed for the potential equation to the Euler equations. Results using this approach for single element airfoils are shown to be competitive with and as accurate as other methods that employ mapped grids. Further, it is demonstrated that this method provides a simple and accurate procedure for solving flow problems involving multielement airfoils.
The conservative full-potential equation is used to study transonic flow over five airfoil sections. The results of the study indicate that once shock waves are present in the flow, the qualitative approximation is different from that observed with the Euler equations. The difference in behavior of the potential eventually leads to multiple solutions.
A finite volume formulation of the Euler equations using Cartesian grids is used to calculate the transonic flow over multielement airfoils and to use the resulting solutions to assess wall interference effects in wind tunnels. Available methods and recommendations for evaluating such effects, which are based on shifts in Mach number and angle of attack, are examined and the results are compared with measurements using the flapped supercritical SKF 1.1 airfoil. Based on the calculations, it is concluded that shifts in Mach number and angle of attack cannot by themselves account for viscous and wall effects on multielement airfoils at transonic speeds.
The Navier-Stokes equations represent an extremely good model of the physical phenomena encountered in most aeronautical problems. However, the computational resource needed to solve the Navier-Stokes equations are so large that even with today's supercomputers, it is necessary to make use of simpler models. A large number of external aerodynamic problems can be accurately described by a simpler model. This model consists of an outer inviscid flow plus a boundary-layer thickness correction for the vehicle shape. The outer inviscid model may be represented by the potential equation or by the Euler equation. The present paper provides the foundations for the numerical solution of the Euler equations. The governing equations are considered, taking into account conservation laws, the medium, the differential form of the conservation laws, generalized solutions, shock-fitting, and characteristics. Attention is also given to initial and boundary conditions, existence and uniqueness, and rotational phenomena.
Report developments in the application of spectral methods to two dimensional compressible flows are reviewed. A brief introduction to spectral methods - their history and especially their implementation - is provided. The stress is on those techniques relevant to transonic flow computation. The spectral multigrid iterative methods are discussed with application to the transonic full potential equation. Discontinuous solutions of the Euler equations are considered. The key element is the shock fitting technique which is briefly explained.
A procedure to study the local stability of planar shock waves is presented. The procedure is applied to a Rankine-Hugoniot shock in a divergent/convergent nozzle, to an isentropic shock in a divergent/convergent nozzle, and to Rankine-Hugoniot shocks attached to wedges and cones. It is shown that for each case, the equation governing the shock motion is equivalent to the damped harmonic oscillator equation.
The time dependent, isentropic, quasi-one-dimensional equations of gas dynamics and other model equations are considered under the constraint of characteristic boundary conditions. Analysis of the time evolution shows how different initial data may lead to different steady states and how seemingly anamolous behavior of the solution may be resolved. Numerical experimentation using time consistent explicit algorithms verifies the conclusions of the analysis. The use of implicit schemes with very large time steps leads to erroneous results.