Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Eigenvalue”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 937 records · Page 52
All sourcesNASA NTRSDOE OSTI
NASA NTRS1986

Multigrid method for nearly singular and slightly indefinite problems

This paper deals with nearly singular, possibly indefinite problems for which the usual multigrid solvers converge very slowly or even diverge. The main difficulty is related to some badly approximated smooth functions which correspond to eigenfunctions with nearly zero eigenvalues. A correction to the usual coarse-grid equations is derived, both in the correction scheme and in the full approximation scheme. The performance of the new algorithm using this correction is essentially as that of usual multigrid for definite problems.

Brandt, A.↗
NASA NTRS1987

Vortex breakdown simulation based on a nonlinear inviscid model

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.

Hafez, M. M.↗
NASA NTRS1987

Semi-implicit and fully implicit shock-capturing methods for hyperbolic conservation laws with stiff source terms

Some numerical aspects of finite-difference algorithms for nonlinear multidimensional hyperbolic conservation laws with stiff nonhomogeneous (source) terms are discussed. If the stiffness is entirely dominated by the source term, a semi-implicit shock-capturing method is proposed provided that the Jacobian of the source terms possesses certain properties. The proposed semi-implicit method can be viewed as a variant of the Bussing and Murman point-implicit scheme with a more appropriate numerical dissipation for the computation of strong shock waves. However, if the stiffness is not solely dominated by the source terms, a fully implicit method would be a better choice. The situation is complicated by problems that are higher than one dimension, and the presence of stiff source terms further complicates the solution procedures for alternating direction implicit (ADI) methods. Several alternatives are discussed. The primary motivation for constructing these schemes was to address thermally and chemically nonequilibrium flows in the hypersonic regime. Due to the unique structure of the eigenvalues and eigenvectors for fluid flows of this type, the computation can be simplified, thus providing a more efficient solution procedure than one might have anticipated.

Yee, H. C.↗
NASA NTRS1987

Removal of spurious modes encountered in solving stability problems by spectral methods

A technique based on the Galerkin approximation is developed to remove spurious roots arising when Chebyshev spectral methods are used to solve eigenvalue problems in hydrodynamic stability. The derivation of Galerkin-Chebyshev approximations is explained, and numerical results for the Orr-Sommerfeld equations of plane Poiseuille flow and a Blasius profile are presented in tables and compared with those obtained by the method of Zebib (1984). It is pointed out that the present method does not increase the size of the algebraic system to be solved.

Zebib, Abdelfattah↗
NASA NTRS1987

Modal decoupling insensitivity

This note deals with the design of an output feedback control system such that selected closed-loop eigenvalues are insensitive, while the corresponding eigenvectors preserve a prescribed decoupling under parameter variation. This results in restricting specified elements of an eigenvector to be insensitive, rather than the entire eigenvector. This requirement is henceforth referred to as modal decoupling insensitivity. Sufficient conditions for modal decoupling insensitivity are derived. These conditions offer the advantage of increased flexibility in eigenvector assignment, but the constraint on the feedback matrix is more restrictive. A numerical example is presented to illustrate the design procedure.

Raman, K. V.↗
NASA NTRS1987

The normal modes of the thermosphere

The linearized momentum, energy, and continuity equations for the thermosphere can be reduced to a form that gives the vertical structure for each horizontal wave mode. The vertical structure equation can be described in terms of the normal modes, or eigenmodes, of the thermosphere. The latter are obtained by using a 27-layer model that includes a realistic temperature profile and the effects of the Lorentz force, viscosity, and heat conduction. The normal modes have one real eigenfrequency for every two complex conjugate eigenfrequency values. The real modes have a dominant rotational wind component and are nonpropagating. The complex modes have comparable divergent and rotational wind components. The complex eigenvalues give vertically propagating modes, primarily associated with the transient response to forcing, and are significantly affected by the dissipation in the upper E region and F region. Results show that the rotational wind component dominates in the steady state when the forcing is due to the two-cell convection pattern at high latitudes and that the normal modes explain the large shears and large winds speeds that are typically observed in the high-latitude E region. The vertical energy flux for the normal modes is also calculated. The results show that the flux is upward above 130 km but downward in the lower E region for the total solution. The downward energy flux is a contribution from the real eigenmode structure.

Larsen, M. F.↗
NASA NTRS1987

The characteristic Rossby frequency

The characteristic Rossby frequency is defined for a fixed zonal wavenumber perturbation as the variational integral of the Rayleigh-Ritz method. It is a measure of the time scale of the disturbance. For a disturbance which locally has the shape of an eigenfunction but is not global in extent, the characteristic Rossby frequency is very close to the true eigenvalue, and additionally remains unchanged under linear inviscid dynamics. Results are presented for the shallow water equations, both with and without a mean zonal wind. The characteristic Rossby frequency of a wavenumber 1 perturbation having the shape of the second symmetric Rossby mode but confined to the Northern Hemisphere is close to the corresponding Rossby frequency. This finding is helpful in understanding the behavior of the observed wavenumber 1 pattern of January 1979, which propagated westward with nearly the pure Rossby frequency but was discernible only in the Northern Hemisphere (as discussed by Daley and Williamson).

Straus, David M.↗
NASA NTRS1985

Implementation of design sensitivity analysis with existing finite element codes

A numerical method is presented to implement structural design sensitivity analysis theory, using the versatility and convenience of existing finite element structural analysis programs. Design variables such as thickness and cross-sectional areas of components of individual members and built-up structures are considered. Structural performance functionals considered include displacement and stress. The method is also applicable for eigenvalue problem design sensitivity analysis. It is shown that calculations can be carried out outside existing finite element codes, using postprocessing data only. Thus, design sensitivity analysis software does not have to be embedded in an existing finite element code. Feasability of the method is shown through analysis of several problems, including a built-up structure. Accurate design sensitivity results are obtained without the uncertainty of numerical accuracy associated with selection of finite difference perturbations.

Choi, Kyung K.↗
NASA NTRS1987

Stability of viscous flow past a circular cylinder

A spectral method which employs trigonometric functions and Chebyshev polynomials is used to compute the steady, incompressible laminar flow past a circular cylinder. Linear stability methods are used to formulate a pair of decoupled generalized eigenvalue problems for the growth of symmetric and asymmetric (about the dividing streamline) perturbations. It is shown that, while the symmetric disturbances are stable, the asymmetric perturbations become unstable at a Reynolds number about 40 with a Strouhal number about 0.12. The critical conditions are found to depend on the size of the computational domain in a manner similar to that observed in the laboratory.

Zebib, A.↗
NASA NTRS1987

On the acceleration of charged particles at relativistic shock fronts

The diffusive acceleration of highly relativistic particles at a shock is reconsidered. Using the same physical assumptions as Blandford and Ostriker (1978), but dropping the restriction to nonrelativistic shock velocities, the authors find approximate solutions of the particle kinetic equation by generalizing the diffusion approximation to higher order terms in the anisotropy of the particle distribution. The general solution of the transport equation on either side of the shock is constructed, which involves the solution of an eigenvalue problem. By matching the two solutions at the shock, the spectral index of the resulting power law is found by taking into account a sufficiently large number of eigenfunctions. Low-order truncation corresponds to the standard diffusion approximation and to a somewhat more general method described by Peacock (1981). In addition to the energy spectrum, the method yields the angular distribution of the particles and its spatial dependence.

Kirk, J. G.↗
NASA NTRS1987

A new method to real-normalize measured complex modes

A time domain subspace iteration technique is presented to compute a set of normal modes from the measured complex modes. By using the proposed method, a large number of physical coordinates are reduced to a smaller number of model or principal coordinates. Subspace free decay time responses are computed using properly scaled complex modal vectors. Companion matrix for the general case of nonproportional damping is then derived in the selected vector subspace. Subspace normal modes are obtained through eigenvalue solution of the (M sub N) sup -1 (K sub N) matrix and transformed back to the physical coordinates to get a set of normal modes. A numerical example is presented to demonstrate the outlined theory.

Wei, Max L.↗
NASA NTRS1987

Simultaneous structural and control optimization via linear quadratic regulator eigenstructure assignment

A method for simultaneous structural and control design of large flexible space structures (LFSS) to reduce vibration generated by disturbances is presented. Desired natural frequencies and damping ratios for the closed loop system are achieved by using a combination of linear quadratic regulator (LQR) synthesis and numerical optimization techniques. The state and control weighing matrices (Q and R) are expressed in terms of structural parameters such as mass and stiffness. The design parameters are selected by numerical optimization so as to minimize the weight of the structure and to achieve the desired closed-loop eigenvalues. An illustrative example of the design of a two bar truss is presented.

Becus, G. A.↗
NASA NTRS1987

Optimal control of large space structures via generalized inverse matrix

Independent Modal Space Control (IMSC) is a control scheme that decouples the space structure into n independent second-order subsystems according to n controlled modes and controls each mode independently. It is well-known that the IMSC eliminates control and observation spillover caused when the conventional coupled modal control scheme is employed. The independent control of each mode requires that the number of actuators be equal to the number of modelled modes, which is very high for a faithful modeling of large space structures. A control scheme is proposed that allows one to use a reduced number of actuators to control all modeled modes suboptimally. In particular, the method of generalized inverse matrices is employed to implement the actuators such that the eigenvalues of the closed-loop system are as closed as possible to those specified by the optimal IMSC. Computer simulation of the proposed control scheme on a simply supported beam is given.

Nguyen, Charles C.↗
NASA NTRS1988

Development of an unsteady wake theory appropriate for aeroelastic analyses of rotors in hover and forward flight

The purpose of this research is the development of an unsteady aerodynamic model for rotors such that it can be used in conventional aeroelastic analysis (e.g., eigenvalue determination and control system design). For this to happen, the model must be in a state-space formulation such that the states of the flow can be defined, calculated and identified as part of the analysis. The fluid mechanics of the problem is given by a closed-form inversion of an acceleration potential. The result is a set of first-order differential equations in time for the unknown flow coefficients. These equations are hierarchical in the sense that they may be truncated at any number of radial or azimuthal terms.

Peters, David A.↗
NASA NTRS1986

Numerical simulation of a controlled boundary layer

The problem of interest is the boundary layer over a flat plate. The three standard laminar flow control (LFC) techniques are pressure gradient, suction, and heating. The parameters used to describe the amount of control in the context of the boundary layer equations are introduced. The numerical method required to find the mean flow, the linear eigenvalues of the Orr-Sommerfeld equation, and the full, nonlinear, 3-D solution of the Navier-Stokes equations are outlined. A secondary instability exists for the parallel boundary subject to uniform pressure gradient, suction, or heating. Selective control of the spanwise mode reduces the secondary instability in the parallel boundary layer at low Reynolds number.

Zang, Thomas A.↗
NASA NTRS1987

Parallels between control PDE's (Partial Differential Equations) and systems of ODE's (Ordinary Differential Equations)

System theorists understand that the same mathematical objects which determine controllability for nonlinear control systems of ordinary differential equations (ODEs) also determine hypoellipticity for linear partial differentail equations (PDEs). Moreover, almost any study of ODE systems begins with linear systems. It is remarkable that Hormander's paper on hypoellipticity of second order linear p.d.e.'s starts with equations due to Kolmogorov, which are shown to be analogous to the linear PDEs. Eigenvalue placement by state feedback for a controllable linear system can be paralleled for a Kolmogorov equation if an appropriate type of feedback is introduced. Results concerning transformations of nonlinear systems to linear systems are similar to results for transforming a linear PDE to a Kolmogorov equation.

Hunt, L. R.↗
NASA NTRS1988

Further developments in exact state reconstruction in deterministic digital control systems

A more general version of the ideal state reconstructor for deterministic digital control systems previously developed is presented. In the original version, measurements prefiltered by a multi-input/multi-output moving-average (MA) process were utilized in the state reconstruction process. In this version, the MA-prefiltered measurements can be supplemented by standard instantaneous measurements. The ideal state reconstructor is so named because: if the plant parameters are known exactly, its output will exactly equal the true state of the plant, not just approximate it. Furthermore, it adds no additional states or eigenvalues to the system. Nor does it affect the plant equation for the system in any way; it affects the measurement equation only. An example is presented which illustrates the procedure for choosing the parameters in it.

Polites, Michael E.↗
NASA NTRS1988

Reduced order models for nonlinear aerodynamics

Reduced order models are needed for reliable, efficient and accurate prediction of aerodynamic forces to analyze fluid structure interaction problems in turbomachinery, including propfans. Here, a finite difference, time marching Navier-Stokes code is validated for unsteady airfoil motion by comparing results with those from classical potential flow. The Navier-Stokes code is then analyzed for calculation of primitive and exact estimates of eigenvalues and eigenvectors associated with fluid-airfoil interaction. A variational formulation for the Euler equations and Navier-Stokes equations will be the basis for reduction of order through an eigenvector transformation.

Mahajan, Aparajit J.↗