SEARCH · Engineering Papers
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.
The advantages of a newly developed code are demonstrated by comparisons of the analysis with existing theoretical data as well as with other available finite element programs. The new program shows promise to significantly reduce the computer time. It also permits accurate and efficient structural analysis of engine hot section components.
A transonic equivalent strip (TES) method was further developed for unsteady flow computations of arbitrary wing planforms. The TES method consists of two consecutive correction steps to a given nonlinear code such as LTRAN2; namely, the chordwise mean flow correction and the spanwise phase correction. The computation procedure requires direct pressure input from other computed or measured data. Otherwise, it does not require airfoil shape or grid generation for given planforms. To validate the computed results, four swept wings of various aspect ratios, including those with control surfaces, are selected as computational examples. Overall trends in unsteady pressures are established with those obtained by XTRAN3S codes, Isogai's full potential code and measured data by NLR and RAE. In comparison with these methods, the TES has achieved considerable saving in computer time and reasonable accuracy which suggests immediate industrial applications.
The method of flux vector splitting used is that of Van Leer. The fluxes split in this manner have the advantage of being continuously differentiable at eigenvalue sign changes and this allows normal shocks to be captured with at most two interior zones, although in practice only one zone is usually observed. The fluxes as originally derived, however did not include the necessary terms appropriate for calculations on a dynamic mesh. The extension of the splitting to include these terms while retaining the advantages of the original splitting is the main purpose of this investigation. In addition, the use of multiple grids to reduce the computer time is investigated. A subiterative procedure to eliminate factorization and linearization error so that larger time steps can be used is also investigated.
Information concerning the programming, maintenance and operation of the G3DXL computer program is presented and the theoretical basis for the code is described. The program computes time domain scattering fields and surface currents and charges induced by a driving function on and within a complex scattering object which may be perfectly conducting or a lossy dielectric. This is accomplished by modeling the object with cells within a three-dimensional, rectangular problem space, enforcing the appropriate boundary conditions and differencing Maxwell's equations in time. In the present version of the program, the driving function can be either the field radiated by a lightning strike or a direct lightning strike. The F-106 B aircraft is used as an example scattering object.
Extendable Integrated Support Environment (EISE) is a real-time computer network consisting of commercially available hardware and software components to support systems level integration, modifications, and enhancement to weapons systems. The EISE approach offers substantial potential savings by eliminating unique support environments in favor of sharing common modules for the support of operational weapon systems. An expert system is being developed that will help support diagnosing faults in this network. This is a multi-level, multi-expert diagnostic system that uses experiential knowledge relating symptoms to faults and also reasons from structural and functional models of the underlying physical model when experiential reasoning is inadequate. The individual expert systems are orchestrated by a supervisory reasoning controller, a meta-level reasoner which plans the sequence of reasoning steps to solve the given specific problem. The overall system, termed the Diagnostic Executive, accesses systems level performance checks and error reports, and issues remote test procedures to formulate and confirm fault hypotheses.
Techniques which may be used to increase the animation update rate of real-time computer raster graphic displays are discussed. They were developed on the ADAGE RDS 3000 graphic system in support of the Advanced Concepts Simulator at the NASA Langley Research Center. These techniques involve the use of a special purpose parallel processor, for high-speed character generation. The description of the parallel processor includes the Barrel Shifter which is part of the hardware and is the key to the high-speed character rendition. The final result of this total effort was a fourfold increase in the update rate of an existing primary flight display from 4 to 16 frames per second.
The STAT program was designed for the optimization of single rotation, tractor propfan designs. New propfan designs, however, generally consist of two counter rotating propfan rotors. STAT is constructed to contain two levels of analysis. An interior loop, consisting of accurate, efficient approximate analyses, is used to perform the primary propfan optimization. Once an optimum design has been obtained, a series of refined analyses are conducted. These analyses, while too computer time expensive for the optimization loop, are of sufficient accuracy to validate the optimized design. Should the design prove to be unacceptable, provisions are made for recalibration of the approximate analyses, for subsequent reoptimization.
Lewis Research Center of NASA, with support from Rocketdyne, was engaged in non-real time computer simulation effort for the Space Station Freedom Electric Power System (EPS) EASY5, a simulation package, is used as the primary tool for this activity. Early in the design of the EPS, two test beds were set up at Lewis. The Integrated Test Bed (ITB), that combines and upgrades these test beds, is in the planning stage. The test beds are designed to functionally represent many of the components of the EPS and their interconnections. The simulation effort is primarily directed towards these test beds. Model verification is performed using test bed data.
A coupling method for efficient vibration mode analysis of aircraft with multiple external store configurations is presented. A set of low-frequency vibration modes, including rigid-body modes, represent the aircraft. Each external store is represented by its vibration modes with clamped boundary conditions, and by its rigid-body inertial properties. The aircraft modes are obtained from a finite-element model loaded by dummy rigid external stores with fictitious masses. The coupling procedure unloads the dummy stores and loads the actual stores instead. The analytical development is presented, the effects of the fictitious mass magnitudes are discussed, and a numerical example is given for a combat aircraft with external wing stores. Comparison with vibration modes obtained by a direct (full-size) eigensolution shows very accurate coupling results. Once the aircraft and stores data bases are constructed, the computer time for analyzing any external store configuration is two to three orders of magnitude less than that of a direct solution.
Methods for multirate digital control system design are discussed. A simple method for sampling rate selection based on control bandwidths is proposed. Methods for generating a discrete-time state model of a sampled-data plant and a discrete-time equivalent to an analog cost function for a sampled-data plant are described. The succesive loop closures and linear quadratic Gaussian synthesis methods are reviewed, and a constrained optimization synthesis method is introduced. The proposed sampling rate selection, discretization, and synthesis methods are applied to two example design problems. Multirate and single-rate compensators synthesized by the different methods are compared, based on closed-loop responses, with compensators having the same real-time computation load.
The shock-capturing integral-equation scheme developed by Kandil and Hu (1987) for the analysis of steady transonic flow over airfoils is extended to the unsteady case. The full potential formulation of the governing equations is reviewed; the solution method is outlined; and results for a NACA 0012 airfoil in forced pitching oscillation at Mach 0.755 are presented in extensive graphs and briefly characterized. The present technique is shown to require significantly less computation time than finite-difference or finite-volume methods, and to give shock-motion predictions in good agreement with those of an implicit finite-volume Euler solver; the surface-pressure peaks are slightly underpredicted.
A direct solution procedure for computing the flutter Mach number and the flutter frequency is applied to the aeroelastic analysis of propfans using a finite element structural model and an unsteady aerodynamic model based on a three-dimensional subsonic compressible lifting surface theory. An approximation to the Jacobian matrix that improves the efficiency of the iterative process is presented. The Jacobian matrix is indirectly approximated from approximate derivatives of the flutter matrix, which are updated only in the direction of the last move. Examples are used to illustrate the convergence properties. The direct solution procedure facilitates the automated flutter analysis in addition to contributing to the efficient use of computer time as well as the analyst's time.
An implicit procedure for solving parabolized Navier-Stokes equations under the assumption of a general equation of state for a gas in chemical equilibrium is given. A general and consistent approach for the evaluation of Jacobian matrices in the implicit operator avoids the use of unnecessary auxiliary quantities and approximations, and leads to a simple expression. Applications to two- and three-dimensional flow problems show efficiency in computer time and economy in storage.
A probabilistic structural analysis method (PSAM) is described which makes an approximate calculation of the structural response of a system, including the associated probabilistic distributions, with minimal computation time and cost, based on a simplified representation of the geometry, loads, and material. The method employs the fast probability integration (FPI) algorithm of Wu and Wirsching. Typical solution strategies are illustrated by formulations for a representative critical component chosen from the Space Shuttle Main Engine (SSME) as part of a major NASA-sponsored program on PSAM. Typical results are presented to demonstrate the role of the methodology in engineering design and analysis.
Similarity solutions are presented for the time-dependent evolution of cooling flows. A cooling flow expands when the cooling time computed from the initial gas distribution increases with radius. If certain simplifying conditions are met, the evolution of the resulting cooling wave becomes self-similar. The similarity solutions obtained here assumed subsonic flow, which is valid in the outer parts of observed cooling flows; they are matched to transonic accretion solutions valid in the central parts. The models are applied to the cooling flow onto M87, with unsatisfactory agreement at small radius. A local linear stability analysis shows that the similarity solution suffers isobaric thermal instability, suggesting that the neglect of star formation may be responsible for the poor agreement. The similarity solution suggests that the cooling flow around M87 was more vigorous in the past.
Two methods, fully- and loosely-coupled, are developed to incorporate nonequilibrium hydrogen-air chemistry into the fluid dynamic implicit flux vector splitting code (F3D). The new code (F3D/Chem) is validated against other existing codes for two cases: nozzle expansion, and shock-induced combustion around a blunt body. The shock-induced combustion case is compared also with an experimental data. The reaction rate constants are varied in an effort to reproduce the experimental data. The fully- and loosely-coupled methods are found to yield comparable results, but the computation time is shorter using the loosely-coupled method. The present method is found to reproduce results obtained using different existing codes. The experimental data was not reproduced with any selected rate coefficients set.
Numerical models based on the discrete exchange factor (DEF) and the zonal methods for radiative analysis of rocket engines containing a radiatively participating medium have been developed. These models implement a new technique for calculating the direct exchange factors to account for possible blockage by the nozzle throat. Given the gas and surface temperature distributions, engine geometry, and radiative properties, the models compute the wall radiative heat fluxes at different axial positions. The results of sample calculations for a typical rocket engine (engine 700 at NASA), which uses RP-1 (a kerosene-type propellant), are presented for a wide range of surface and gas properties. It is found that the heat transfer by radiation can reach up to 50 percent of that due to convection. The maximum radiative heat flux is at the inner side of the engine, where the gas temperature is the highest. While the results of both models are in excellent agreement, the computation time of the DEF method is found to be much smaller.
An analytical/numerical model for the deflection and mixing of a single gaseous jet in a supersonic crossflow is presented. The jet cross-section is described in terms of the compressible vortex pair resulting from viscous and impulsive forces acting at the jet periphery, and the vortex pair data are combined with data for the mass and momentum balance along the jet axis in order to model the trajectory and mixing of the injected fluid. A numerical technique is employed to solve for the inviscid outer flow and the position of the bow shock which envelopes the jet. The model is shown to be capable of predicting overall jet penetration (for perfectly or slightly underexpanded jets) to within 10 percent of experimental findings, while requiring only a few seconds of computer time.