ALBERT

Description: Basic algorithms for unstructured mesh generation and fluid flow calculation are discussed. In particular the following are addressed: preliminaries of graphs and meshes; duality and data structures; basic graph operations important in CFD (Computational Fluid Dynamics); triangulation methods, including Varonoi diagrams and Delaunay triangulation; maximum principle analysis; finite volume schemes for scalar conservation law equations; finite volume schemes for the Euler and Navier-Stokes equations; and convergence acceleration for steady state calculations.

Description: It is shown that to satisfy the general accepted compressible law of the wall derived from the Van Driest transformation, turbulence modeling coefficients must actually be functions of density gradients. The transformed velocity profiles obtained by using standard turbulence model constants have too small a value of the effective von Karman constant kappa in the log-law region (inner layer). Thus, if the model is otherwise accurate, the wake component is overpredicted and the predicted skin friction is lower than the expected value.

Description: The compressible dynamic stall flowfield over a NACA 0012 airfoil transiently pitching from 0 to 60 deg at a constant rate under compressible flow conditions has been studied using real-time interferometry. A quantitative description of the overall flowfield, including the finer details of dynamic stall vortex formation, growth, and the concomitant changes in the airfoil pressure distribution, has been provided by analyzing the interferograms. For Mach numbers above 0.4, small multiple shocks appear near the leading edge and are present through the initial stages of dynamic stall. Dynamic stall was found to occur coincidentally with the bursting of the separation bubble over the airfoil. Compressibility was found to confine the dynamic stall vortical structure closer to the airfoil surface. The measurements show that the peak suction pressure coefficient drops with increasing freestream Mach number, and also it lags the steady flow values at any given angle of attack. As the dynamic stall vortex is shed, an anti-clockwise vortex is induced near the trailing edge, which actively interacts with the post-stall flow.

Description: The effect of the porous leading edge of an airfoil on the blade-vortex interaction noise, which dominates the far-field acoustic spectrum of the helicopter, is investigated. The thin-layer Navier-Stokes equations are solved with a high-order upwind-biased scheme and a multizonal grid system. The Baldwin-Lomax turbulence model is modified for considering transpiration on the surface. The amplitudes of the propagating acoustic wave in the near field are calculated directly from the computation. The porosity effect on the surface is modeled in two ways: (1) imposition of prescribed transpiration velocity distribution and (2) calculation of transpiration velocity distribution by Darcy's law. Results show leading-edge transpiration can suppress pressure fluctuations at the leading edge during blade-vortex interaction and consequently reduce the amplitude of propagating noise by 30% at a maximum in the near field.

Description: A method has been developed for calculating the viscous flow about airfoils with and without deflected flaps at -90 deg incidence. This method provides for the solution of the unsteady incompressible Navier-Stokes equations by means of an implicit technique. The solution is calculated on a body-fitted computational mesh using a staggered-grid method. The vorticity is defined at the node points, and the velocity components are defined at the mesh-cell sides. The staggered-grid orientation provides for accurate representation of vorticity at the node points and the continuity equation at the mesh-cell centers. The method provides for the noniterative solution of the flowfield and satisfies the continuity equation to machine zero at each time step. The method is evaluated in terms of its stability to predict two-dimensional flow about an airfoil at -90-deg incidence for varying Reynolds number and laminar/turbulent models. The variations of the average loading and surface pressure distribution due to flap deflection, Reynolds number, and laminar or turbulent flow are presented and compared with experimental results. The comparisom indicate that the calculated drag and drag reduction caused by flap deflection and the calculated average surface pressure are in excellent agreement with the measured results at a similar Reynolds number.

Description: The multigrid method has been applied to an existing three-dimensional compressible Euler solver to accelerate the convergence of the implicit symmetric relaxation scheme. This lower-upper symmetric Gauss-Seidel implicit scheme is shown to be an effective multigrid driver in three dimensions. A grid refinement study is performed including the effects of large cell aspect ratio meshes. Performance figures of the present multigrid code on Cray computers including the new C90 are presented. A reduction of three orders of magnitude in the residual for a three-dimensional transonic inviscid flow using 920 k grid points is obtained in less than 4 min on a Cray C90.

Description: The present paper explores the use of large-eddy simulations as a tool for predicting noise from first principles. A high-order numerical scheme is used to perform large-eddy simulations of a supersonic jet flow with emphasis on capturing the time-dependent flow structure representating the sound source. The wavelike nature of this structure under random inflow disturbances is demonstrated. This wavelike structure is then enhanced by taking the inflow disturbances to be purely harmonic. Application of Lighthill's theory to calculate the far-field noise, with the sound source obtained from the calculated time-dependent near field, is demonstrated. Alternative approaches to coupling the near-field sound source to the far-field sound are discussed.

Description: The steady state solution of the system of equations consisting of the full Navier-Stokes equations and two turbulence equations has been obtained using a multigrid strategy of unstructured meshes. The flow equations and turbulence equations are solved in a loosely coupled manner. The flow equations are advanced in time using a multistage Runge-Kutta time-stepping scheme with a stability-bound local time step, while turbulence equations are advanced in a point-implicit scheme with a time step which guarantees stability and positivity. Low-Reynolds-number modifications to the original two-equation model are incorporated in a manner which results in well-behaved equations for arbitrarily small wall distances. A variety of aerodynamic flows are solved, initializing all quantities with uniform freestream values. Rapid and uniform convergence rates for the flow and turbulence equations are observed.

Description: Rotor noise prediction codes predict the thickness and loading noise produced by a helicopter rotor, given the blade motion, rotor operating conditions, and fluctuating force distribution over the blade surface. However, the criticality of these various inputs, and their respective effects on the predicted acoustic field, have never been fully addressed. This paper examines the importance of these inputs, and the sensitivity of the acoustic predicitions to a variation of each parameter. The effects of collective and cyclic pitch, as well as coning and cyclic flapping, are presented. Blade loading inputs are examined to determine the necessary spatial and temporal resolution, as well as the importance of the chordwise distribution. The acoustic predictions show regions in the acoustic field where significant errors occur when simplified blade motions or blade loadings are used. An assessment of the variation in the predicted acoustic field is balanced by a consideration of Central Processing Unit (CPU) time necessary for the various approximations.

Description: The U.S. National Aeronautics and Space Administration (NASA) Balloon Program has been highly successful since recovering from the catastrophic balloon failure problems of the early to mid 1980s. Balloons have continued to perform at unprecedented success rates. The comprehensive research and development (R&D) effort has continued with advances being made across the spectrum of balloon related disciplines. The long duration balloon project will be transitioning from a development effort to an operational capability this year. Recently, emphasis has been placed on the development and implementation of new support systems and facilities. A new permanent launch facility at Fort Sumner, New Mexico has been established. New ground station support equipment is being implemented, and a new heavy load launch vehicle is scheduled to be implemented in 1992. The progress, status and future plans for these and other aspects of the NASA program will be presented.

Description: The catastrophic balloon failure during the first half of the 1980's identified the need for a comprehensive and continuing balloon research and development (R&D) commitment by NASA. Technical understanding was lacking in many of the disciplines and processes associated with scientific ballooning. A comprehensive balloon R&D plan was developed in 1986 and implemented in 1987. The objectives were to develop the understanding of balloon system performance, limitations, and failure mechanisms. The program consisted of five major technical areas: structures, performance and analysis, materials, chemistry and processing, and quality control. Research activitites have been conducted at NASA/Goddard Space Flight Center (GSFC)-Wallops Flight Facility (WFF), other NASA centers and government facilities, universities, and the balloon manufacturers. Several new and increased capabilities and resources have resulted from this activity. The findings, capabilities, and plan of the balloon R&D program are presented.

Description: Caps have been used to structurally reinforce scientific research balloons since the late 1950's. The scientific research balloons used by the National Aeronautics and Space Administration (NASA) use internal caps. A NASA cap placement specification does not exist since no empirical information exisits concerning cap placement. To develop a cap placement specification, NASA has completed two in-hangar inflation tests comparing the structural contributions of internal caps and external caps. The tests used small scale test balloons designed to develop the highest possible stresses within the constraints of the hangar and balloon materials. An externally capped test balloon and an internally capped test balloon were designed, built, inflated and simulated to determine the structural contributions and benefits of each. The results of the tests and simulations are presented.

Description: The analysis of balloon envelopes by the finite element (FE) method is plagued by convergence problems. A pratical FE analysis approach is based on the fact that in thin shells with non-zero Gaussian curvature the membrane solution component is essentially decoupled from the bending solution component. A proxy-problem is solved by using a small artificial bending stiffness that assures convergence without significantly affecting the membrane solution component. This approach has been previously validated on slightly overpressurized balloon envelopes. Extensions of this approach to more difficult problems in the structural analysis of balloon envelopes are presented. The convergence forcing modelling measures are discussed. Implications of the findings of the analysis results to future balloon designs are also discussed.

Description: Numerical results obtained with direct simulation Monte Carlo and Navier-Stokes methods are presented for a Mach-20 nitrogen flow about a 70-deg blunted cone. The flow conditions simuulated are those that can be obtained in existing low-density hypersonic wind tunnels. Three sets of flow conditions are considered with freestream Knudsen numbers ranging from 0.03 to 0.001. The focus is on the wake structure: how the wake structure changes as a function of rarefaction, what the afterbody levels of heating are, and to what limits the continuum models are realistic as rarefaction in the wake is progressively increased. Calculations are made with and without an afterbody sting. Results for the after body sting are emphasizes in anticipation of an experimental study for the current flow conditions and model configuration. The Navier-Stokes calculations were made with and without slip boundary conditions. Comparisons of the results obtained with the two simulation methodologies are made for both flowfield structure and surface quantities.

Description: The purpose of this Note is to present results from an analytic/experimental study that investigated the potential for passively changing blade twist through the use of extension-twist coupling. A set of composite model rotor blades was manufactured from existing blade molds for a low-twist metal helicopter rotor blade, with a view toward establishing a preliminary proof concept for extension-twist-coupled rotor blades. Data were obtained in hover for both a ballasted and unballasted blade configuration in sea-level atmospheric conditions. Test data were compared with results obtained from a geometrically nonlinear analysis of a detailed finite element model of the rotor blade developed in MSC/NASTRAN.

Description: The paper considers the compressible Rayleigh equation as a model for the Mach wave emission mechanism associated with high-temperature supersonic jets. Solutions to the compressible Rayleigh equation reveal the existence of several families of supersonically convecting instability waves. These waves directly radiate noise to the jet far field. The predicted noise characteristics are compared to previously acquired experimental data for an axisymmetric Mach 2 fully pressure balanced jet operating over a range of jet total temperatures from ambient to 1370 K. The results of this comparison show that the first-order supersonic instability wave and the Kelvin-Hemlhlotz first-, second-, and third-order modes have directional radiation characteristics that are in agreement with observed data. The assumption of equal initial amplitudes for all of the waves leads to the conclusion that the flapping mode of instability dominates the noise radiatio process of supersonic jets. At a jet temperature of 1370 K, supersonic instability waves are predicted to dominate the noise radiated at high frequency at narrow angles to the jet axis.

Description: The objective of the present work is to study the mixing characteristics of a linear array of supersonic rectangular jets under conditions of screech synchronization. The screech synchronization at a fully expanded jet Mach number of 1.61 is achieved by a precise adjustment of the internozzle spacing. To our knowledge, such an experiment on the resonant mixing of screech synchronized multiple rectangular jets has not been reported before. The results are compared with the case where the screech was suppressed in the multijet configuration.

Description: The objective of the present investigation is to assess the effect of the spatial order of accuracy used for the evaluation of the inviscid fluxes on the resolution of higher order quantitites, such as velocity gradients. The viscous terms are computed as second-order accurate with central difference formulas, even though for the explicit part of the algorithm higher order approximations may be used. A viscous/inviscid method is used, and the outer part of the flowfield is computed with the inviscid flow equations. The viscous boundary-layer type flow region close to the body surface is computed with an algebraic eddy viscosity model. Results obtained with the conservative and nonconservative formulations and the viscous/inviscid approach are compared with available experimental data. The effect of grid refinement on the accuracy of the solution is also presented.

Description: The transformation validity question utilizing resulting data from direct numerical simulations (DNS) of supersonic, isothermal cold wall channel flow was investigated. The DNS results stood for a wide scope of parameter and were suitable for the purpose of examining the generality of Van Driest transformation. The Van Driest law of the wall can be obtained from the inner-layer similarity arguments. It was demonstrated that the Van Driest transformation cannot be incorporated to collapse the sublayer and log-layer velocity profiles simultaneously. Velocity and temperature predictions according to the preceding composite mixing-length model were presented. Despite satisfactory congruity with the DNS data, the model must be perceived as an engineering guide and not as a rigorous analysis.

Description: A Monte Carlo solution technique has been formulated to predict the radiative heat transfer in three-dimensional, inhomogeneous participating media which exhibit spectrally dependent emission and absorption and anisotropic scattering. Details of the technique and selected numerical sensitivities are discussed. The technique was applied to a problem involving a medium composed of a gas mixture of carbon dioxide and nitrogen and suspended carbon particles. A homogeneous medium was modeled to examine the effect of total pressure and carbon-particle concentration on radiative heat transfer. Variation in total pressure, over the range studied, had minimal effect on the amount of heat radiated to the enclosure walls and on the radiative-flux distribution within the medium. Increases in the carbon particle concentration produced significantly higher heat fluxes at the boundaries and altered the radiative flux distribution. The technique was then applied to an inhomogeneous medium to examine effects of specific temperature and carbon particle concentration distributions on radiative heat transfer. For the inhomogeneous conditions examined, the largest radiative flux divergence occurs near the center of the medium and the regions near some enclosure walls act as energy sinks.

Description: Numerical results obtained with direct simulation Monte Carlo and Navier-Stokes methods are presented for a Mach-20 nitrogen flow about a 70-deg blunted cone. The flow conditions simulated are those that can be obtained in existing low-density hypersonic wind tunnels. Three sets of flow conditions are considered with freestream Knudsen numbers ranging from 0.03 to 0.001. The focus is on the wake structure: how the wake structure changes as a function of rare faction, what the afterbody levels of heating are, and to what limits the continuum models are realistic as rarefunction in the wake is progressively increased. Calculations are made with and without an afterbody sting. Results for the afterbody sting are emphasized in anticipation of an experimental study for the current flow conditions and model configuration. The Navier-Stokes calculations were made with and without slip boundary conditions. Comparisons of the results obtained with the two simulation methodologies are made for both flowfield structure and surface quantities.

Description: The Simplified Shuttle Payload Thermal Analyzer program (SSPTA) was developed to aid in the evaluation of thermal design concepts of instruments to be flown in the Space Shuttle cargo bay. SSPTA consists of a collection of programs that are currently used in the thermal analysis of spacecraft and have been modified for quick, preliminary analysis of payloads. SSPTA includes a reduced math model of the Shuttle cargo bay to simplify use of the program for payload analysis. One of the prime objectives in developing SSPTA was to create a program which was easy to use. With SSPTA, the user required input is simple and the user is free from many of the concerns of computer usage such as disk space handling, tape usage, and complicated program control. Although SSPTA was designed primarily to analyze Shuttle payloads, it can easily be used to perform thermal analysis in other situations. SSPTA is comprised of a system of data files called 'bins', a master program, and a set of thermal subprograms. The bin system is a collection of disk files which contain data required by or computed by the thermal subprograms. SSPTA currently has the capability of handling 50 bins. The master program serves primarily as a manager for the bin system and its interaction with the thermal subprograms. Input to the master program consists of simple user commands which direct the data manipulation procedures, prepare the data for these procedures, and call the appropriate thermal subprograms. The subprograms of SSPTA are all based on programs which have been used extensively in the analysis of orbiting spacecraft and space hardware. Subprogram CONSHAD uses the user supplied geometric radiation model to compute black body view factors, shadow factors, and a description of the surface model. The subprogram WORKSHEET uses the surface model description, optical property data, and node assignment data to prepare input for SCRIPTF. Subprogram SCRIPTF computes the inverses of the infrared (IR) and ultraviolet (UV) radiation transfer equations; it also computes the radiation coupling between nodes in the thermal model. Subprogram ORBITAL uses the shadow tables to compute incident flux intensities on each surface in the geometric model. Subprogram ABSORB uses these flux intensities combined with the IR and UV inverses to compute the IR and UV fluxes absorbed by each surface. The radiation couplings from SCRIPTF and the absorbed fluxes from ABSORB are used by subprogram TTA to compute the temperature and power balance for each node in the thermal model. Output consists of tabulated data from each of the subprograms executed during a particular analysis. Due to the modular form of SSPTA, analyses may be run in whole or in part, and new subprograms may be added by the user. SSPTA is written in FORTRAN for use on a DEC VAX-11/780. SSPTA was originally developed in 1977 for use on IBM 370 series computers. This version is an update which was ported to the VAX in 1980.

Description: ARC2D is a computational fluid dynamics program developed at the NASA Ames Research Center specifically for airfoil computations. The program uses implicit finite-difference techniques to solve two-dimensional Euler equations and thin layer Navier-Stokes equations. It is based on the Beam and Warming implicit approximate factorization algorithm in generalized coordinates. The methods are either time accurate or accelerated non-time accurate steady state schemes. The evolution of the solution through time is physically realistic; good solution accuracy is dependent on mesh spacing and boundary conditions. The mathematical development of ARC2D begins with the strong conservation law form of the two-dimensional Navier-Stokes equations in Cartesian coordinates, which admits shock capturing. The Navier-Stokes equations can be transformed from Cartesian coordinates to generalized curvilinear coordinates in a manner that permits one computational code to serve a wide variety of physical geometries and grid systems. ARC2D includes an algebraic mixing length model to approximate the effect of turbulence. In cases of high Reynolds number viscous flows, thin layer approximation can be applied. ARC2D allows for a variety of solutions to stability boundaries, such as those encountered in flows with shocks. The user has considerable flexibility in assigning geometry and developing grid patterns, as well as in assigning boundary conditions. However, the ARC2D model is most appropriate for attached and mildly separated boundary layers; no attempt is made to model wake regions and widely separated flows. The techniques have been successfully used for a variety of inviscid and viscous flowfield calculations. The Cray version of ARC2D is written in FORTRAN 77 for use on Cray series computers and requires approximately 5Mb memory. The program is fully vectorized. The tape includes variations for the COS and UNICOS operating systems. Also included is a sample routine for CONVEX computers to emulate Cray system time calls, which should be easy to modify for other machines as well. The standard distribution media for this version is a 9-track 1600 BPI ASCII Card Image format magnetic tape. The Cray version was developed in 1987. The IBM ES/3090 version is an IBM port of the Cray version. It is written in IBM VS FORTRAN and has the capability of executing in both vector and parallel modes on the MVS/XA operating system and in vector mode on the VM/XA operating system. Various options of the IBM VS FORTRAN compiler provide new features for the ES/3090 version, including 64-bit arithmetic and up to 2 GB of virtual addressability. The IBM ES/3090 version is available only as a 9-track, 1600 BPI IBM IEBCOPY format magnetic tape. The IBM ES/3090 version was developed in 1989. The DEC RISC ULTRIX version is a DEC port of the Cray version. It is written in FORTRAN 77 for RISC-based Digital Equipment platforms. The memory requirement is approximately 7Mb of main memory. It is available in UNIX tar format on TK50 tape cartridge. The port to DEC RISC ULTRIX was done in 1990. COS and UNICOS are trademarks and Cray is a registered trademark of Cray Research, Inc. IBM, ES/3090, VS FORTRAN, MVS/XA, and VM/XA are registered trademarks of International Business Machines. DEC and ULTRIX are registered trademarks of Digital Equipment Corporation.

Description: Panel method computer programs are software tools of moderate cost used for solving a wide range of engineering problems. The panel code PMARC_12 (Panel Method Ames Research Center, version 12) can compute the potential flow field around complex three-dimensional bodies such as complete aircraft models. PMARC_12 is a well-documented, highly structured code with an open architecture that facilitates modifications and the addition of new features. Adjustable arrays are used throughout the code, with dimensioning controlled by a set of parameter statements contained in an include file; thus, the size of the code (i.e. the number of panels that it can handle) can be changed very quickly. This allows the user to tailor PMARC_12 to specific problems and computer hardware constraints. In addition, PMARC_12 can be configured (through one of the parameter statements in the include file) so that the code's iterative matrix solver is run entirely in RAM, rather than reading a large matrix from disk at each iteration. This significantly increases the execution speed of the code, but it requires a large amount of RAM memory. PMARC_12 contains several advanced features, including internal flow modeling, a time-stepping wake model for simulating either steady or unsteady (including oscillatory) motions, a Trefftz plane induced drag computation, off-body and on-body streamline computations, and computation of boundary layer parameters using a two-dimensional integral boundary layer method along surface streamlines. In a panel method, the surface of the body over which the flow field is to be computed is represented by a set of panels. Singularities are distributed on the panels to perturb the flow field around the body surfaces. PMARC_12 uses constant strength source and doublet distributions over each panel, thus making it a low order panel method. Higher order panel methods allow the singularity strength to vary linearly or quadratically across each panel. Experience has shown that low order panel methods can provide nearly the same accuracy as higher order methods over a wide range of cases with significantly reduced computation times; hence, the low order formulation was adopted for PMARC_12. The flow problem is solved by modeling the body as a closed surface dividing space into two regions: the region external to the surface in which an unknown velocity potential exists representing the flow field of interest, and the region internal to the surface in which a known velocity potential (representing a fictitious flow) is prescribed as a boundary condition. Both velocity potentials are required to satisfy Laplace's equation. A surface integral equation for the unknown potential external to the surface can be written by applying Green's Theorem to the external region. Using the internal potential and zero flow through the surface as boundary conditions, the unknown potential external to the surface can be solved for. When the internal flow option, which allows the analysis of closed ducts, wind tunnels, and similar internal flow problems, is selected, the geometry is modeled such that the flow field of interest is inside the geometry and the fictitious flow is outside the geometry. Items such as wings, struts, or aircraft models can be included in the internal flow problem. The time-stepping wake model gives PMARC_12 the ability to model both steady and unsteady flow problems. The wake is convected downstream from the wake-separation line by the local velocity field. With each time step, a new row of wake panels is added to the wake at the wake-separation line. Time stepping can start from time t=0 (no initial wake) or from time t=t0 (an initial wake is specified). A wide range of motions can be prescribed, including constant rates of translation, constant rate of rotation about an arbitrary axis, oscillatory translation, and oscillatory rotation about any of the three coordinate axes. Investigators interested in a visual representation of the phenomenon they are studying with PMARC_12 may want to consider obtaining the program GVS (ARC-13361), the General Visualization System. GVS is a Silicon Graphics IRIS program which was created for the purpose of supporting the scientific visualization needs of PMARC_12. GVS is available separately from COSMIC. PMARC_12 is written in standard FORTRAN 77, with the exception of the NAMELIST extension used for input. This makes the code fairly machine independent. A compiler which supports the NAMELIST extension is required. The amount of free disk space and RAM memory required for PMARC_12 will vary depending on how the code is dimensioned using the parameter statements in the include file. The recommended minimum requirements are 20Mb of free disk space and 4Mb of RAM. PMARC_12 has been successfully implemented on a Macintosh II running System 6.0.7 or 7.0 (using MPW/Language Systems Fortran 3.0), a Sun SLC running SunOS 4.1.1, an HP 720 running HP-UX 8.07, an SGI IRIS running IRIX 4.0 (it will not run under IRIX 3.x.x without modifications), an IBM RS/6000 running AIX, a DECstation 3100 running ULTRIX, and a CRAY-YMP running UNICOS 6.0 or later. Due to its memory requirements, this program does not readily lend itself to implementation on MS-DOS based machines. The standard distribution medium for PMARC_12 is a set of three 3.5 inch 800K Macintosh format diskettes and one 3.5 inch 1.44Mb Macintosh format diskette which contains an electronic copy of the documentation in MS Word 5.0 format for the Macintosh. Alternate distribution media and formats are available upon request, but these will not include the electronic version of the document. No executables are included on the distribution media. This program is an update to PMARC version 11, which was released in 1989. PMARC_12 was released in 1993. It is available only for use by United States citizens.

Description: Remote Interactive Particle-tracing (RIP) is a distributed-graphics program which computes particle traces for computational fluid dynamics (CFD) solution data sets. A particle trace is a line which shows the path a massless particle in a fluid will take; it is a visual image of where the fluid is going. The program is able to compute and display particle traces at a speed of about one trace per second because it runs on two machines concurrently. The data used by the program is contained in two files. The solution file contains data on density, momentum and energy quantities of a flow field at discrete points in three-dimensional space, while the grid file contains the physical coordinates of each of the discrete points. RIP requires two computers. A local graphics workstation interfaces with the user for program control and graphics manipulation, and a remote machine interfaces with the solution data set and performs time-intensive computations. The program utilizes two machines in a distributed mode for two reasons. First, the data to be used by the program is usually generated on the supercomputer. RIP avoids having to convert and transfer the data, eliminating any memory limitations of the local machine. Second, as computing the particle traces can be computationally expensive, RIP utilizes the power of the supercomputer for this task. Although the remote site code was developed on a CRAY, it is possible to port this to any supercomputer class machine with a UNIX-like operating system. Integration of a velocity field from a starting physical location produces the particle trace. The remote machine computes the particle traces using the particle-tracing subroutines from PLOT3D/AMES, a CFD post-processing graphics program available from COSMIC (ARC-12779). These routines use a second-order predictor-corrector method to integrate the velocity field. Then the remote program sends graphics tokens to the local machine via a remote-graphics library. The local machine interprets the graphics tokens and draws the particle traces. The program is menu driven. RIP is implemented on the silicon graphics IRIS 3000 (local workstation) with an IRIX operating system and on the CRAY2 (remote station) with a UNICOS 1.0 or 2.0 operating system. The IRIS 4D can be used in place of the IRIS 3000. The program is written in C (67%) and FORTRAN 77 (43%) and has an IRIS memory requirement of 4 MB. The remote and local stations must use the same user ID. PLOT3D/AMES unformatted data sets are required for the remote machine. The program was developed in 1988.

Description: This software package includes two programs, the KPD12 and the KPD12P. Both programs utilizes the vortex-blob method to simulate flow around solid bodies, in an unbounded domain using the KPD12, with periodicity in one direction using the KPD12P. The main advantage of the vortex-blob method is the ability to handle situations involving arbitrary shapes including multiple bodies. The user just supplies points on the solid boundaries; there is no grid. The KPD12 program has worked successfully on bluff bodies, stalled wings, and multiple-element airfoils. The KPD12P program has been used successfully on high-solidity separated cascades and on cases of rotating stall in cascades of thin airfoils. However, they do not capture subtle viscous effects such as incipient separation and friction drag. The KPD12 and the KPD12P programs apply the vortex-blob method to time-dependent, high-Reynolds-number flows around solid bodies. Both programs solve the two-dimensional incompressible Navier-Stokes equations, neglecting the viscous effects away from the walls. By creating new vortices along the wall at every time step, they treat the no-penetration and no-slip boundary conditions while using an influence matrix. The code automatically controls the number of vortices. Furthermore, the code has the option of treating the boundary layers by simple integral methods to determine the separation points. The KPD12 outputs forces, moments, and pressure distributions on the bodies. The KPD12P also outputs the turning angle and loss of total pressure. The source code is in Cray FORTRAN and contains a few calls to Cray vector functions which are vectorized with the Cray compiler. However, substitutes for these vector functions are provided. The code is set up to plot the bodies, vortex positions, and streamlines using the DISSPLA graphics software. The software requires a mainframe computer with at least 589k of memory available running under COS 1.16. KPD12 was developed in 1988.

Description: This program determines the supersonic flowfield surrounding three-dimensional wing-body configurations of a delta wing. It was designed to provide the numerical computation of three dimensional inviscid, flowfields of either perfect or real gases about supersonic or hypersonic airplanes. The governing equations in conservation law form are solved by a finite difference method using a second order noncentered algorithm between the body and the outermost shock wave, which is treated as a sharp discontinuity. Secondary shocks which form between these boundaries are captured automatically. The flowfield between the body and outermost shock is treated in a shock capturing fashion and therefore allows for the correct formation of secondary internal shocks . The program operates in batch mode, is in CDC update format, has been implemented on the CDC 7600, and requires more than 140K (octal) word locations.

Description: The Steady State Thermal Analysis Program (STEADY) provides the thermal designer with a quick and convenient method for calculating heat loads and temperatures. STEADY can be used on small nodal networks for conceptual or preliminary thermal design and analysis. STEADY will accept up to 20 nodes of fixed or variable temperature, with constant or temperature-dependent thermal conductivities, and any set of consistent units. In a steady state thermal network, the heat balance on each variable temperature node must sum to zero. The general heat transfer equations are solved with a Newton-Raphson technique and refined by a fourth order quartic solution. Input data includes the number of nodes, number of boundary nodes, the fixed temperatures at all boundary nodes, initial temperature guesses for variable nodes, impressed heat loads, conduction and radiation coefficients, and control parameters such as convergence criteria, maximum iterations, and damping factors. The output is stored in a print file and tabulates final temperatures and heat flows for all nodes. STEADY is menu driven and allows the user to save files for future modification. STEADY is written in FORTRAN 77 (Ryan McFarland's RMFORTRAN) for interactive execution and has been implemented on the IBM PC computer series under DOS with a central memory requirement of approximately 92K of 8 bit bytes using a math coprocessor, and 103K bytes without the coprocessor. This program was developed in 1987.

Description: The Composite Plate Buckling Analysis Program (COMPPAP) was written to help engineers determine buckling loads of orthotropic (or isotropic) irregularly shaped plates without requiring hand calculations from design curves or extensive finite element modeling. COMPPAP is a one element finite element program that utilizes high-order displacement functions. The high order of the displacement functions enables the user to produce results more accurate than traditional h-finite elements. This program uses these high-order displacement functions to perform a plane stress analysis of a general plate followed by a buckling calculation based on the stresses found in the plane stress solution. The current version assumes a flat plate (constant thickness) subject to a constant edge load (normal or shear) on one or more edges. COMPPAP uses the power method to find the eigenvalues of the buckling problem. The power method provides an efficient solution when only one eigenvalue is desired. Once the eigenvalue is found, the eigenvector, which corresponds to the plate buckling mode shape, results as a by-product. A positive feature of the power method is that the dominant eigenvalue is the first found, which is this case is the plate buckling load. The reported eigenvalue expresses a load factor to induce plate buckling. COMPPAP is written in ANSI FORTRAN 77. Two machine versions are available from COSMIC: a PC version (MSC-22428), which is for IBM PC 386 series and higher computers and compatibles running MS-DOS; and a UNIX version (MSC-22286). The distribution medium for both machine versions includes source code for both single and double precision versions of COMPPAP. The PC version includes source code which has been optimized for implementation within DOS memory constraints as well as sample executables for both the single and double precision versions of COMPPAP. The double precision versions of COMPPAP have been successfully implemented on an IBM PC 386 compatible running MS-DOS, a Sun4 series computer running SunOS, an HP-9000 series computer running HP-UX, and a CRAY X-MP series computer running UNICOS. COMPPAP requires 1Mb of RAM and the BLAS and LINPACK math libraries, which are included on the distribution medium. The COMPPAP documentation provides instructions for using the commercial post-processing package PATRAN for graphical interpretation of COMPPAP output. The UNIX version includes two electronic versions of the documentation: one in LaTex format and one in PostScript format. The standard distribution medium for the PC version (MSC-22428) is a 5.25 inch 1.2Mb MS-DOS format diskette. The standard distribution medium for the UNIX version (MSC-22286) is a .25 inch streaming magnetic tape cartridge (Sun QIC-24) in UNIX tar format. For the UNIX version, alternate distribution media and formats are available upon request. COMPPAP was developed in 1992.

Description: The Composite Plate Buckling Analysis Program (COMPPAP) was written to help engineers determine buckling loads of orthotropic (or isotropic) irregularly shaped plates without requiring hand calculations from design curves or extensive finite element modeling. COMPPAP is a one element finite element program that utilizes high-order displacement functions. The high order of the displacement functions enables the user to produce results more accurate than traditional h-finite elements. This program uses these high-order displacement functions to perform a plane stress analysis of a general plate followed by a buckling calculation based on the stresses found in the plane stress solution. The current version assumes a flat plate (constant thickness) subject to a constant edge load (normal or shear) on one or more edges. COMPPAP uses the power method to find the eigenvalues of the buckling problem. The power method provides an efficient solution when only one eigenvalue is desired. Once the eigenvalue is found, the eigenvector, which corresponds to the plate buckling mode shape, results as a by-product. A positive feature of the power method is that the dominant eigenvalue is the first found, which is this case is the plate buckling load. The reported eigenvalue expresses a load factor to induce plate buckling. COMPPAP is written in ANSI FORTRAN 77. Two machine versions are available from COSMIC: a PC version (MSC-22428), which is for IBM PC 386 series and higher computers and compatibles running MS-DOS; and a UNIX version (MSC-22286). The distribution medium for both machine versions includes source code for both single and double precision versions of COMPPAP. The PC version includes source code which has been optimized for implementation within DOS memory constraints as well as sample executables for both the single and double precision versions of COMPPAP. The double precision versions of COMPPAP have been successfully implemented on an IBM PC 386 compatible running MS-DOS, a Sun4 series computer running SunOS, an HP-9000 series computer running HP-UX, and a CRAY X-MP series computer running UNICOS. COMPPAP requires 1Mb of RAM and the BLAS and LINPACK math libraries, which are included on the distribution medium. The COMPPAP documentation provides instructions for using the commercial post-processing package PATRAN for graphical interpretation of COMPPAP output. The UNIX version includes two electronic versions of the documentation: one in LaTex format and one in PostScript format. The standard distribution medium for the PC version (MSC-22428) is a 5.25 inch 1.2Mb MS-DOS format diskette. The standard distribution medium for the UNIX version (MSC-22286) is a .25 inch streaming magnetic tape cartridge (Sun QIC-24) in UNIX tar format. For the UNIX version, alternate distribution media and formats are available upon request. COMPPAP was developed in 1992.

Description: Structural flaws and cracks may grow under fatigue inducing loads and, upon reaching a critical size, cause structural failure to occur. The growth of these flaws and cracks may occur at load levels well below the ultimate load bearing capability of the structure. The Fatigue Crack Growth Computer Program, NASA/FLAGRO, was developed as an aid in predicting the growth of pre-existing flaws and cracks in structural components of space systems. The earlier version of the program, FLAGRO4, was the primary analysis tool used by Rockwell International and the Shuttle subcontractors for fracture control analysis on the Space Shuttle. NASA/FLAGRO is an enhanced version of the program and incorporates state-of-the-art improvements in both fracture mechanics and computer technology. NASA/FLAGRO provides the fracture mechanics analyst with a computerized method of evaluating the "safe crack growth life" capabilities of structural components. NASA/FLAGRO could also be used to evaluate the damage tolerance aspects of a given structural design. The propagation of an existing crack is governed by the stress field in the vicinity of the crack tip. The stress intensity factor is defined in terms of the relationship between the stress field magnitude and the crack size. The propagation of the crack becomes catastrophic when the local stress intensity factor reaches the fracture toughness of the material. NASA/FLAGRO predicts crack growth using a two-dimensional model which predicts growth independently in two directions based on the calculation of stress intensity factors. The analyst can choose to use either a crack growth rate equation or a nonlinear interpolation routine based on tabular data. The growth rate equation is a modified Forman equation which can be converted to a Paris or Walker equation by substituting different values into the exponent. This equation provides accuracy and versatility and can be fit to data using standard least squares methods. Stress-intensity factor numerical values can be computed for making comparisons or checks of solutions. NASA/FLAGRO can check for failure of a part-through crack in the mode of a through crack when net ligament yielding occurs. NASA/FLAGRO has a number of special subroutines and files which provide enhanced capabilities and easy entry of data. These include crack case solutions, cyclic load spectrums, nondestructive examination initial flaw sizes, table interpolation, and material properties. The materials properties files are divided into two types, a user defined file and a fixed file. Data is entered and stored in the user defined file during program execution, while the fixed file contains already coded-in property value data for many different materials. Prompted input from CRT terminals consists of initial crack definition (which can be defined automatically), rate solution type, flaw type and geometry, material properties (if they are not in the built-in tables of material data), load spectrum data (if not included in the loads spectrum file), and design limit stress levels. NASA/FLAGRO output includes an echo of the input with any error or warning messages, the final crack size, whether or not critical crack size has been reached for the specified stress level, and a life history profile of the crack propagation. NASA/FLAGRO is modularly designed to facilitate revisions and operation on minicomputers. The program was implemented on a DEC VAX 11/780 with the VMS operating system. NASA/FLAGRO is written in FORTRAN77 and has a memory requirement of 1.4 MB. The program was developed in 1986.

Description: The Thermal Radiation Analyzer System, TRASYS, is a computer software system with generalized capability to solve the radiation related aspects of thermal analysis problems. TRASYS computes the total thermal radiation environment for a spacecraft in orbit. The software calculates internode radiation interchange data as well as incident and absorbed heat rate data originating from environmental radiant heat sources. TRASYS provides data of both types in a format directly usable by such thermal analyzer programs as SINDA/FLUINT (available from COSMIC, program number MSC-21528). One primary feature of TRASYS is that it allows users to write their own driver programs to organize and direct the preprocessor and processor library routines in solving specific thermal radiation problems. The preprocessor first reads and converts the user's geometry input data into the form used by the processor library routines. Then, the preprocessor accepts the user's driving logic, written in the TRASYS modified FORTRAN language. In many cases, the user has a choice of routines to solve a given problem. Users may also provide their own routines where desirable. In particular, the user may write output routines to provide for an interface between TRASYS and any thermal analyzer program using the R-C network concept. Input to the TRASYS program consists of Options and Edit data, Model data, and Logic Flow and Operations data. Options and Edit data provide for basic program control and user edit capability. The Model data describe the problem in terms of geometry and other properties. This information includes surface geometry data, documentation data, nodal data, block coordinate system data, form factor data, and flux data. Logic Flow and Operations data house the user's driver logic, including the sequence of subroutine calls and the subroutine library. Output from TRASYS consists of two basic types of data: internode radiation interchange data, and incident and absorbed heat rate data. The flexible structure of TRASYS allows considerable freedom in the definition and choice of solution method for a thermal radiation problem. The program's flexible structure has also allowed TRASYS to retain the same basic input structure as the authors update it in order to keep up with changing requirements. Among its other important features are the following: 1) up to 3200 node problem size capability with shadowing by intervening opaque or semi-transparent surfaces; 2) choice of diffuse, specular, or diffuse/specular radiant interchange solutions; 3) a restart capability that minimizes recomputing; 4) macroinstructions that automatically provide the executive logic for orbit generation that optimizes the use of previously completed computations; 5) a time variable geometry package that provides automatic pointing of the various parts of an articulated spacecraft and an automatic look-back feature that eliminates redundant form factor calculations; 6) capability to specify submodel names to identify sets of surfaces or components as an entity; and 7) subroutines to perform functions which save and recall the internodal and/or space form factors in subsequent steps for nodes with fixed geometry during a variable geometry run. There are two machine versions of TRASYS v27: a DEC VAX version and a Cray UNICOS version. Both versions require installation of the NASADIG library (MSC-21801 for DEC VAX or COS-10049 for CRAY), which is available from COSMIC either separately or bundled with TRASYS. The NASADIG (NASA Device Independent Graphics Library) plot package provides a pictorial representation of input geometry, orbital/orientation parameters, and heating rate output as a function of time. NASADIG supports Tektronix terminals. The CRAY version of TRASYS v27 is written in FORTRAN 77 for batch or interactive execution and has been implemented on CRAY X-MP and CRAY Y-MP series computers running UNICOS. The standard distribution medium for MSC-21959 (CRAY version without NASADIG) is a 1600 BPI 9-track magnetic tape in UNIX tar format. The standard distribution medium for COS-10040 (CRAY version with NASADIG) is a set of two 6250 BPI 9-track magnetic tapes in UNIX tar format. Alternate distribution media and formats are available upon request. The DEC VAX version of TRASYS v27 is written in FORTRAN 77 for batch execution (only the plotting driver program is interactive) and has been implemented on a DEC VAX 8650 computer under VMS. Since the source codes for MSC-21030 and COS-10026 are in VAX/VMS text library files and DEC Command Language files, COSMIC will only provide these programs in the following formats: MSC-21030, TRASYS (DEC VAX version without NASADIG) is available on a 1600 BPI 9-track magnetic tape in VAX BACKUP format (standard distribution medium) or in VAX BACKUP format on a TK50 tape cartridge; COS-10026, TRASYS (DEC VAX version with NASADIG), is available in VAX BACKUP format on a set of three 6250 BPI 9-track magnetic tapes (standard distribution medium) or a set of three TK50 tape cartridges in VAX BACKUP format. TRASYS was last updated in 1993.

Description: TDIGG is a fast and versatile program for generating two-dimensional computational grids for use with finite-difference flow-solvers. Both algebraic and elliptic grid generation systems are included. The method for grid generation by algebraic transformation is based on an interpolation algorithm and the elliptic grid generation is established by solving the partial differential equation (PDE). Non-uniform grid distributions are carried out using a hyperbolic tangent stretching function. For algebraic grid systems, interpolations in one direction (univariate) and two directions (bivariate) are considered. These interpolations are associated with linear or cubic Lagrangian/Hermite/Bezier polynomial functions. The algebraic grids can subsequently be smoothed using an elliptic solver. For elliptic grid systems, the PDE can be in the form of Laplace (zero forcing function) or Poisson. The forcing functions in the Poisson equation come from the boundary or the entire domain of the initial algebraic grids. A graphics interface procedure using the Silicon Graphics (GL) Library is included to allow users to visualize the grid variations at each iteration. This will allow users to interactively modify the grid to match their applications. TDIGG is written in FORTRAN 77 for Silicon Graphics IRIS series computers running IRIX. This package requires either MIT's X Window System, Version 11 Revision 4 or SGI (Motif) Window System. A sample executable is provided on the distribution medium. It requires 148K of RAM for execution. The standard distribution medium is a .25 inch streaming magnetic IRIX tape cartridge in UNIX tar format. This program was developed in 1992.

Description: The Systems Improved Numerical Fluids Analysis Code, SINFAC, consists of additional routines added to the April 1983 revision of SINDA, a general thermal analyzer program. The purpose of the additional routines is to allow for the modeling of active heat transfer loops. The modeler can simulate the steady-state and pseudo-transient operations of 16 different heat transfer loop components including radiators, evaporators, condensers, mechanical pumps, reservoirs and many types of valves and fittings. In addition, the program contains a property analysis routine that can be used to compute the thermodynamic properties of 20 different refrigerants. SINFAC can simulate the response to transient boundary conditions. SINFAC was first developed as a method for computing the steady-state performance of two phase systems. It was then modified using CNFRWD, SINDA's explicit time-integration scheme, to accommodate transient thermal models. However, SINFAC cannot simulate pressure drops due to time-dependent fluid acceleration, transient boil-out, or transient fill-up, except in the accumulator. SINFAC also requires the user to be familiar with SINDA. The solution procedure used by SINFAC is similar to that which an engineer would use to solve a system manually. The solution to a system requires the determination of all of the outlet conditions of each component such as the flow rate, pressure, and enthalpy. To obtain these values, the user first estimates the inlet conditions to the first component of the system, then computes the outlet conditions from the data supplied by the manufacturer of the first component. The user then estimates the temperature at the outlet of the third component and computes the corresponding flow resistance of the second component. With the flow resistance of the second component, the user computes the conditions down stream, namely the inlet conditions of the third. The computations follow for the rest of the system, back to the first component. On the first pass, the user finds that the calculated outlet conditions of the last component do not match the estimated inlet conditions of the first. The user then modifies the estimated inlet conditions of the first component in an attempt to match the calculated values. The user estimated values are called State Variables. The differences between the user estimated values and calculated values are called the Error Variables. The procedure systematically changes the State Variables until all of the Error Variables are less than the user-specified iteration limits. The solution procedure is referred to as SCX. It consists of two phases, the Systems phase and the Controller phase. The X is to imply experimental. SCX computes each next set of State Variables in two phases. In the first phase, SCX fixes the controller positions and modifies the other State Variables by the Newton-Raphson method. This first phase is the Systems phase. Once the Newton-Raphson method has solved the problem for the fixed controller positions, SCX next calculates new controller positions based on Newton's method while treating each sensor-controller pair independently but allowing all to change in one iteration. This phase is the Controller phase. SINFAC is available by license for a period of ten (10) years to approved licensees. The licenced program product includes the source code for the additional routines to SINDA, the SINDA object code, command procedures, sample data and supporting documentation. Additional documentation may be purchased at the price below. SINFAC was created for use on a DEC VAX under VMS. Source code is written in FORTRAN 77, requires 180k of memory, and should be fully transportable. The program was developed in 1988.

Description: This theoretical aerodynamics program, TAD, was developed to predict the aerodynamic characteristics of vehicles with sounding rocket configurations. These slender, axisymmetric finned vehicle configurations have a wide range of aeronautical applications from rockets to high speed armament. Over a given range of Mach numbers, TAD will compute the normal force coefficient derivative, the center-of-pressure, the roll forcing moment coefficient derivative, the roll damping moment coefficient derivative, and the pitch damping moment coefficient derivative of a sounding rocket configured vehicle. The vehicle may consist of a sharp pointed nose of cone or tangent ogive shape, up to nine other body divisions of conical shoulder, conical boattail, or circular cylinder shape, and fins of trapezoid planform shape with constant cross section and either three or four fins per fin set. The characteristics computed by TAD have been shown to be accurate to within ten percent of experimental data in the supersonic region. The TAD program calculates the characteristics of separate portions of the vehicle, calculates the interference between separate portions of the vehicle, and then combines the results to form a total vehicle solution. Also, TAD can be used to calculate the characteristics of the body or fins separately as an aid in the design process. Input to the TAD program consists of simple descriptions of the body and fin geometries and the Mach range of interest. Output includes the aerodynamic characteristics of the total vehicle, or user-selected portions, at specified points over the mach range. The TAD program is written in FORTRAN IV for batch execution and has been implemented on an IBM 360 computer with a central memory requirement of approximately 123K of 8 bit bytes. The TAD program was originally developed in 1967 and last updated in 1972.

Description: The Thermal Radiation Analyzer System, TRASYS, is a computer software system with generalized capability to solve the radiation related aspects of thermal analysis problems. TRASYS computes the total thermal radiation environment for a spacecraft in orbit. The software calculates internode radiation interchange data as well as incident and absorbed heat rate data originating from environmental radiant heat sources. TRASYS provides data of both types in a format directly usable by such thermal analyzer programs as SINDA/FLUINT (available from COSMIC, program number MSC-21528). One primary feature of TRASYS is that it allows users to write their own driver programs to organize and direct the preprocessor and processor library routines in solving specific thermal radiation problems. The preprocessor first reads and converts the user's geometry input data into the form used by the processor library routines. Then, the preprocessor accepts the user's driving logic, written in the TRASYS modified FORTRAN language. In many cases, the user has a choice of routines to solve a given problem. Users may also provide their own routines where desirable. In particular, the user may write output routines to provide for an interface between TRASYS and any thermal analyzer program using the R-C network concept. Input to the TRASYS program consists of Options and Edit data, Model data, and Logic Flow and Operations data. Options and Edit data provide for basic program control and user edit capability. The Model data describe the problem in terms of geometry and other properties. This information includes surface geometry data, documentation data, nodal data, block coordinate system data, form factor data, and flux data. Logic Flow and Operations data house the user's driver logic, including the sequence of subroutine calls and the subroutine library. Output from TRASYS consists of two basic types of data: internode radiation interchange data, and incident and absorbed heat rate data. The flexible structure of TRASYS allows considerable freedom in the definition and choice of solution method for a thermal radiation problem. The program's flexible structure has also allowed TRASYS to retain the same basic input structure as the authors update it in order to keep up with changing requirements. Among its other important features are the following: 1) up to 3200 node problem size capability with shadowing by intervening opaque or semi-transparent surfaces; 2) choice of diffuse, specular, or diffuse/specular radiant interchange solutions; 3) a restart capability that minimizes recomputing; 4) macroinstructions that automatically provide the executive logic for orbit generation that optimizes the use of previously completed computations; 5) a time variable geometry package that provides automatic pointing of the various parts of an articulated spacecraft and an automatic look-back feature that eliminates redundant form factor calculations; 6) capability to specify submodel names to identify sets of surfaces or components as an entity; and 7) subroutines to perform functions which save and recall the internodal and/or space form factors in subsequent steps for nodes with fixed geometry during a variable geometry run. There are two machine versions of TRASYS v27: a DEC VAX version and a Cray UNICOS version. Both versions require installation of the NASADIG library (MSC-21801 for DEC VAX or COS-10049 for CRAY), which is available from COSMIC either separately or bundled with TRASYS. The NASADIG (NASA Device Independent Graphics Library) plot package provides a pictorial representation of input geometry, orbital/orientation parameters, and heating rate output as a function of time. NASADIG supports Tektronix terminals. The CRAY version of TRASYS v27 is written in FORTRAN 77 for batch or interactive execution and has been implemented on CRAY X-MP and CRAY Y-MP series computers running UNICOS. The standard distribution medium for MSC-21959 (CRAY version without NASADIG) is a 1600 BPI 9-track magnetic tape in UNIX tar format. The standard distribution medium for COS-10040 (CRAY version with NASADIG) is a set of two 6250 BPI 9-track magnetic tapes in UNIX tar format. Alternate distribution media and formats are available upon request. The DEC VAX version of TRASYS v27 is written in FORTRAN 77 for batch execution (only the plotting driver program is interactive) and has been implemented on a DEC VAX 8650 computer under VMS. Since the source codes for MSC-21030 and COS-10026 are in VAX/VMS text library files and DEC Command Language files, COSMIC will only provide these programs in the following formats: MSC-21030, TRASYS (DEC VAX version without NASADIG) is available on a 1600 BPI 9-track magnetic tape in VAX BACKUP format (standard distribution medium) or in VAX BACKUP format on a TK50 tape cartridge; COS-10026, TRASYS (DEC VAX version with NASADIG), is available in VAX BACKUP format on a set of three 6250 BPI 9-track magnetic tapes (standard distribution medium) or a set of three TK50 tape cartridges in VAX BACKUP format. TRASYS was last updated in 1993.

Description: INS3D computes steady-state solutions to the incompressible Navier-Stokes equations. The INS3D approach utilizes pseudo-compressibility combined with an approximate factorization scheme. This computational fluid dynamics (CFD) code has been verified on problems such as flow through a channel, flow over a backwardfacing step and flow over a circular cylinder. Three dimensional cases include flow over an ogive cylinder, flow through a rectangular duct, wind tunnel inlet flow, cylinder-wall juncture flow and flow through multiple posts mounted between two plates. INS3D uses a pseudo-compressibility approach in which a time derivative of pressure is added to the continuity equation, which together with the momentum equations form a set of four equations with pressure and velocity as the dependent variables. The equations' coordinates are transformed for general three dimensional applications. The equations are advanced in time by the implicit, non-iterative, approximately-factored, finite-difference scheme of Beam and Warming. The numerical stability of the scheme depends on the use of higher-order smoothing terms to damp out higher-frequency oscillations caused by second-order central differencing. The artificial compressibility introduces pressure (sound) waves of finite speed (whereas the speed of sound would be infinite in an incompressible fluid). As the solution converges, these pressure waves die out, causing the derivation of pressure with respect to time to approach zero. Thus, continuity is satisfied for the incompressible fluid in the steady state. Computational efficiency is achieved using a diagonal algorithm. A block tri-diagonal option is also available. When a steady-state solution is reached, the modified continuity equation will satisfy the divergence-free velocity field condition. INS3D is capable of handling several different types of boundaries encountered in numerical simulations, including solid-surface, inflow and outflow, and far-field boundaries. Three machine versions of INS3D are available. INS3D for the CRAY is written in CRAY FORTRAN for execution on a CRAY X-MP under COS, INS3D for the IBM is written in FORTRAN 77 for execution on an IBM 3090 under the VM or MVS operating system, and INS3D for DEC RISC-based systems is written in RISC FORTRAN for execution on a DEC workstation running RISC ULTRIX 3.1 or later. The CRAY version has a central memory requirement of 730279 words. The central memory requirement for the IBM is 150Mb. The memory requirement for the DEC RISC ULTRIX version is 3Mb of main memory. INS3D was developed in 1987. The port to the IBM was done in 1990. The port to the DECstation 3100 was done in 1991. CRAY is a registered trademark of Cray Research Inc. IBM is a registered trademark of International Business Machines. DEC, DECstation, and ULTRIX are trademarks of the Digital Equipment Corporation.

Description: SAGE, Self Adaptive Grid codE, is a flexible tool for adapting and restructuring both 2D and 3D grids. Solution-adaptive grid methods are useful tools for efficient and accurate flow predictions. In supersonic and hypersonic flows, strong gradient regions such as shocks, contact discontinuities, shear layers, etc., require careful distribution of grid points to minimize grid error and produce accurate flow-field predictions. SAGE helps the user obtain more accurate solutions by intelligently redistributing (i.e. adapting) the original grid points based on an initial or interim flow-field solution. The user then computes a new solution using the adapted grid as input to the flow solver. The adaptive-grid methodology poses the problem in an algebraic, unidirectional manner for multi-dimensional adaptations. The procedure is analogous to applying tension and torsion spring forces proportional to the local flow gradient at every grid point and finding the equilibrium position of the resulting system of grid points. The multi-dimensional problem of grid adaption is split into a series of one-dimensional problems along the computational coordinate lines. The reduced one dimensional problem then requires a tridiagonal solver to find the location of grid points along a coordinate line. Multi-directional adaption is achieved by the sequential application of the method in each coordinate direction. The tension forces direct the redistribution of points to the strong gradient region. To maintain smoothness and a measure of orthogonality of grid lines, torsional forces are introduced that relate information between the family of lines adjacent to one another. The smoothness and orthogonality constraints are direction-dependent, since they relate only the coordinate lines that are being adapted to the neighboring lines that have already been adapted. Therefore the solutions are non-unique and depend on the order and direction of adaption. Non-uniqueness of the adapted grid is acceptable since it makes possible an overall and local error reduction through grid redistribution. SAGE includes the ability to modify the adaption techniques in boundary regions, which substantially improves the flexibility of the adaptive scheme. The vectorial approach used in the analysis also provides flexibility. The user has complete choice of adaption direction and order of sequential adaptions without concern for the computational data structure. Multiple passes are available with no restraint on stepping directions; for each adaptive pass the user can choose a completely new set of adaptive parameters. This facility, combined with the capability of edge boundary control, enables the code to individually adapt multi-dimensional multiple grids. Zonal grids can be adapted while maintaining continuity along the common boundaries. For patched grids, the multiple-pass capability enables complete adaption. SAGE is written in FORTRAN 77 and is intended to be machine independent; however, it requires a FORTRAN compiler which supports NAMELIST input. It has been successfully implemented on Sun series computers, SGI IRIS's, DEC MicroVAX computers, HP series computers, the Cray YMP, and IBM PC compatibles. Source code is provided, but no sample input and output files are provided. The code reads three datafiles: one that contains the initial grid coordinates (x,y,z), one that contains corresponding flow-field variables, and one that contains the user control parameters. It is assumed that the first two datasets are formatted as defined in the plotting software package PLOT3D. Several machine versions of PLOT3D are available from COSMIC. The amount of main memory is dependent on the size of the matrix. The standard distribution medium for SAGE is a 5.25 inch 360K MS-DOS format diskette. It is also available on a .25 inch streaming magnetic tape cartridge in UNIX tar format or on a 9-track 1600 BPI ASCII CARD IMAGE format magnetic tape. SAGE was developed in 1989, first released as a 2D version in 1991 and updated to 3D in 1993.

Description: Delaminations near the outer surface of a laminate are susceptible to local buckling and buckling-induced delamination propagation when the laminate is subjected to transverse impact loading. This results in a loss of stiffness and strength. TRBUCKL is an unique dynamic delamination buckling and delamination propagation analysis capability that can be incorporated into the structural analysis program, NASTRAN. This capability will aid engineers in the design of structures incorporating composite laminates. The capability consists of: (1) a modification of the direct time integration solution sequence which provides a new analysis algorithm that can be used to predict delamination buckling in a laminate subjected to dynamic loading; and (2) a new method of modeling the composite laminate using plate bending elements and multipoint constraints. The capability now exists to predict the time at which the onset of dynamic delamination buckling occurs, the dynamic buckling mode shape, and the dynamic delamination strain energy release rate. A procedure file for NASTRAN, TRBUCKL predicts both impact induced buckling in composite laminates with initial delaminations and the strain energy release rate due to extension of the delamination. In addition, the file is useful in calculating the dynamic delamination strain energy release rate for a composite laminate under impact loading. This procedure simplifies the simulation of progressive crack extension. TRBUCKL has been incorporated into COSMIC NASTRAN. TRBUCKL is a DMAP Alter for NASTRAN. It is intended for use only with the COSMIC NASTRAN Direct Transient Analysis (RF 9) solution sequence. The program is available as a listing only. TRBUCKL was developed in 1987.

Description: INS3D computes steady-state solutions to the incompressible Navier-Stokes equations. The INS3D approach utilizes pseudo-compressibility combined with an approximate factorization scheme. This computational fluid dynamics (CFD) code has been verified on problems such as flow through a channel, flow over a backwardfacing step and flow over a circular cylinder. Three dimensional cases include flow over an ogive cylinder, flow through a rectangular duct, wind tunnel inlet flow, cylinder-wall juncture flow and flow through multiple posts mounted between two plates. INS3D uses a pseudo-compressibility approach in which a time derivative of pressure is added to the continuity equation, which together with the momentum equations form a set of four equations with pressure and velocity as the dependent variables. The equations' coordinates are transformed for general three dimensional applications. The equations are advanced in time by the implicit, non-iterative, approximately-factored, finite-difference scheme of Beam and Warming. The numerical stability of the scheme depends on the use of higher-order smoothing terms to damp out higher-frequency oscillations caused by second-order central differencing. The artificial compressibility introduces pressure (sound) waves of finite speed (whereas the speed of sound would be infinite in an incompressible fluid). As the solution converges, these pressure waves die out, causing the derivation of pressure with respect to time to approach zero. Thus, continuity is satisfied for the incompressible fluid in the steady state. Computational efficiency is achieved using a diagonal algorithm. A block tri-diagonal option is also available. When a steady-state solution is reached, the modified continuity equation will satisfy the divergence-free velocity field condition. INS3D is capable of handling several different types of boundaries encountered in numerical simulations, including solid-surface, inflow and outflow, and far-field boundaries. Three machine versions of INS3D are available. INS3D for the CRAY is written in CRAY FORTRAN for execution on a CRAY X-MP under COS, INS3D for the IBM is written in FORTRAN 77 for execution on an IBM 3090 under the VM or MVS operating system, and INS3D for DEC RISC-based systems is written in RISC FORTRAN for execution on a DEC workstation running RISC ULTRIX 3.1 or later. The CRAY version has a central memory requirement of 730279 words. The central memory requirement for the IBM is 150Mb. The memory requirement for the DEC RISC ULTRIX version is 3Mb of main memory. INS3D was developed in 1987. The port to the IBM was done in 1990. The port to the DECstation 3100 was done in 1991. CRAY is a registered trademark of Cray Research Inc. IBM is a registered trademark of International Business Machines. DEC, DECstation, and ULTRIX are trademarks of the Digital Equipment Corporation.

Description: The ability to treat arbitrary boundary shapes is one of the most desirable characteristics of a method for generating grids, including those about airfoils. In a grid used for computing aerodynamic flow over an airfoil, or any other body shape, the surface of the body is usually treated as an inner boundary and often cannot be easily represented as an analytic function. The GRAPE computer program was developed to incorporate a method for generating two-dimensional finite-difference grids about airfoils and other shapes by the use of the Poisson differential equation. GRAPE can be used with any boundary shape, even one specified by tabulated points and including a limited number of sharp corners. The GRAPE program has been developed to be numerically stable and computationally fast. GRAPE can provide the aerodynamic analyst with an efficient and consistent means of grid generation. The GRAPE procedure generates a grid between an inner and an outer boundary by utilizing an iterative procedure to solve the Poisson differential equation subject to geometrical restraints. In this method, the inhomogeneous terms of the equation are automatically chosen such that two important effects are imposed on the grid. The first effect is control of the spacing between mesh points along mesh lines intersecting the boundaries. The second effect is control of the angles with which mesh lines intersect the boundaries. Along with the iterative solution to Poisson's equation, a technique of coarse-fine sequencing is employed to accelerate numerical convergence. GRAPE program control cards and input data are entered via the NAMELIST feature. Each variable has a default value such that user supplied data is kept to a minimum. Basic input data consists of the boundary specification, mesh point spacings on the boundaries, and mesh line angles at the boundaries. Output consists of a dataset containing the grid data and, if requested, a plot of the generated mesh. The GRAPE program is written in FORTRAN IV for batch execution and has been implemented on a CDC 6000 series computer with a central memory requirement of approximately 135K (octal) of 60 bit words. For plotted output the commercially available DISSPLA graphics software package is required. The GRAPE program was developed in 1980.

Description: This program, which is called 'AOFA', determines the complete viscous and inviscid flow around a body of revolution at a given angle of attack and traveling at supersonic speeds. The viscous calculations from this program agree with experimental values for surface and pitot pressures and with surface heating rates. At high speeds, lee-side flows are important because the local heating is difficult to correlate and because the shed vortices can interact with vehicle components such as a canopy or a vertical tail. This program should find application in the design analysis of any high speed vehicle. Lee-side flows are difficult to calculate because thin-boundary-layer theory is not applicable and the concept of matching inviscid and viscous flow is questionable. This program uses the parabolic approximation to the compressible Navier-Stokes equations and solves for the complete inviscid and viscous regions of flow, including the pressure. The parabolic approximation results from the assumption that the stress derivatives in the streamwise direction are small in comparison with derivatives in the normal and circumferential directions. This assumption permits the equation to be solved by an implicit finite difference marching technique which proceeds downstream from the initial data point, provided the inviscid portion of flow is supersonic. The viscous cross-flow separation is also determined as part of the solution. To use this method it is necessary to first determine an initial data point in a region where the inviscid portion of the flow is supersonic. Input to this program consists of two parts. Problem description is conveyed to the program by namelist input. Initial data is acquired by the program as formatted data. Because of the large amount of run time this program can consume the program includes a restart capability. Output is in printed format and magnetic tape for further processing. This program is written in FORTRAN IV and has been implemented on a CDC 7600 with a central memory requirement of approximately 35K (octal) of 60 bit words.

Description: Predictions of fatigue crack growth behavior can be made with the Fatigue Crack Growth Structural Analysis (FASTRAN II) computer program. As cyclic loads are applied to a selected crack configuration with an initial crack size, FASTRAN II predicts crack growth as a function of cyclic load history until either a desired crack size is reached or failure occurs. FASTRAN II is based on plasticity-induced crack-closure behavior of cracks in metallic materials and accounts for load-interaction effects, such as retardation and acceleration, under variable-amplitude loading. The closure model is based on the Dugdale model with modifications to allow plastically deformed material to be left along the crack surfaces as the crack grows. Plane stress and plane strain conditions, as well as conditions between these two, can be simulated in FASTRAN II by using a constraint factor on tensile yielding at the crack front to approximately account for three-dimensional stress states. FASTRAN II contains seventeen predefined crack configurations (standard laboratory fatigue crack growth rate specimens and many common crack configurations found in structures); and the user can define one additional crack configuration. The baseline crack growth rate properties (effective stress-intensity factor against crack growth rate) may be given in either equation or tabular form. For three-dimensional crack configurations, such as surface cracks or corner cracks at holes or notches, the fatigue crack growth rate properties may be different in the crack depth and crack length directions. Final failure of the cracked structure can be modelled with fracture toughness properties using either linear-elastic fracture mechanics (brittle materials), a two-parameter fracture criterion (brittle to ductile materials), or plastic collapse (extremely ductile materials). The crack configurations in FASTRAN II can be subjected to either constant-amplitude, variable-amplitude or spectrum loading. The applied loads may be either tensile or compressive. Several standardized aircraft flight-load histories, such as TWIST, Mini-TWIST, FALSTAFF, Inverted FALSTAFF, Felix and Gaussian, are included as options. FASTRAN II also includes two other methods that will help the user input spectrum load histories. The two methods are: (1) a list of stress points, and (2) a flight-by-flight history of stress points. Examples are provided in the user manual. Developed as a research program, FASTRAN II has successfully predicted crack growth in many metallic materials under various aircraft spectrum loading. A computer program DKEFF which is a part of the FASTRAN II package was also developed to analyze crack growth rate data from laboratory specimens to obtain the effective stress-intensity factor against crack growth rate relations used in FASTRAN II. FASTRAN II is written in standard FORTRAN 77. It has been successfully compiled and implemented on Sun4 series computers running SunOS and on IBM PC compatibles running MS-DOS using the Lahey F77L FORTRAN compiler. Sample input and output data are included with the FASTRAN II package. The UNIX version requires 660K of RAM for execution. The standard distribution medium for the UNIX version (LAR-14865) is a .25 inch streaming magnetic tape cartridge in UNIX tar format. It is also available on a 3.5 inch diskette in UNIX tar format. The standard distribution medium for the MS-DOS version (LAR-14944) is a 5.25 inch 360K MS-DOS format diskette. The contents of the diskette are compressed using the PKWARE archiving tools. The utility to unarchive the files, PKUNZIP.EXE, is included. The program was developed in 1984 and revised in 1992. Sun4 and SunOS are trademarks of Sun Microsystems, Inc. IBM PC is a trademark of International Business Machines Corp. MS-DOS is a trademark of Microsoft, Inc. F77L is a trademark of the Lahey Computer Systems, Inc. UNIX is a registered trademark of AT&T Bell Laboratories. PKWARE and PKUNZIP are trademarks of PKWare, Inc.

Description: The Panel Analysis and Sizing Code (PASCO) was developed for the buckling and vibration analysis and sizing of prismatic structures having an arbitrary cross section. PASCO is primarily intended for analyzing and sizing stiffened panels made of laminated orthotropic materials and is of particular value in analyzing and sizing filamentary composite structures. When used in the analysis mode, PASCO calculates laminate stiffnesses, lamina stress and strains (including the effects of temperature and panel bending), buckling loads, vibration frequencies, and overall panel stiffness. When used in the sizing mode, PASCO adjusts sizing variables to provide a low-mass panel design that carries a set of specified loadings without exceeding buckling or material strength allowables and that meets other design requirements such as upper and lower bounds on sizing variables, upper and lower bounds on overall bending, extensional and shear stiffnesses, and lower bounds on vibration frequencies. Although emphasis in PASCO is placed on flat panels having several identical bays, the only restriction on configuration modeling is that the structure is assumed to be prismatic. In addition, it is assumed that loads and temperatures do not vary along the length of a panel. Because of their wide application in aerospace structures, stiffened panels are readily handled by PASCO. The panel cross section may be composed of an arbitrary assemblage of thin, flat, rectangular plate elements that are connected together along their longitudinal edges. Each plate element consists of a balanced symmetric laminate of any number of layers of orthotropic material. Any group of element widths, layer thicknesses, and layer orientation angles may be selected as sizing variables. Substructuring is available to increase the efficiency of the analysis and to simplify the modeling of complicated structures. The Macintosh version of PASCO includes an interactive, graphic preprocessor called MacPASCO. The main objective of MacPASCO is to make the use of PASCO faster, simpler, and less error-prone. By using a graphical user interface (GUI), MacPASCO simplifies the specification of panel geometry and reduces user input errors, thus making the modeling and analysis of panel designs more efficient. The user draws the initial structural geometry on the computer screen, then uses a combination of graphic and text inputs to: refine the structural geometry, specify information required for analysis such as panel load conditions, and define design variables and constraints for minimum-mass optimization. Composite panel design is an ideal application because the graphical user interface can: serve as a visual aid, eliminate the tedious aspects of text-based input, and eliminate many sources of input errors. The current version of MacPASCO does not implement all the modeling features of PASCO, but has been found to be sufficient for many users. Many difficulties common to text-based inputs are avoided because MacPASCO uses a GUI. First, the graphic displays eliminate syntax errors, like misplaced commas and incorrect command names, because there is no text-based syntax. Second, graphic displays allow the user to see the geometry as it is created and immediately detect and correct any errors. Third, MacPASCO's drawing tools have been designed to avoid modeling errors. Fourth, the graphic displays make revisions to existing structural designs much easier and less error-prone by eliminating the need for the user to conceptualize the text input as geometry. The user can work directly with the geometry displayed on the screen. Finally, MacPASCO automatically generates the correct PASCO input from the geometry displayed on the screen. This input file can be used with any machine version of PASCO to actually perform the analysis and sizing and to output results. The DEC VAX version of PASCO is written in FORTRAN IV for batch execution and has been implemented on a DEC VAX series computer. The Macintosh version of PASCO was developed for Macintosh II series computers with at least 2Mb of RAM running MPW Pascal 3.0 and Language Systems FORTRAN 2.0 under the MPW programming environment. It includes MPW compatible makefiles for compiling the source code. The Macintosh version uses input files compatible with versions of PASCO running on different platforms. MacPASCO is written in Macintosh Programmers Workbench 3.0, MPW Pascal 3.0, and MacAPP 2.0. The Pascal source code is included on the distribution diskette. MacAPP is a development library which is not included. MacPASCO requires a Mac Plus, SE/30, or MacII, IIx, IIcx, IIci, or IIfx running System 6.0 or greater. MacPASCO is System 7.0 compatible. A minimum of 2Mb of RAM is required for execution. The Macintosh version of PASCO is distributed on four 3.5 inch 800K Macintosh format diskettes. The DEC VAX version is distributed on a 9-track 1600 BPI magnetic tape. The PASCO program was developed in 1981, adapted to the DEC VAX in 1983 and to the Macintosh in 1991. MacPASCO was released in 1992.

Description: The Thermal Radiation Analyzer System, TRASYS, is a computer software system with generalized capability to solve the radiation related aspects of thermal analysis problems. TRASYS computes the total thermal radiation environment for a spacecraft in orbit. The software calculates internode radiation interchange data as well as incident and absorbed heat rate data originating from environmental radiant heat sources. TRASYS provides data of both types in a format directly usable by such thermal analyzer programs as SINDA/FLUINT (available from COSMIC, program number MSC-21528). One primary feature of TRASYS is that it allows users to write their own driver programs to organize and direct the preprocessor and processor library routines in solving specific thermal radiation problems. The preprocessor first reads and converts the user's geometry input data into the form used by the processor library routines. Then, the preprocessor accepts the user's driving logic, written in the TRASYS modified FORTRAN language. In many cases, the user has a choice of routines to solve a given problem. Users may also provide their own routines where desirable. In particular, the user may write output routines to provide for an interface between TRASYS and any thermal analyzer program using the R-C network concept. Input to the TRASYS program consists of Options and Edit data, Model data, and Logic Flow and Operations data. Options and Edit data provide for basic program control and user edit capability. The Model data describe the problem in terms of geometry and other properties. This information includes surface geometry data, documentation data, nodal data, block coordinate system data, form factor data, and flux data. Logic Flow and Operations data house the user's driver logic, including the sequence of subroutine calls and the subroutine library. Output from TRASYS consists of two basic types of data: internode radiation interchange data, and incident and absorbed heat rate data. The flexible structure of TRASYS allows considerable freedom in the definition and choice of solution method for a thermal radiation problem. The program's flexible structure has also allowed TRASYS to retain the same basic input structure as the authors update it in order to keep up with changing requirements. Among its other important features are the following: 1) up to 3200 node problem size capability with shadowing by intervening opaque or semi-transparent surfaces; 2) choice of diffuse, specular, or diffuse/specular radiant interchange solutions; 3) a restart capability that minimizes recomputing; 4) macroinstructions that automatically provide the executive logic for orbit generation that optimizes the use of previously completed computations; 5) a time variable geometry package that provides automatic pointing of the various parts of an articulated spacecraft and an automatic look-back feature that eliminates redundant form factor calculations; 6) capability to specify submodel names to identify sets of surfaces or components as an entity; and 7) subroutines to perform functions which save and recall the internodal and/or space form factors in subsequent steps for nodes with fixed geometry during a variable geometry run. There are two machine versions of TRASYS v27: a DEC VAX version and a Cray UNICOS version. Both versions require installation of the NASADIG library (MSC-21801 for DEC VAX or COS-10049 for CRAY), which is available from COSMIC either separately or bundled with TRASYS. The NASADIG (NASA Device Independent Graphics Library) plot package provides a pictorial representation of input geometry, orbital/orientation parameters, and heating rate output as a function of time. NASADIG supports Tektronix terminals. The CRAY version of TRASYS v27 is written in FORTRAN 77 for batch or interactive execution and has been implemented on CRAY X-MP and CRAY Y-MP series computers running UNICOS. The standard distribution medium for MSC-21959 (CRAY version without NASADIG) is a 1600 BPI 9-track magnetic tape in UNIX tar format. The standard distribution medium for COS-10040 (CRAY version with NASADIG) is a set of two 6250 BPI 9-track magnetic tapes in UNIX tar format. Alternate distribution media and formats are available upon request. The DEC VAX version of TRASYS v27 is written in FORTRAN 77 for batch execution (only the plotting driver program is interactive) and has been implemented on a DEC VAX 8650 computer under VMS. Since the source codes for MSC-21030 and COS-10026 are in VAX/VMS text library files and DEC Command Language files, COSMIC will only provide these programs in the following formats: MSC-21030, TRASYS (DEC VAX version without NASADIG) is available on a 1600 BPI 9-track magnetic tape in VAX BACKUP format (standard distribution medium) or in VAX BACKUP format on a TK50 tape cartridge; COS-10026, TRASYS (DEC VAX version with NASADIG), is available in VAX BACKUP format on a set of three 6250 BPI 9-track magnetic tapes (standard distribution medium) or a set of three TK50 tape cartridges in VAX BACKUP format. TRASYS was last updated in 1993.

Description: The Comprehensive Analytical Model of Rotorcraft Aerodynamics, CAMRAD, program is designed to calculate rotor performance, loads, and noise; helicopter vibration and gust response; flight dynamics and handling qualities; and system aeroelastic stability. The analysis is a consistent combination of structural, inertial, and aerodynamic models applicable to a wide range of problems and a wide class of vehicles. The CAMRAD analysis can be applied to articulated, hingeless, gimballed, and teetering rotors with an arbitrary number of blades. The rotor degrees of freedom included are blade/flap bending, rigid pitch and elastic torsion, and optionally gimbal or teeter motion. General two-rotor aircrafts can be modeled. Single main-rotor and tandem helicopter and sideby-side tilting proprotor aircraft configurations can be considered. The case of a rotor or helicopter in a wind tunnel can also be modeled. The aircraft degrees of freedom included are the six rigid body motion, elastic airframe motions, and the rotor/engine speed perturbations. CAMRAD calculates the load and motion of helicopters and airframes in two stages. First the trim solution is obtained; then the flutter, flight dynamics, and/or transient behavior can be calculated. The trim operating conditions considered include level flight, steady climb or descent, and steady turns. The analysis of the rotor includes nonlinear inertial and aerodynamic models, applicable to large blade angles and a high inflow ratio, The rotor aerodynamic model is based on two-dimensional steady airfoil characteristics with corrections for three-dimensional and unsteady flow effects, including a dynamic stall model. In the flutter analysis, the matrices are constructed that describe the linear differential equations of motion, and the equations are analyzed. In the flight dynamics analysis, the stability derivatives are calculated and the matrices are constructed that describe the linear differential equations of motion. These equations are analyzed. In the transient analysis, the rigid body equations of motion are numerically integrated, for a prescribed transient gust or control input. The CAMRAD program product is available by license for a period of ten years to domestic U.S. licensees. The licensed program product includes the CAMRAD source code, command procedures, sample applications, and one set of supporting documentation. Copies of the documentation may be purchased separately at the price indicated below. CAMRAD is written in FORTRAN 77 for the DEC VAX under VMS 4.6 with a recommended core memory of 4.04 megabytes. The DISSPLA package is necessary for graphical output. CAMRAD was developed in 1980.

Description: ARC2D is a computational fluid dynamics program developed at the NASA Ames Research Center specifically for airfoil computations. The program uses implicit finite-difference techniques to solve two-dimensional Euler equations and thin layer Navier-Stokes equations. It is based on the Beam and Warming implicit approximate factorization algorithm in generalized coordinates. The methods are either time accurate or accelerated non-time accurate steady state schemes. The evolution of the solution through time is physically realistic; good solution accuracy is dependent on mesh spacing and boundary conditions. The mathematical development of ARC2D begins with the strong conservation law form of the two-dimensional Navier-Stokes equations in Cartesian coordinates, which admits shock capturing. The Navier-Stokes equations can be transformed from Cartesian coordinates to generalized curvilinear coordinates in a manner that permits one computational code to serve a wide variety of physical geometries and grid systems. ARC2D includes an algebraic mixing length model to approximate the effect of turbulence. In cases of high Reynolds number viscous flows, thin layer approximation can be applied. ARC2D allows for a variety of solutions to stability boundaries, such as those encountered in flows with shocks. The user has considerable flexibility in assigning geometry and developing grid patterns, as well as in assigning boundary conditions. However, the ARC2D model is most appropriate for attached and mildly separated boundary layers; no attempt is made to model wake regions and widely separated flows. The techniques have been successfully used for a variety of inviscid and viscous flowfield calculations. The Cray version of ARC2D is written in FORTRAN 77 for use on Cray series computers and requires approximately 5Mb memory. The program is fully vectorized. The tape includes variations for the COS and UNICOS operating systems. Also included is a sample routine for CONVEX computers to emulate Cray system time calls, which should be easy to modify for other machines as well. The standard distribution media for this version is a 9-track 1600 BPI ASCII Card Image format magnetic tape. The Cray version was developed in 1987. The IBM ES/3090 version is an IBM port of the Cray version. It is written in IBM VS FORTRAN and has the capability of executing in both vector and parallel modes on the MVS/XA operating system and in vector mode on the VM/XA operating system. Various options of the IBM VS FORTRAN compiler provide new features for the ES/3090 version, including 64-bit arithmetic and up to 2 GB of virtual addressability. The IBM ES/3090 version is available only as a 9-track, 1600 BPI IBM IEBCOPY format magnetic tape. The IBM ES/3090 version was developed in 1989. The DEC RISC ULTRIX version is a DEC port of the Cray version. It is written in FORTRAN 77 for RISC-based Digital Equipment platforms. The memory requirement is approximately 7Mb of main memory. It is available in UNIX tar format on TK50 tape cartridge. The port to DEC RISC ULTRIX was done in 1990. COS and UNICOS are trademarks and Cray is a registered trademark of Cray Research, Inc. IBM, ES/3090, VS FORTRAN, MVS/XA, and VM/XA are registered trademarks of International Business Machines. DEC and ULTRIX are registered trademarks of Digital Equipment Corporation.

Description: INS3D computes steady-state solutions to the incompressible Navier-Stokes equations. The INS3D approach utilizes pseudo-compressibility combined with an approximate factorization scheme. This computational fluid dynamics (CFD) code has been verified on problems such as flow through a channel, flow over a backwardfacing step and flow over a circular cylinder. Three dimensional cases include flow over an ogive cylinder, flow through a rectangular duct, wind tunnel inlet flow, cylinder-wall juncture flow and flow through multiple posts mounted between two plates. INS3D uses a pseudo-compressibility approach in which a time derivative of pressure is added to the continuity equation, which together with the momentum equations form a set of four equations with pressure and velocity as the dependent variables. The equations' coordinates are transformed for general three dimensional applications. The equations are advanced in time by the implicit, non-iterative, approximately-factored, finite-difference scheme of Beam and Warming. The numerical stability of the scheme depends on the use of higher-order smoothing terms to damp out higher-frequency oscillations caused by second-order central differencing. The artificial compressibility introduces pressure (sound) waves of finite speed (whereas the speed of sound would be infinite in an incompressible fluid). As the solution converges, these pressure waves die out, causing the derivation of pressure with respect to time to approach zero. Thus, continuity is satisfied for the incompressible fluid in the steady state. Computational efficiency is achieved using a diagonal algorithm. A block tri-diagonal option is also available. When a steady-state solution is reached, the modified continuity equation will satisfy the divergence-free velocity field condition. INS3D is capable of handling several different types of boundaries encountered in numerical simulations, including solid-surface, inflow and outflow, and far-field boundaries. Three machine versions of INS3D are available. INS3D for the CRAY is written in CRAY FORTRAN for execution on a CRAY X-MP under COS, INS3D for the IBM is written in FORTRAN 77 for execution on an IBM 3090 under the VM or MVS operating system, and INS3D for DEC RISC-based systems is written in RISC FORTRAN for execution on a DEC workstation running RISC ULTRIX 3.1 or later. The CRAY version has a central memory requirement of 730279 words. The central memory requirement for the IBM is 150Mb. The memory requirement for the DEC RISC ULTRIX version is 3Mb of main memory. INS3D was developed in 1987. The port to the IBM was done in 1990. The port to the DECstation 3100 was done in 1991. CRAY is a registered trademark of Cray Research Inc. IBM is a registered trademark of International Business Machines. DEC, DECstation, and ULTRIX are trademarks of the Digital Equipment Corporation.

Description: Panel methods are moderate cost tools for solving a wide range of engineering problems. PMARC (Panel Method Ames Research Center) is a potential flow panel code that numerically predicts flow fields around complex three-dimensional geometries. PMARC's predecessor was a panel code named VSAERO which was developed for NASA by Analytical Methods, Inc. PMARC is a new program with many additional subroutines and a well-documented code suitable for powered-lift aerodynamic predictions. The program's open architecture facilitates modifications or additions of new features. Another improvement is the adjustable size code which allows for an optimum match between the computer hardware available to the user and the size of the problem being solved. PMARC can be resized (the maximum number of panels can be changed) in a matter of minutes. Several other state-of-the-art PMARC features include internal flow modeling for ducts and wind tunnel test sections, simple jet plume modeling essential for the analysis and design of powered-lift aircraft, and a time-stepping wake model which allows the study of both steady and unsteady motions. PMARC is a low-order panel method, which means the singularities are distributed with constant strength over each panel. In many cases low-order methods can provide nearly the same accuracy as higher order methods (where the singularities are allowed to vary linearly or quadratically over each panel). Low-order methods have the advantage of a shorter computation time and do not require exact matching between panels. The flow problem is solved by assuming that the body is at rest in a moving flow field. The body is modeled as a closed surface which divides space into two regions -- one region contains the flow field of interest and the other contains a fictitious flow. External flow problems, such as a wing in a uniform stream, have the external region as the flow field of interest and the internal flow as the fictitious flow. This arrangement is reversed for internal flow problems where the internal region contains the flow field of interest and the external flow field is fictitious. In either case it is assumed that the velocity potentials in both regions satisfy Laplace's equation. PMARC has extensive geometry modeling capabilities for handling complex, three-dimensional surfaces. As with all panel methods, the geometry must be modeled by a set of panels. For convenience, the geometry is usually subdivided into several pieces and modeled with sets of panels called patches. A patch may be folded over on itself so that opposing sides of the patch form a common line. For example, wings are normally modeled with a folded patch to form the trailing edge of the wing. PMARC also has the capability to automatically generate a closing tip patch. In the case of a wing, a tip patch could be generated to close off the wing's third side. PMARC has a simple jet model for simulating a jet plume in a crossflow. The jet plume shape, trajectory, and entrainment velocities are computed using the Adler/Baron jet in crossflow code. This information is then passed back to PMARC. The wake model in PMARC is a time-stepping wake model. The wake is convected downstream from the wake separation line by the local velocity flowfield. With each time step, a new row of wake panels is added to the wake at the wake separation line. PMARC also allows an initial wake to be specified if desired, or, as a third option, no wakes need be modeled. The effective presentation of results for aerodynamics problems requires the generation of report-quality graphics. PMAPP (ARC-12751), the Panel Method Aerodynamic Plotting Program, (Sterling Software), was written for scientists at NASA's Ames Research Center to plot the aerodynamic analysis results (flow data) from PMARC. PMAPP is an interactive, color-capable graphics program for the DEC VAX or MicroVAX running VMS. It was designed to work with a variety of terminal types and hardcopy devices. PMAPP is available separately from COSMIC. PMARC was written in standard FORTRAN77 using adjustable size arrays throughout the code. Redimensioning PMARC will change the amount of disk space and memory the code requires to be able to run; however, due to its memory requirements, this program does not readily lend itself to implementation on MS-DOS based machines. The program was implemented on an Apple Macintosh (using 2.5 MB of memory) and tested on a VAX/VMS computer. The program is available on a 3.5 inch Macintosh format diskette (standard media) or in VAX BACKUP format on TK50 tape cartridge or 9-track magnetic tape. PMARC was developed in 1989.

Description: The Transonic Airfoil analysis computer code, TAIR, was developed to employ a fast, fully implicit algorithm to solve the conservative full-potential equation for the steady transonic flow field about an arbitrary airfoil immersed in a subsonic free stream. The full-potential formulation is considered exact under the assumptions of irrotational, isentropic, and inviscid flow. These assumptions are valid for a wide range of practical transonic flows typical of modern aircraft cruise conditions. The primary features of TAIR include: a new fully implicit iteration scheme which is typically many times faster than classical successive line overrelaxation algorithms; a new, reliable artifical density spatial differencing scheme treating the conservative form of the full-potential equation; and a numerical mapping procedure capable of generating curvilinear, body-fitted finite-difference grids about arbitrary airfoil geometries. Three aspects emphasized during the development of the TAIR code were reliability, simplicity, and speed. The reliability of TAIR comes from two sources: the new algorithm employed and the implementation of effective convergence monitoring logic. TAIR achieves ease of use by employing a "default mode" that greatly simplifies code operation, especially by inexperienced users, and many useful options including: several airfoil-geometry input options, flexible user controls over program output, and a multiple solution capability. The speed of the TAIR code is attributed to the new algorithm and the manner in which it has been implemented. Input to the TAIR program consists of airfoil coordinates, aerodynamic and flow-field convergence parameters, and geometric and grid convergence parameters. The airfoil coordinates for many airfoil shapes can be generated in TAIR from just a few input parameters. Most of the other input parameters have default values which allow the user to run an analysis in the default mode by specifing only a few input parameters. Output from TAIR may include aerodynamic coefficients, the airfoil surface solution, convergence histories, and printer plots of Mach number and density contour maps. The TAIR program is written in FORTRAN IV for batch execution and has been implemented on a CDC 7600 computer with a central memory requirement of approximately 155K (octal) of 60 bit words. The TAIR program was developed in 1981.

Description: PLAN2D is a FORTRAN computer program for the plastic analysis of planar rigid frame structures. Given a structure and loading pattern as input, PLAN2D calculates the ultimate load that the structure can sustain before collapse. Element moments and plastic hinge rotations are calculated for the ultimate load. The location of hinges required for a collapse mechanism to form are also determined. The program proceeds in an iterative series of linear elastic analyses. After each iteration the resulting elastic moments in each member are compared to the reserve plastic moment capacity of that member. The member or members that have moments closest to their reserve capacity will determine the minimum load factor and the site where the next hinge is to be inserted. Next, hinges are inserted and the structural stiffness matrix is reformulated. This cycle is repeated until the structure becomes unstable. At this point the ultimate collapse load is calculated by accumulating the minimum load factor from each previous iteration and multiplying them by the original input loads. PLAN2D is based on the program STAN, originally written by Dr. E.L. Wilson at U.C. Berkeley. PLAN2D has several limitations: 1) Although PLAN2D will detect unloading of hinges it does not contain the capability to remove hinges; 2) PLAN2D does not allow the user to input different positive and negative moment capacities and 3) PLAN2D does not consider the interaction between axial and plastic moment capacity. Axial yielding and buckling is ignored as is the reduction in moment capacity due to axial load. PLAN2D is written in FORTRAN and is machine independent. It has been tested on an IBM PC and a DEC MicroVAX. The program was developed in 1988.

Description: Predictions of fatigue crack growth behavior can be made with the Fatigue Crack Growth Structural Analysis (FASTRAN II) computer program. As cyclic loads are applied to a selected crack configuration with an initial crack size, FASTRAN II predicts crack growth as a function of cyclic load history until either a desired crack size is reached or failure occurs. FASTRAN II is based on plasticity-induced crack-closure behavior of cracks in metallic materials and accounts for load-interaction effects, such as retardation and acceleration, under variable-amplitude loading. The closure model is based on the Dugdale model with modifications to allow plastically deformed material to be left along the crack surfaces as the crack grows. Plane stress and plane strain conditions, as well as conditions between these two, can be simulated in FASTRAN II by using a constraint factor on tensile yielding at the crack front to approximately account for three-dimensional stress states. FASTRAN II contains seventeen predefined crack configurations (standard laboratory fatigue crack growth rate specimens and many common crack configurations found in structures); and the user can define one additional crack configuration. The baseline crack growth rate properties (effective stress-intensity factor against crack growth rate) may be given in either equation or tabular form. For three-dimensional crack configurations, such as surface cracks or corner cracks at holes or notches, the fatigue crack growth rate properties may be different in the crack depth and crack length directions. Final failure of the cracked structure can be modelled with fracture toughness properties using either linear-elastic fracture mechanics (brittle materials), a two-parameter fracture criterion (brittle to ductile materials), or plastic collapse (extremely ductile materials). The crack configurations in FASTRAN II can be subjected to either constant-amplitude, variable-amplitude or spectrum loading. The applied loads may be either tensile or compressive. Several standardized aircraft flight-load histories, such as TWIST, Mini-TWIST, FALSTAFF, Inverted FALSTAFF, Felix and Gaussian, are included as options. FASTRAN II also includes two other methods that will help the user input spectrum load histories. The two methods are: (1) a list of stress points, and (2) a flight-by-flight history of stress points. Examples are provided in the user manual. Developed as a research program, FASTRAN II has successfully predicted crack growth in many metallic materials under various aircraft spectrum loading. A computer program DKEFF which is a part of the FASTRAN II package was also developed to analyze crack growth rate data from laboratory specimens to obtain the effective stress-intensity factor against crack growth rate relations used in FASTRAN II. FASTRAN II is written in standard FORTRAN 77. It has been successfully compiled and implemented on Sun4 series computers running SunOS and on IBM PC compatibles running MS-DOS using the Lahey F77L FORTRAN compiler. Sample input and output data are included with the FASTRAN II package. The UNIX version requires 660K of RAM for execution. The standard distribution medium for the UNIX version (LAR-14865) is a .25 inch streaming magnetic tape cartridge in UNIX tar format. It is also available on a 3.5 inch diskette in UNIX tar format. The standard distribution medium for the MS-DOS version (LAR-14944) is a 5.25 inch 360K MS-DOS format diskette. The contents of the diskette are compressed using the PKWARE archiving tools. The utility to unarchive the files, PKUNZIP.EXE, is included. The program was developed in 1984 and revised in 1992. Sun4 and SunOS are trademarks of Sun Microsystems, Inc. IBM PC is a trademark of International Business Machines Corp. MS-DOS is a trademark of Microsoft, Inc. F77L is a trademark of the Lahey Computer Systems, Inc. UNIX is a registered trademark of AT&T Bell Laboratories. PKWARE and PKUNZIP are trademarks of PKWare, Inc.

Description: The ability to treat arbitrary boundary shapes is one of the most desirable characteristics of a method for generating grids. 3DGRAPE is designed to make computational grids in or about almost any shape. These grids are generated by the solution of Poisson's differential equations in three dimensions. The program automatically finds its own values for inhomogeneous terms which give near-orthogonality and controlled grid cell height at boundaries. Grids generated by 3DGRAPE have been applied to both viscous and inviscid aerodynamic problems, and to problems in other fluid-dynamic areas. 3DGRAPE uses zones to solve the problem of warping one cube into the physical domain in real-world computational fluid dynamics problems. In a zonal approach, a physical domain is divided into regions, each of which maps into its own computational cube. It is believed that even the most complicated physical region can be divided into zones, and since it is possible to warp a cube into each zone, a grid generator which is oriented to zones and allows communication across zonal boundaries (where appropriate) solves the problem of topological complexity. 3DGRAPE expects to read in already-distributed x,y,z coordinates on the bodies of interest, coordinates which will remain fixed during the entire grid-generation process. The 3DGRAPE code makes no attempt to fit given body shapes and redistribute points thereon. Body-fitting is a formidable problem in itself. The user must either be working with some simple analytical body shape, upon which a simple analytical distribution can be easily effected, or must have available some sophisticated stand-alone body-fitting software. 3DGRAPE does not require the user to supply the block-to-block boundaries nor the shapes of the distribution of points. 3DGRAPE will typically supply those block-to-block boundaries simply as surfaces in the elliptic grid. Thus at block-to-block boundaries the following conditions are obtained: (1) grids lines will match up as they approach the block-to-block boundary from either side, (2) grid lines will cross the boundary with no slope discontinuity, (3) the spacing of points along the line piercing the boundary will be continuous, (4) the shape of the boundary will be consistent with the surrounding grid, and (5) the distribution of points on the boundary will be reasonable in view of the surrounding grid. 3DGRAPE offers a powerful building-block approach to complex 3-D grid generation, but is a low-level tool. Users may build each face of each block as they wish, from a wide variety of resources. 3DGRAPE uses point-successive-over-relaxation (point-SOR) to solve the Poisson equations. This method is slow, although it does vectorize nicely. Any number of sophisticated graphics programs may be used on the stored output file of 3DGRAPE though it lacks interactive graphics. Versatility was a prominent consideration in developing the code. The block structure allows a great latitude in the problems it can treat. As the acronym implies, this program should be able to handle just about any physical region into which a computational cube or cubes can be warped. 3DGRAPE was written in FORTRAN 77 and should be machine independent. It was originally developed on a Cray under COS and tested on a MicroVAX 3200 under VMS 5.1.

Description: The TAWFIVE program calculates transonic flow over a transport-type wing and fuselage. Although more complex Euler and Navier-Stokes methods are available, TAWFIVE combines a multi-grid acceleration technique in the iterative solution of the potential equation with the use of integral-form boundary-layer equations to provide a computationally efficient and sufficiently accurate design tool. TAWFIVE simplifies the solution process by breaking the problem into a loosely coupled set of modified equations. The inviscid method, using standard inviscid equations (nonlinear full potential), is valid in the "outer" region away from the wing, whereas the boundary-layer equations are valid in the thin region near the solid surface of the wing. The two types of equations are coupled by a technique of modifying surface boundary conditions for the inviscid equations. This interaction process starts with a solution of the outer flow field. Pressures are computed at the wing surface and are used to calculate the boundary layer. The boundary-layer and wake properties are then computed using a three-dimensional integral method, and the computed displacement thickness is added to the surface of the "hard" geometry. This new displaced wing surface is then regridded and the inviscid flowfield is recomputed. New values of the inviscid pressures are then used by the boundary-layer method to predict a new displacement thickness distribution. An under-relaxed update of the previously predicted displacement thickness is then made to obtain a new displacement thickness correction that is added to the "hard" geometry. These global iterations are continued until suitable convergence is obtained. Input to TAWFIVE is limited to geometric definition of the configuration, free-stream flow quantities, and iteration control parameters. The geometric input consists of the definition of a series of airfoil sections to define the wing and a series of fuselage cross sections to model the fuselage. High-aspect-ratio wings are modeled more accurately than low-aspect-ratio wings since no special provisions are made to accurately model the wing-fuselage juncture or the wingtip region. The user can specify the solution either in terms of lift or in terms of angle of attack. TAWFIVE can produce tabular output and input files for PLOT3D (COSMIC program number ARC-12779). TAWFIVE is written in FORTRAN 77 for CRAY series computers running UNICOS. The main memory requirement is 2.7Mb for execution. This program is available on a 9-track 1600 BPI UNIX tar format magnetic tape. TAWFIVE was under development from 1979 to 1989 and first released by COSMIC in 1991. CRAY and UNICOS are registered trademarks of Cray Research, Inc.

Description: This computer program is designed to calculate the flow fields in two-dimensional and three-dimensional axisymmetric supersonic inlets. The method of characteristics is used to compute arrays of points in the flow field. At each point the total pressure, local Mach number, local flow angle, and static pressure are calculated. This program can be used to design and analyze supersonic inlets by determining the surface compression rates and throat flow properties. The program employs the method of characteristics for a perfect gas. The basic equation used in the program is the compatibility equation which relates the change in stream angle to the change in entropy and the change in velocity. In order to facilitate the computation, the flow field behind the bow shock wave is broken into regions bounded by shock waves. In each region successive rays are computed from a surface to a shock wave until the shock wave intersects a surface or falls outside the cowl lip. As soon as the intersection occurs a new region is started and the previous region continued only in the area in which it is needed, thus eliminating unnecessary calculations. The maximum number of regions possible in the program is ten, which allows for the simultaneous calculations of up to nine shock waves. Input to this program consists of surface contours, free-stream Mach number, and various calculation control parameters. Output consists of printed and/or plotted results. For plotted results an SC-4020 or similar plotting device is required. This program is written in FORTRAN IV to be executed in the batch mode and has been implemented on a CDC 7600 with a central memory requirement of approximately 27k (octal) of 60 bit words.

Description: The SNIP program is a FORTRAN computer code that generates NASTRAN structural model thermal loads when given SINDA (or similar thermal model) temperature results. SNIP correlates thermal nodes to structural elements to interface SINDA finite difference thermal models with NASTRAN finite element structural models. Node-to-element correlation includes determining which SINDA nodes should be related to each NASTRAN element and calculating a weighing factor for temperatures associated with each element-related thermal node. SNIP provides structural model thermal loads that accurately reflect thermal model results while reducing the time required to interface thermal and structural models as compared to other methods. SNIP uses thermal model geometry to search the three-dimensional space around each structural element for the nearest thermal nodes. Thermal model geometry is the combination of standard thermal model temperature results from SINDA and structural model geometry from NASTRAN. Thermal and structural models must both be defined in the same, single Cartesian coordinate system. The thermal nodes located nearest each element are used to determine element temperature for thermal distortion and stress analysis. The program shapes the three-dimensional search region while the user controls the size. With these region specifications, the numerical coding of thermal nodes, and the structural element numbers; the code can provide for the separation of substructures during correlation. The input to SNIP contains a file of thermal model temperature results and a physical location of each thermal node in three-dimensional space, combined in a SNIP-unique format. The input also contains a standard NASTRAN input deck for a model made up of plate, shell, beam, and bar elements. SNIP supports the CTRIA, CQUAD, CBAR, and CBEAM elements of NASTRAN. The user adjusts the input parameters in the source code which control the node-to-element correlation. The program outputs NASTRAN element temperature load cards for each element and NASTRAN case control cards for each temperature load set. SNIP also outputs a list of elements that contains the numbers of the SINDA nodes related to each NASTRAN element and the weight that is given to each node in temperature calculations. SNIP is written in ANSI standard FORTRAN 77. The PC version requires a PC FORTRAN compiler and has compiled successfully using Lahey FORTRAN v. 3.0. A core memory of 300k is recommended. The program was developed in 1987.

Description: This program was developed to predict turbine stage performance taking into account the effects of complex passage geometries. The method uses a quasi-3D inviscid-flow analysis iteratively coupled to calculated losses so that changes in losses result in changes in the flow distribution. In this manner the effects of both the geometry on the flow distribution and the flow distribution on losses are accounted for. The flow may be subsonic or shock-free transonic. The blade row may be fixed or rotating, and the blades may be twisted and leaned. This program has been applied to axial and radial turbines, and is helpful in the analysis of mixed flow machines. This program is a combination of the flow analysis programs MERIDL and TSONIC coupled to the boundary layer program BLAYER. The subsonic flow solution is obtained by a finite difference, stream function analysis. Transonic blade-to-blade solutions are obtained using information from the finite difference, stream function solution with a reduced flow factor. Upstream and downstream flow variables may vary from hub to shroud and provision is made to correct for loss of stagnation pressure. Boundary layer analyses are made to determine profile and end-wall friction losses. Empirical loss models are used to account for incidence, secondary flow, disc windage, and clearance losses. The total losses are then used to calculate stator, rotor, and stage efficiency. This program is written in FORTRAN IV for batch execution and has been implemented on an IBM 370/3033 under TSS with a central memory requirement of approximately 4.5 Megs of 8 bit bytes. This program was developed in 1985.

Description: This algorithm has been developed for calculating both the quantity of compressor bleed flow required to cool a turbine and the resulting decrease in efficiency due to cooling air injected into the gas stream. Because of the trend toward higher turbine inlet temperatures, it is important to accurately predict the required cooling flow. This program is intended for use with axial flow, air-breathing jet propulsion engines with a variety of airfoil cooling configurations. The algorithm results have compared extremely well with figures given by major engine manufacturers for given bulk metal temperatures and cooling configurations. The program calculates the required cooling flow and corresponding decrease in stage efficiency for each row of airfoils throughout the turbine. These values are combined with the thermodynamic efficiency of the uncooled turbine to predict the total bleed airflow required and the altered turbine efficiency. There are ten airfoil cooling configurations and the algorithm allows a different option for each row of cooled airfoils. Materials technology is incorporated and requires the date of the first year of service for the turbine stator vane and rotor blade. The user must specify pressure, temperatures, and gas flows into the turbine. This program is written in FORTRAN IV for batch execution and has been implemented on an IBM 3080 series computer with a central memory requirement of approximately 61K of 8 bit bytes. This program was developed in 1980.

Description: Turbomachinery components are often connected by ducts, which are usually annular. The configurations and aerodynamic characteristics of these ducts are crucial to the optimum performance of the turbomachinery blade rows. The ANDUCT computer program was developed to calculate the velocity distribution along an arbitrary line between the inner and outer walls of an annular duct with axisymmetric swirling flow. Although other programs are available for duct analysis, the use of the velocity gradient method makes the ANDUCT program fast and convenient while requiring only modest computer resources. A fast and easy method of analyzing the flow through a duct with axisymmetric flow is the velocity gradient method, also known as the stream filament or streamline curvature method. This method has been used extensively for blade passages but has not been widely used for ducts, except for the radial equilibrium equation. In ANDUCT, a velocity gradient equation derived from the momentum equation is used to determine the velocity variation along an arbitrary straight line between the inner and outer wall of an annular duct. The velocity gradient equation is used with an assumed variation of meridional streamline curvature. Upstream flow conditions may vary between the inner and outer walls, and an assumed total pressure distribution may be specified. ANDUCT works best for well-guided passages and where the curvature of the walls is small as compared to the width of the passage. The ANDUCT program is written in FORTRAN IV for batch execution and has been implemented on an IBM 370 series computer with a central memory requirement of approximately 60K of 8 bit bytes. The ANDUCT program was developed in 1982.

Description: The Panel Code for Planar Cascades was developed as an aid for the designer of turbomachinery blade rows. The effective design of turbomachinery blade rows relies on the use of computer codes to model the flow on blade-to-blade surfaces. Most of the currently used codes model the flow as inviscid, irrotational, and compressible with solutions being obtained by finite difference or finite element numerical techniques. While these codes can yield very accurate solutions, they usually require an experienced user to manipulate input data and control parameters. Also, they often limit a designer in the types of blade geometries, cascade configurations, and flow conditions that can be considered. The Panel Code for Planar Cascades accelerates the design process and gives the designer more freedom in developing blade shapes by offering a simple blade-to-blade flow code. Panel, or integral equation, solution techniques have been used for several years by external aerodynamicists who have developed and refined them into a primary design tool of the aircraft industry. The Panel Code for Planar Cascades adapts these same techniques to provide a versatile, stable, and efficient calculation scheme for internal flow. The code calculates the compressible, inviscid, irrotational flow through a planar cascade of arbitrary blade shapes. Since the panel solution technique is for incompressible flow, a compressibility correction is introduced to account for compressible flow effects. The analysis is limited to flow conditions in the subsonic and shock-free transonic range. Input to the code consists of inlet flow conditions, blade geometry data, and simple control parameters. Output includes flow parameters at selected control points. This program is written in FORTRAN IV for batch execution and has been implemented on an IBM 370 series computer with a central memory requirement of approximately 590K of 8 bit bytes. This program was developed in 1982.

Description: An exact, full-potential-equation model for the steady, irrotational, homoentropic, and homoenergetic flow of a compressible, inviscid fluid through a two-dimensional planar cascade together with its appropriate boundary conditions has been derived. The CAS2D computer program numerically solves an artificially time-dependent form of the actual full-potential-equation, providing a nonrotating blade-to-blade, steady, potential transonic cascade flow analysis code. Comparisons of results with test data and theoretical solutions indicate very good agreement. In CAS2D, the governing equation is discretized by using type-dependent, rotated finite differencing and the finite area technique. The flow field is discretized by providing a boundary-fitted, nonuniform computational mesh. This mesh is generated by using a sequence of conformal mapping, nonorthogonal coordinate stretching, and local, isoparametric, bilinear mapping functions. The discretized form of the full-potential equation is solved iteratively by using successive line over relaxation. Possible isentropic shocks are captured by the explicit addition of an artificial viscosity in a conservative form. In addition, a four-level, consecutive, mesh refinement feature makes CAS2D a reliable and fast algorithm for the analysis of transonic, two-dimensional cascade flows. The results from CAS2D are not directly applicable to three-dimensional, potential, rotating flows through a cascade of blades because CAS2D does not consider the effects of the Coriolis force that would be present in the three-dimensional case. This program is written in FORTRAN IV for batch execution and has been implemented on an IBM 370 series computer with a central memory requirement of approximately 200K of 8 bit bytes. The CAS2D program was developed in 1980.

Description: A fast algorithm has been developed for accurately generating boundary-conforming, three-dimensional consecutively refined computational grids applicable to arbitrary wing-body and axial turbomachinery geometries. This algorithm has been incorporated into the GRID3O computer program. The method employed in GRID3O is based on using an analytic function to generate two-dimensional grids on a number of coaxial axisymmetric surfaces positioned between the centerbody and the outer radial boundary. These grids are of the O-type and are characterized by quasi-orthogonality, geometric periodicity, and an adequate resolution throughout the flow field. Because the built-in nonorthogonal coordinate stretching and shearing cause the grid lines leaving the blade or wing trailing-edge to end at downstream infinity, use of the generated grid simplifies the numerical treatment of three-dimensional trailing vortex sheets. The GRID3O program is written in FORTRAN IV for batch execution and has been implemented on an IBM 370 series computer with a central memory requirement of approximately 450K of 8 bit bytes. The GRID3O program was developed in 1981.

Description: A computer program, QSONIC, has been developed for calculating the full potential, transonic quasi-three-dimensional flow through a rotating turbomachinery blade row. The need for lighter, more efficient turbomachinery components has led to the consideration of machines with fewer stages, each with blades capable of higher speeds and higher loading. As speeds increase, the numerical problems inherent in the transonic regime have to be resolved. These problems include the calculation of imbedded shock discontinuities and the dual nature of the governing equations, which are elliptic in the subcritical flow regions but become hyperbolic for supersonic zones. QSONIC provides the flow analyst with a fast and reliable means of obtaining the transonic potential flow distribution on a blade-to-blade stream surface of a stationary or rotating turbomachine blade row. QSONIC combines several promising transonic analysis techniques. The full potential equation in conservative form is discretized at each point on a body-fitted period mesh. A mass balance is calculated through the finite volume surrounding each point. Each local volume is corrected in the third dimension for any change in stream-tube thickness along the stream tube. The nonlinear equations for all volumes are of mixed type (elliptic or hyperbolic) depending on the local Mach number. The final result is a block-tridiagonal matrix formulation involving potential corrections at each grid point as the unknowns. The residual of each system of equations is solved along each grid line. At points where the Mach number exceeds unity, the density at the forward (sweeping) edge of the volume is replaced by an artificial density. This method calculates the flow field about a cascade of arbitrary two-dimensional airfoils. Three-dimensional flow is approximated in a turbomachinery blade row by correcting for stream-tube convergence and radius change in the through flow direction. Several significant assumptions were made in developing the QSONIC program, including: (1) the flow is inviscid and adiabatic, (2) the flow relative to the blade is steady, (3) the fluid is a perfect gas with constant specific heat, (4) the flow is isentropic and any discontinuities (shocks) are weak enough to be approximated as isentropic jumps, (5) there is no velocity component normal to the stream surface, and (6) the flow relative to a fixed frame in space (absolute velocity) is completely irrotational. These assumptions place some limitations on the application of QSONIC. Sharp leading edges at high incidence and high-Mach-number turbine blade trailing edges with substantial deviation will both cause large velocity peaks on the blade. In addition, the program may have difficulty converging if the passage is nearly choked. Input to QSONIC consists of case control parameters, a geometry description, upstream boundary conditions, and a rotor description. Output includes solution scheme parameters and flow field parameters. A data file is also output which contains data on the solution mesh, surface Mach numbers, surface static pressures, isomachs, and the velocity vector field. This data may be used for further processing or for plotting. The QSONIC is written in FORTRAN IV for batch execution and has been implemented on an IBM 370 series computer with a central memory requirement of approximately 500K of 8 bit bytes. QSONIC was developed in 1982.

Description: This computer program, WIND, was developed to numerically solve the exact, full-potential equation for three-dimensional, steady, inviscid flow through an isolated wind turbine rotor. The program automatically generates a three-dimensional, boundary-conforming grid and iteratively solves the full-potential equation while fully accounting for both the rotating and Coriolis effects. WIND is capable of numerically analyzing the flow field about a given blade shape of the horizontal-axis type wind turbine. The rotor hub is assumed representable by a doubly infinite circular cylinder. An arbitrary number of blades may be attached to the hub and these blades may have arbitrary spanwise distributions of taper and of the twist, sweep, and dihedral angles. An arbitrary number of different airfoil section shapes may be used along the span as long as the spanwise variation of all the geometeric parameters is reasonably smooth. The numerical techniques employed in WIND involve rotated, type-dependent finite differencing, a finite volume method, artificial viscosity in conservative form, and a successive overrelaxation combined with the sequential grid refinement procedure to accelerate the iterative convergence rate. Consequently, WIND is cabable of accurately analyzing incompressible and compressible flows, including those that are locally transonic and terminated by weak shocks. Along with the three-dimensional results, WIND provides the results of the two-dimensional calculations to aid the user in locating areas of possible improvement in the aerodynamic design of the blade. Output from WIND includes the chordwise distribution of the coefficient of pressure, the Mach number, the density, and the relative velocity components at spanwise stations along the blade. In addition, the results specify local values of the lift coefficient and the tangent and axial aerodynamic force components. These are also given in integrated form expressing the total torque and the total axial force acting on the shaft. WIND can also be used to analyze the flow around isolated aircraft propellers and helicopter rotors in hover as long as the relative oncoming flow is subsonic. The WIND program is written in FORTRAN IV for batch execution and has been implemented on an IBM 370 series computer with a central memory requirement of approximately 253K of 8 bit bytes. WIND was developed in 1980.

Description: This computer program calculates the flow field in the supersonic portion of a mixed-compression aircraft inlet at non-zero angle of attack. This approach is based on the method of characteristics for steady three-dimensional flow. The results of this program agree with those produced by the two-dimensional method of characteristics when axisymmetric flow fields are calculated. Except in regions of high viscous interaction and boundary layer removal, the results agree well with experimental data obtained for threedimensional flow fields. The flow field in a variety of axisymmetric mixed compression inlets can be calculated using this program. The bow shock wave and the internal shock wave system are calculated using a discrete shock wave fitting procedure. The internal flow field can be calculated either with or without the discrete fitting of the internal shock wave system. The influence of molecular transport can be included in the calculation of the external flow about the forebody and in the calculation of the internal flow when internal shock waves are not discretely fitted. The viscous and thermal diffussion effects are included by treating them as correction terms in the method of characteristics procedure. Dynamic viscosity is represented by Sutherland's law and thermal conductivity is represented as a quadratic function of temperature. The thermodynamic model used is that of a thermally and calorically perfect gas. The program assumes that the cowl lip is contained in a constant plane and that the centerbody contour and cowl contour are smooth and have continuous first partial derivatives. This program cannot calculate subsonic flow, the external flow field if the bow shock wave does not exist entirely around the forebody, or the internal flow field if the bow flow field is injected into the annulus. Input to the program consists of parameters to control execution, to define the geometry, and the vehicle orientation. Output consists of a list of parameters used, solution planes, and a description of the shock waves. This program is written in FORTRAN IV for batch execution and has been implemented on a CDC 6000 series machine with a central memory requirement of 110K (octal) of 60 bit words when it is overlayed. This flow analysis program was developed in 1978.

Description: As turbine-engine core operating conditions become more severe, designers must develop more effective means of cooling blades and vanes. In order to design reliable, cooled turbine blades, advanced transient thermal calculation techniques are required. The TACT1 computer program was developed to perform transient and steady-state heat-transfer and coolant-flow analyses for cooled blades, given the outside hot-gas boundary condition, the coolant inlet conditions, the geometry of the blade shell, and the cooling configuration. TACT1 can analyze turbine blades, or vanes, equipped with a central coolant-plenum insert from which coolant-air impinges on the inner surface of the blade shell. Coolant-side heat-transfer coefficients are calculated with the heat transfer mode at each station being user specified as either impingement with crossflow, forced convection channel flow, or forced convection over pin fins. A limited capability to handle film cooling is also available in the program. The TACT1 program solves for the blade temperature distribution using a transient energy equation for each node. The nodal energy balances are linearized, one-dimensional, heat-conduction equations which are applied at the wall-outer-surface node, at the junction of the cladding and the metal node, and at the wall-inner-surface node. At the mid-metal node a linear, three-dimensional, heat-conduction equation is used. Similarly, the coolant pressure distribution is determined by solving the set of transfer momentum equations for the one-dimensional flow between adjacent fluid nodes. In the coolant channel, energy and momentum equations for one-dimensional compressible flow, including friction and heat transfer, are used for the elemental channel length between two coolant nodes. The TACT1 program first obtains a steady-state solution using iterative calculations to obtain convergence of stable temperatures, pressures, coolant-flow split, and overall coolant mass balance. Transient calculations are based on the steady-state solutions obtained. Input to the TACT1 program includes a geometrical description of the blade and insert, the nodal spacing to be used, and the boundary conditions describing the outside hot-gas and the coolant-inlet conditions. The program output includes the value of nodal temperatures and pressures at each iteration. The final solution output includes the temperature at each coolant node, and the coolant flow rates and Reynolds numbers. This program is written in FORTRAN IV for batch execution and has been implemented on an IBM 360 computer with a central memory requirement of approximately 480K of 8 bit bytes. The TACT1 program was developed in 1978.

Description: This computer program was developed for calculating the subsonic or transonic flow on the hub-shroud mid-channel stream surface of a single blade row of a turbomachine. The design and analysis of blades for compressors and turbines ideally requires methods for analyzing unsteady, three-dimensional, turbulent viscous flow through a turbomachine. Since an exact solution is impossible at present, solutions on two-dimensional surfaces are calculated to obtain a quasi-three dimensional solution. When three-dimensional effects are important, significant information can be obtained from a solution on a cross-sectional surface of the passage normal to the flow. With this program, a solution to the equations of flow on the meridional surface can be carried out. This solution is chosen when the turbomachine under consideration has significant variation in flow properties in the hubshroud direction, especially when input is needed for use in blade-to-blade calculations. The program can also perform flow calculations for annular ducts without blades. This program should prove very useful in the design and analysis of any turbomachine. This program calculates a solution for two-dimensional, adiabatic shockfree flow. The flow must be essentially subsonic, but there may be local areas of supersonic flow. To obtain the solution, this program uses both the finite difference and the quasi-orthogonal (velocity gradient) methods combined in a way that takes maximum advantage of both. The finite-difference method solves a finite-difference equation along the meridional stream surface in a very efficient manner but is limited to subsonic velocities. This approach must be used in cases where the blade aspect ratios are above one, cases where the passage is curved, and cases with low hub-tip-ratio blades. The quasi-orthogonal method solves the velocity gradient equation on the meridional surface and is used if it is necessary to extend the range of solutions into the transonic regime. In general the blade row may be fixed or rotating and the blades may be twisted and leaned. The flow may be axial, radial, or mixed. The upstream and downstream flow conditions can vary from hub to shroud with provisions made for an approximate correction for loss of stagnation pressure. Also, viscous forces are neglected along solution mesh lines running from hub to tip. The capabilities of this program include handling of nonaxial flows without restriction, annular ducts without blades, and specified streamwise loss distributions. This program is written in FORTRAN IV for batch execution and has been implemented on an IBM 360 computer with a central memory requirement of approximately 700K of 8 bit bytes. This core requirement can be reduced depending on the size of the problem and the desired solution accuracy. This program was developed in 1977.

Description: This is a finite-difference program for calculating the viscous compressible boundary layer flow over either planar or axisymmetric surfaces. The flow may be initially laminar and progress through a transitional zone to a fully turbulent flow, or it may remain laminar, depending on the imposed boundary conditions, laws of viscosity, and numerical solution of the momentum and energy equations. The flow may also be forced into a turbulent flow at a chosen spot by the data input. The input may contain factors of arbitrary Reynolds number, free-stream Mach number, free stream turbulence, wall heating or cooling, longitudinal wall curvature, wall suction or blowing, and wall roughness. The solution may start from an initial Falkner-Skan similarity profile, an approximate equilibrium turbulent profile, or an initial arbitrary input profile. This program has been implemented on the IBM 7094/7044 Direct Couple System. This program is written in FORTRAN IV and was developed in 1974.

Description: A computer program has been developed for the design of supersonic rotor blades where losses are accounted for by correcting the ideal blade geometry for boundary layer displacement thickness. The ideal blade passage is designed by the method of characteristics and is based on establishing vortex flow within the passage. Boundary-layer parameters (displacement and momentum thicknesses) are calculated for the ideal passage, and the final blade geometry is obtained by adding the displacement thicknesses to the ideal nozzle coordinates. The boundary-layer parameters are also used to calculate the aftermixing conditions downstream of the rotor blades assuming the flow mixes to a uniform state. The computer program input consists essentially of the rotor inlet and outlet Mach numbers, upper- and lower-surface Mach numbers, inlet flow angle, specific heat ratio, and total flow conditions. The program gas properties are set up for air. Additional gases require changes to be made to the program. The computer output consists of the corrected rotor blade coordinates, the principal boundary-layer parameters, and the aftermixing conditions. This program is written in FORTRAN IV for batch execution and has been implemented on an IBM 7094. This program was developed in 1971.

Description: A computer program has been developed for the design of sharp-edged throat supersonic nozzles where losses are accounted for by correcting the ideal nozzle geometry for boundary layer displacement thickness. The ideal nozzle is designed by the method of characteristics to produce uniform parallel flow at the nozzle exit in the smallest possible distance. Boundary-layer parameters (displacement and momentum thicknesses) are calculated for the ideal nozzle, and the final nozzle geometry is obtained by adding the displacement thickness to the ideal nozzle coordinates. The boundary layer parameters are also used to calculate the aftermixing conditions downstream of the nozzle assuming the flow mixes to a uniform state. The computer program input consists essentially of the nozzle-exit Mach number, specific-heat ratio, nozzle angle, throat half-height, nozzle subsonic section coordinates and corresponding pressure ratios, total temperature and pressure, gas constant, and initial momentum or displacement thickness. The program gas properties are set up for air; for other gases, changes are required to the program. The computer program output consists of the corrected nozzle coordinates, the principal boundary-layer parameters, and the aftermixing conditions. This program has been implemented on the IBM 7094/7044 Direct Couple System.

Description: A computer program has been developed which analyzes by means of mathematical models the temperature profiles in the contents of a filled propellant tank. In designing space vehicles using cryogenic liquid propellants, it is necessary to know how heat transferred from the tank walls and heat absorbed internally affect the temperature distribution with the tank contents. The mathematical flow model is based on results from small-scale experiments. The results showed that when a subcooled fluid is subject to both nonuniform internal heating and wall heating, two distinct temperature regions are developed. In the lower region, the fluid is thoroughly mixed and maintains a uniform temperature profile. In the upper region, a stratified layer develops, and a temperature gradient is formed from the accumulation of warm fluid from the boundary layer along the tank walls; it also indicated that the temperature profiles in the stratified layer exhibited similarity. This concept was developed primarily for internal heating caused by nuclear radiation. However, the theory and computer program are applicable for any form of internal or bulk heating. This program is written in FORTRAN IV for batch execution and has been implemented on the IBM 7094. This program was developed in 1970.

Description: This program solves the two-dimensional, compressible laminar or turbulent boundary-layer equations in an arbitrary pressure gradient. Cohen and Reshotko's method is used for the laminar boundary layer, Sasman and Cresci's method for the turbulent boundary layer, and the Schlichting-Ulrich-Granville method to predict transition. Transition may also be forced at any point by the user. Separation, if it occurs, is predicted for both laminar and turbulent flow. The user may begin values for displacement thickness and momentum thickness in either laminar or turbulent flow. This program was implemented on the IBM 7094.

Description: This program obtains a transonic flow solution on a blade-to-blade surface between blades of a turbomachine. The flow must be essentially subsonic, but there may be locally supersonic flow. The solution is two-dimensional, isentropic, and shock free. The blades may be fixed or rotating. The flow may be axial, radial, or mixed, and there may be a change in stream-channel thickness in the through-flow direction. A loss in relative stagnation pressure may be accounted for. The program input consists of blade and stream-channel geometry, stagnation flow conditions, inlet and outlet flow angles, and blade-to-blade stream-channel weight flow. The output includes blade surface velocities, velocity magnitude and direction at all interior mesh points in the blade-to-blade passage, and streamline coordinates throughout the passage. The transonic solution is obtained by a combination of a finite-difference, stream-function solution and a velocity-gradient solution. The finite-difference solution at a reduced weight flow provides information needed to obtain a velocity-gradient solution. This program is written in FORTRAN IV for batch execution and has been implemented on the IBM 360 computer with a central memory requirement of approximately 36K of 8 bit bytes. This program was developed in 1969 and last updated in 1979.

Description: This program is a revision of an existing program for blade-to-blade aerodynamic analysis of turbomachine blades and it is a simpler program while consistent with related programs. The analysis is for two-dimensional, subsonic, compressible (or incompressible), nonviscous flow in a circular or straight infinite cascade of blades, which may be fixed or rotating. The flow may be axial, radial, or mixed, and the stream channel thickness may change in the through-flow direction. The program input consists of blade and stream channel geometry, total flow conditions, inlet and outlet flow angles, and blade-to-blade stream channel weight flow. The output includes blade surface velocities, velocity magnitude and direction at all interior mesh points in the blade-to-blade passage, and streamline coordinates throughout the passage. This program was developed on an IBM 7094/7044 DCS.

Description: This computer program gives the blade-to-blade solution of the two-dimensional, subsonic, compressible (or incompressible), nonviscous flow problem for a circular or straight infinite cascade of tandem or slotted turbomachine blades. The blades may be fixed or rotating. The flow may be axial, radial , or mixed. The method of solution is based on the stream function using an iterative solution of nonlinear finite-difference equations. These equations are solved using two major levels of iteration. The inner iteration consists of the solution of simultaneous linear equations by successive over-relaxation, using an estimated optimum over-relaxation factor. The outer iteration then changes the coefficients of the simultaneous equations to correct for compressibility. The program input consists of the basic blade geometry, the meridional stream channel coordinates, fluid stagnation conditions, weight flow and flow split through the slot, and inlet and outlet flow angles. The output includes blade surface velocities, velocity magnitude and direction throughout the passage, and the streamline coordinates.

Description: This FORTRAN IV computer program which incorporates the method of characteristics was written to assist in the design of supersonic inlets. There were two objectives: (1) to study a greater variety of supersonic inlet configurations and (2) to reduce the time required for trial-and-error procedures to arrive at optimum inlet design. The computer program was written with the intention of being able to construct a variety of inlet configurations by interchanging specific subroutines. In this manner, greater flexibility of choice was attained, and the time required to program a specific inlet configuration was greatly reduced. The second objective was accomplished by a reformulation of the boundary value problem for hyperbolic equations. By this reformulation of the boundary data, the engineering design quantities, throat Mach number and flow angle, were introduced as direct input quantities to the computer program. As a consequence of introducing the engineering parameters as input, the computer program will calculate the surface contours required to satisfy the specific throat conditions. Inviscid flow is assumed and the method used to calculate the inlet contour results in minimum distortion to the flow in the throat. This program was developed on an IBM 7094.

Description: This program represents a subsonic aerodynamic method for determining the mean camber surface of trimmed noncoplaner planforms with minimum vortex drag. With this program, multiple surfaces can be designed together to yield a trimmed configuration with minimum induced drag at some specified lift coefficient. The method uses a vortex-lattice and overcomes previous difficulties with chord loading specification. A Trefftz plane analysis is used to determine the optimum span loading for minimum drag. The program then solves for the mean camber surface of the wing associated with this loading. Pitching-moment or root-bending-moment constraints can be employed at the design lift coefficient. Sensitivity studies of vortex-lattice arrangements have been made with this program and comparisons with other theories show generally good agreement. The program is very versatile and has been applied to isolated wings, wing-canard configurations, a tandem wing, and a wing-winglet configuration. The design problem solved with this code is essentially an optimization one. A subsonic vortex-lattice is used to determine the span load distribution(s) on bent lifting line(s) in the Trefftz plane. A Lagrange multiplier technique determines the required loading which is used to calculate the mean camber slopes, which are then integrated to yield the local elevation surface. The problem of determining the necessary circulation matrix is simplified by having the chordwise shape of the bound circulation remain unchanged across each span, though the chordwise shape may vary from one planform to another. The circulation matrix is obtained by calculating the spanwise scaling of the chordwise shapes. A chordwise summation of the lift and pitching-moment is utilized in the Trefftz plane solution on the assumption that the trailing wake does not roll up and that the general configuration has specifiable chord loading shapes. VLMD is written in FORTRAN for IBM PC series and compatible computers running MS-DOS. This program requires 360K of RAM for execution. The Ryan McFarland FORTRAN compiler and PLINK86 are required to recompile the source code; however, a sample executable is provided on the diskette. The standard distribution medium for VLMD is a 5.25 inch 360K MS-DOS format diskette. VLMD was originally developed for use on CDC 6000 series computers in 1976. It was originally ported to the IBM PC in 1986, and, after minor modifications, the IBM PC port was released in 1993.

Description: This code was developed to aid design engineers in the selection and evaluation of aerodynamically efficient wing-canard and wing-horizontal-tail configurations that may employ simple hinged-flap systems. Rapid estimates of the longitudinal aerodynamic characteristics of conceptual airplane lifting surface arrangements are provided. The method is particularly well suited to configurations which, because of high speed flight requirements, must employ thin wings with highly swept leading edges. The code is applicable to wings with either sharp or rounded leading edges. The code provides theoretical pressure distributions over the wing, the canard or horizontal tail, and the deflected flap surfaces as well as estimates of the wing lift, drag, and pitching moments which account for attainable leading edge thrust and leading edge separation vortex forces. The wing planform information is specified by a series of leading edge and trailing edge breakpoints for a right hand wing panel. Up to 21 pairs of coordinates may be used to describe both the leading edge and the trailing edge. The code has been written to accommodate 2000 right hand panel elements, but can easily be modified to accommodate a larger or smaller number of elements depending on the capacity of the target computer platform. The code provides solutions for wing surfaces composed of all possible combinations of leading edge and trailing edge flap settings provided by the original deflection multipliers and by the flap deflection multipliers. Up to 25 pairs of leading edge and trailing edge flap deflection schedules may thus be treated simultaneously. The code also provides for an improved accounting of hinge-line singularities in determination of wing forces and moments. To determine lifting surface perturbation velocity distributions, the code provides for a maximum of 70 iterations. The program is constructed so that successive runs may be made with a given code entry. To make additional runs, it is necessary only to add an identification record and the namelist data that are to be changed from the previous run. This code was originally developed in 1989 in FORTRAN V on a CDC 6000 computer system, and was later ported to an MS-DOS environment. Both versions are available from COSMIC. There are only a few differences between the PC version (LAR-14458) and CDC version (LAR-14178) of AERO2S distributed by COSMIC. The CDC version has one main source code file while the PC version has two files which are easier to edit and compile on a PC. The PC version does not require a FORTRAN compiler which supports NAMELIST because a special INPUT subroutine has been added. The CDC version includes two MODIFY decks which can be used to improve the code and prevent the possibility of some infrequently occurring errors while PC-version users will have to make these code changes manually. The PC version includes an executable which was generated with the Ryan McFarland/FORTRAN compiler and requires 253K RAM and an 80x87 math co-processor. Using this executable, the sample case requires about four hours to execute on an 8MHz AT-class microcomputer with a co-processor. The source code conforms to the FORTRAN 77 standard except that it uses variables longer than six characters. With two minor modifications, the PC version should be portable to any computer with a FORTRAN compiler and sufficient memory. The CDC version of AERO2S is available in CDC NOS Internal format on a 9-track 1600 BPI magnetic tape. The PC version is available on a set of two 5.25 inch 360K MS-DOS format diskettes. IBM AT is a registered trademark of International Business Machines. MS-DOS is a registered trademark of Microsoft Corporation. CDC is a registered trademark of Control Data Corporation. NOS is a trademark of Control Data Corporation.

Description: This program provides a wing design algorithm based on modified linear theory which takes into account the effects of attainable leading-edge thrust. A primary objective of the WINGDES2 approach is the generation of a camber surface as mild as possible to produce drag levels comparable to those attainable with full theoretical leading-edge thrust. WINGDES2 provides both an analysis and a design capability and is applicable to both subsonic and supersonic flow. The optimization can be carried out for designated wing portions such as leading and trailing edge areas for the design of mission-adaptive surfaces, or for an entire planform such as a supersonic transport wing. This program replaces an earlier wing design code, LAR-13315, designated WINGDES. WINGDES2 incorporates modifications to improve numerical accuracy and provides additional capabilities. A means of accounting for the presence of interference pressure fields from airplane components other than the wing and a direct process for selection of flap surfaces to approach the performance levels of the optimized wing surfaces are included. An increased storage capacity allows better numerical representation of those configurations that have small chord leading-edge or trailing-edge design areas. WINGDES2 determines an optimum combination of a series of candidate surfaces rather than the more commonly used candidate loadings. The objective of the design is the recovery of unrealized theoretical leading-edge thrust of the input flat surface by shaping of the design surface to create a distributed thrust and thus minimize drag. The input consists of airfoil section thickness data, leading and trailing edge planform geometry, and operational parameters such as Mach number, Reynolds number, and design lift coefficient. Output includes optimized camber surface ordinates, pressure coefficient distributions, and theoretical aerodynamic characteristics. WINGDES2 is written in FORTRAN V for batch execution and has been implemented on a CDC CYBER computer operating under NOS 2.7.1 with a central memory requirement of approximately 344K (octal) of 60 bit words. This program was developed in 1984, and last updated in 1990. CDC and CYBER are trademarks of Control Data Corporation.

Description: NPLOT is an interactive computer graphics program for plotting undeformed and deformed NASTRAN finite element models (FEMs). Although there are many commercial codes already available for plotting FEMs, these have limited use due to their cost, speed, and lack of features to view BAR elements. NPLOT was specifically developed to overcome these limitations. On a vector type graphics device the two best ways to show depth are by hidden line plotting or haloed line plotting. A hidden line algorithm generates views of models with all hidden lines removed, and a haloed line algorithm displays views with aft lines broken in order to show depth while keeping the entire model visible. A haloed line algorithm is especially useful for plotting models composed of many line elements and few surface elements. The most important feature of NPLOT is its ability to create both hidden line and haloed line views accurately and much more quickly than with any other existing hidden or haloed line algorithms. NPLOT is also capable of plotting a normal wire frame view to display all lines of a model. NPLOT is able to aid in viewing all elements, but it has special features not generally available for plotting BAR elements. These features include plotting of TRUE LENGTH and NORMALIZED offset vectors and orientation vectors. Standard display operations such as rotation and perspective are possible, but different view planes such as X-Y, Y-Z, and X-Z may also be selected. Another display option is the Z-axis cut which allows a portion of the fore part of the model to be cut away to reveal details of the inside of the model. A zoom function is available to terminals with a locator (graphics cursor, joystick, etc.). The user interface of NPLOT is designed to make the program quick and easy to use. A combination of menus and commands with help menus for detailed information about each command allows experienced users greater speed and efficiency. Once a plot is on the screen the interface becomes command driven, enabling the user to manipulate the display or execute a command without having to return to the menu. NPLOT is also able to plot deformed shapes allowing it to perform post-processing. The program can read displacements, either static displacements or eigenvectors, from a MSC/NASTRAN F06 file or a UAI/NASTRAN PRT file. The displacements are written into a unformatted scratch file where they are available for rapid access when the user wishes to display a deformed shape. All subcases or mode shapes can be read in at once. Then it is easy to enable the deformed shape, to change subcases or mode shapes and to change the scale factor for subsequent plots. NPLOT is written in VAX FORTRAN for DEC VAX series computers running VMS. As distributed, the NPLOT source code makes calls to the DI3000 graphics package from Precision Visuals; however, a set of interface routines is provided to translate the DI3000 calls into Tektronix PLOT10/TCS graphics library calls so that NPLOT can use the standard Tektronix 4010 which many PC terminal emulation software programs support. NPLOT is available in VAX BACKUP format on a 9-track 1600 BPI DEC VAX BACKUP format magnetic tape (standard media) or a TK50 tape cartridge. This program was developed in 1991. DEC, VAX, VMS, and TK50 are trademarks of Digital Equipment Corporation. Tektronix, PLOT10, and TCS are trademarks of Tektronix, Inc. DI3000 is a registered trademark of Precision Visuals, Inc. NASTRAN is a registered trademark of the National Aeronautics and Space Administration. MSC/ is a trademark of MacNeal-Schwendler Corporation. UAI is a trademark of Universal Analytics, Inc.

Description: New research on hypersonic vehicles, such as the National Aero-Space Plane (NASP), has raised concerns about the effects of shock-wave interference on various structural components of the craft. State-of-the-art aerothermal analysis software is inadequate to predict local flow and heat flux in areas of extremely high heat transfer, such as the surface impingement of an Edney-type supersonic jet. EASI revives and updates older computational methods for calculating inviscid flow field and maximum heating from shock wave interference. The program expands these methods to solve problems involving the six shock-wave interference patterns on a two-dimensional cylindrical leading edge with an equilibrium chemically reacting gas mixture (representing, for example, the scramjet cowl of the NASP). The inclusion of gas chemistry allows for a more accurate prediction of the maximum pressure and heating loads by accounting for the effects of high temperature on the air mixture. Caloric imperfections and specie dissociation of high-temperature air cause shock-wave angles, flow deflection angles, and thermodynamic properties to differ from those calculated by a calorically perfect gas model. EASI contains pressure- and temperature-dependent thermodynamic and transport properties to determine heating rates, and uses either a calorically perfect air model or an 11-specie, 7-reaction reacting air model at equilibrium with temperatures up to 15,000 K for the inviscid flowfield calculations. EASI solves the flow field and the associated maximum surface pressure and heat flux for the six common types of shock wave interference. Depending on the type of interference, the program solves for shock-wave/boundary-layer interaction, expansion-fan/boundary-layer interaction, attaching shear layer or supersonic jet impingement. Heat flux predictions require a knowledge (from experimental data or relevant calculations) of a pertinent length scale of the interaction. Output files contain flow-field information for the various shock-wave interference patterns and their associated maximum surface pressure and heat flux predictions. EASI is written in FORTRAN 77 for a DEC VAX 8500 series computer using the VAX/VMS operating system, and requires 75K of memory. The program is available on a 9-track 1600 BPI magnetic tape in DEC VAX BACKUP format. EASI was developed in 1989. DEC, VAX, and VMS are registered trademarks of the Digital Equipment Corporation.

Description: This code was developed to aid design engineers in the selection and evaluation of aerodynamically efficient wing-canard and wing-horizontal-tail configurations that may employ simple hinged-flap systems. Rapid estimates of the longitudinal aerodynamic characteristics of conceptual airplane lifting surface arrangements are provided. The method is particularly well suited to configurations which, because of high speed flight requirements, must employ thin wings with highly swept leading edges. The code is applicable to wings with either sharp or rounded leading edges. The code provides theoretical pressure distributions over the wing, the canard or horizontal tail, and the deflected flap surfaces as well as estimates of the wing lift, drag, and pitching moments which account for attainable leading edge thrust and leading edge separation vortex forces. The wing planform information is specified by a series of leading edge and trailing edge breakpoints for a right hand wing panel. Up to 21 pairs of coordinates may be used to describe both the leading edge and the trailing edge. The code has been written to accommodate 2000 right hand panel elements, but can easily be modified to accommodate a larger or smaller number of elements depending on the capacity of the target computer platform. The code provides solutions for wing surfaces composed of all possible combinations of leading edge and trailing edge flap settings provided by the original deflection multipliers and by the flap deflection multipliers. Up to 25 pairs of leading edge and trailing edge flap deflection schedules may thus be treated simultaneously. The code also provides for an improved accounting of hinge-line singularities in determination of wing forces and moments. To determine lifting surface perturbation velocity distributions, the code provides for a maximum of 70 iterations. The program is constructed so that successive runs may be made with a given code entry. To make additional runs, it is necessary only to add an identification record and the namelist data that are to be changed from the previous run. This code was originally developed in 1989 in FORTRAN V on a CDC 6000 computer system, and was later ported to an MS-DOS environment. Both versions are available from COSMIC. There are only a few differences between the PC version (LAR-14458) and CDC version (LAR-14178) of AERO2S distributed by COSMIC. The CDC version has one main source code file while the PC version has two files which are easier to edit and compile on a PC. The PC version does not require a FORTRAN compiler which supports NAMELIST because a special INPUT subroutine has been added. The CDC version includes two MODIFY decks which can be used to improve the code and prevent the possibility of some infrequently occurring errors while PC-version users will have to make these code changes manually. The PC version includes an executable which was generated with the Ryan McFarland/FORTRAN compiler and requires 253K RAM and an 80x87 math co-processor. Using this executable, the sample case requires about four hours to execute on an 8MHz AT-class microcomputer with a co-processor. The source code conforms to the FORTRAN 77 standard except that it uses variables longer than six characters. With two minor modifications, the PC version should be portable to any computer with a FORTRAN compiler and sufficient memory. The CDC version of AERO2S is available in CDC NOS Internal format on a 9-track 1600 BPI magnetic tape. The PC version is available on a set of two 5.25 inch 360K MS-DOS format diskettes. IBM AT is a registered trademark of International Business Machines. MS-DOS is a registered trademark of Microsoft Corporation. CDC is a registered trademark of Control Data Corporation. NOS is a trademark of Control Data Corporation.

Description: LATDYN is a computer code for modeling the Large Angle Transient DYNamics of flexible articulating structures and mechanisms involving joints about which members rotate through large angles. LATDYN extends and brings together some of the aspects of Finite Element Structural Analysis, Multi-Body Dynamics, and Control System Analysis; three disciplines that have been historically separate. It combines significant portions of their distinct capabilities into one single analysis tool. The finite element formulation for flexible bodies in LATDYN extends the conventional finite element formulation by using a convected coordinate system for constructing the equation of motion. LATDYN's formulation allows for large displacements and rotations of finite elements subject to the restriction that deformations within each are small. Also, the finite element approach implemented in LATDYN provides a convergent path for checking solutions simply by increasing mesh density. For rigid bodies and joints LATDYN borrows extensively from methodology used in multi-body dynamics where rigid bodies may be defined and connected together through joints (hinges, ball, universal, sliders, etc.). Joints may be modeled either by constraints or by adding joint degrees of freedom. To eliminate error brought about by the separation of structural analysis and control analysis, LATDYN provides symbolic capabilities for modeling control systems which are integrated with the structural dynamic analysis itself. Its command language contains syntactical structures which perform symbolic operations which are also interfaced directly with the finite element structural model, bypassing the modal approximation. Thus, when the dynamic equations representing the structural model are integrated, the equations representing the control system are integrated along with them as a coupled system. This procedure also has the side benefit of enabling a dramatic simplification of the user interface for modeling control systems. Three FORTRAN computer programs, the LATDYN Program, the Preprocessor, and the Postprocessor, make up the collective LATDYN System. The Preprocessor translates user commands into a form which can be used while the LATDYN program provides the computational core. The Postprocessor allows the user to interactively plot and manage a database of LATDYN transient analysis results. It also includes special facilities for modeling control systems and for programming changes to the model which take place during analysis sequence. The documentation includes a Demonstration Problem Manual for the evaluation and verification of results and a Postprocessor guide. Because the program should be viewed as a byproduct of research on technology development, LATDYN's scope is limited. It does not have a wide library of finite elements, and 3-D Graphics are not available. Nevertheless, it does have a measure of "user friendliness". The LATDYN program was developed over a period of several years and was implemented on a CDC NOS/VE & Convex Unix computer. It is written in FORTRAN 77 and has a virtual memory requirement of 1.46 MB. The program was validated on a DEC MICROVAX operating under VMS 5.2.

Description: The increasing number of applications of fiber-reinforced composites in industry demands a detailed understanding of their material properties and behavior. A three-dimensional finite-element computer program called PAFAC (Plastic and Failure Analysis of Composites) has been developed for the elastic-plastic analysis of fiber-reinforced composite materials and structures. The evaluation of stresses and deformations at edges, cut-outs, and joints is essential in understanding the strength and failure for metal-matrix composites since the onset of plastic yielding starts very early in the loading process as compared to the composite's ultimate strength. Such comprehensive analysis can only be achieved by a finite-element program like PAFAC. PAFAC is particularly suited for the analysis of laminated metal-matrix composites. It can model the elastic-plastic behavior of the matrix phase while the fibers remain elastic. Since the PAFAC program uses a three-dimensional element, the program can also model the individual layers of the laminate to account for thickness effects. In PAFAC, the composite is modeled as a continuum reinforced by cylindrical fibers of vanishingly small diameter which occupy a finite volume fraction of the composite. In this way, the essential axial constraint of the phases is retained. Furthermore, the local stress and strain fields are uniform. The PAFAC finite-element solution is obtained using the displacement method. Solution of the nonlinear equilibrium equations is obtained with a Newton-Raphson iteration technique. The elastic-plastic behavior of composites consisting of aligned, continuous elastic filaments and an elastic-plastic matrix is described in terms of the constituent properties, their volume fractions, and mutual constraints between phases indicated by the geometry of the microstructure. The program uses an iterative procedure to determine the overall response of the laminate, then from the overall response determines the stress state in each phase of the composite material. Failure of the fibers or matrix within an element can also be modeled by PAFAC. PAFAC is written in FORTRAN IV for batch execution and has been implemented on a CDC CYBER 170 series computer with a segmented memory requirement of approximately 66K (octal) of 60 bit words. PAFAC was developed in 1982.

Description: Since its early beginnings, NASA has been actively involved in the design and testing of airfoil sections for a wide variety of applications. Recently a set of programs has been developed to smooth and scale arbitrary airfoil coordinates. The smoothing program, AFSMO, utilizes both least-squares polynomial and least-squares cubic-spline techniques to iteratively smooth the second derivatives of the y-axis airfoil coordinates with respect to a transformed x-axis system which unwraps the airfoil and stretches the nose and trailing-edge regions. The corresponding smooth airfoil coordinates are then determined by solving a tridiagonal matrix of simultaneous cubic-spline equations relating the y-axis coordinates and their corresponding second derivatives. The camber and thickness distribution of the smooth airfoil are also computed. The scaling program, AFSCL, may then be used to scale the thickness distribution generated by the smoothing program to a specified maximum thickness. Once the thickness distribution has been scaled, it is combined with the camber distribution to obtain the final scaled airfoil contour. The airfoil smoothing and scaling programs are written in FORTRAN IV for batch execution and have been implemented on a CDC CYBER 170 series computer with a central memory requirement of approximately 70K (octal) of 60 bit words. Both programs generate plotted output via CALCOMP type plotting calls. These programs were developed in 1983.

Description: The Panel Analysis and Sizing Code (PASCO) was developed for the buckling and vibration analysis and sizing of prismatic structures having an arbitrary cross section. PASCO is primarily intended for analyzing and sizing stiffened panels made of laminated orthotropic materials and is of particular value in analyzing and sizing filamentary composite structures. When used in the analysis mode, PASCO calculates laminate stiffnesses, lamina stress and strains (including the effects of temperature and panel bending), buckling loads, vibration frequencies, and overall panel stiffness. When used in the sizing mode, PASCO adjusts sizing variables to provide a low-mass panel design that carries a set of specified loadings without exceeding buckling or material strength allowables and that meets other design requirements such as upper and lower bounds on sizing variables, upper and lower bounds on overall bending, extensional and shear stiffnesses, and lower bounds on vibration frequencies. Although emphasis in PASCO is placed on flat panels having several identical bays, the only restriction on configuration modeling is that the structure is assumed to be prismatic. In addition, it is assumed that loads and temperatures do not vary along the length of a panel. Because of their wide application in aerospace structures, stiffened panels are readily handled by PASCO. The panel cross section may be composed of an arbitrary assemblage of thin, flat, rectangular plate elements that are connected together along their longitudinal edges. Each plate element consists of a balanced symmetric laminate of any number of layers of orthotropic material. Any group of element widths, layer thicknesses, and layer orientation angles may be selected as sizing variables. Substructuring is available to increase the efficiency of the analysis and to simplify the modeling of complicated structures. The Macintosh version of PASCO includes an interactive, graphic preprocessor called MacPASCO. The main objective of MacPASCO is to make the use of PASCO faster, simpler, and less error-prone. By using a graphical user interface (GUI), MacPASCO simplifies the specification of panel geometry and reduces user input errors, thus making the modeling and analysis of panel designs more efficient. The user draws the initial structural geometry on the computer screen, then uses a combination of graphic and text inputs to: refine the structural geometry, specify information required for analysis such as panel load conditions, and define design variables and constraints for minimum-mass optimization. Composite panel design is an ideal application because the graphical user interface can: serve as a visual aid, eliminate the tedious aspects of text-based input, and eliminate many sources of input errors. The current version of MacPASCO does not implement all the modeling features of PASCO, but has been found to be sufficient for many users. Many difficulties common to text-based inputs are avoided because MacPASCO uses a GUI. First, the graphic displays eliminate syntax errors, like misplaced commas and incorrect command names, because there is no text-based syntax. Second, graphic displays allow the user to see the geometry as it is created and immediately detect and correct any errors. Third, MacPASCO's drawing tools have been designed to avoid modeling errors. Fourth, the graphic displays make revisions to existing structural designs much easier and less error-prone by eliminating the need for the user to conceptualize the text input as geometry. The user can work directly with the geometry displayed on the screen. Finally, MacPASCO automatically generates the correct PASCO input from the geometry displayed on the screen. This input file can be used with any machine version of PASCO to actually perform the analysis and sizing and to output results. The DEC VAX version of PASCO is written in FORTRAN IV for batch execution and has been implemented on a DEC VAX series computer. The Macintosh version of PASCO was developed for Macintosh II series computers with at least 2Mb of RAM running MPW Pascal 3.0 and Language Systems FORTRAN 2.0 under the MPW programming environment. It includes MPW compatible makefiles for compiling the source code. The Macintosh version uses input files compatible with versions of PASCO running on different platforms. MacPASCO is written in Macintosh Programmers Workbench 3.0, MPW Pascal 3.0, and MacAPP 2.0. The Pascal source code is included on the distribution diskette. MacAPP is a development library which is not included. MacPASCO requires a Mac Plus, SE/30, or MacII, IIx, IIcx, IIci, or IIfx running System 6.0 or greater. MacPASCO is System 7.0 compatible. A minimum of 2Mb of RAM is required for execution. The Macintosh version of PASCO is distributed on four 3.5 inch 800K Macintosh format diskettes. The DEC VAX version is distributed on a 9-track 1600 BPI magnetic tape. The PASCO program was developed in 1981, adapted to the DEC VAX in 1983 and to the Macintosh in 1991. MacPASCO was released in 1992.

Description: The Supersonic Wing Nonlinear Aerodynamics computer program, LTSTAR, was developed to provide for the estimation of the nonlinear aerodynamic characteristics of a wing at supersonic speeds. This corrected linearized-theory method accounts for nonlinearities in the variation of basic pressure loadings with local surface slopes, predicts the degree of attainment of theoretical leading-edge thrust forces, and provides an estimate of detached leading-edge vortex loadings that result when the theoretical thrust forces are not fully realized. Comparisons of LTSTAR computations with experimental results show significant improvements in detailed wing pressure distributions, particularly for large angles of attack and for regions of the wing where the flow is highly three-dimensional. The program provides generally improved predictions of the wing overall force and moment coefficients. LTSTAR could be useful in design studies aimed at aerodynamic performance optimization and for providing more realistic trade-off information for selection of wing planform geometry and airfoil section parameters. Input to the LTSTAR program includes wing planform data, freestream conditions, wing camber, wing thickness, scaling options, and output options. Output includes pressure coefficients along each chord, section normal and axial force coefficients, and the spanwise distribution of section force coefficients. With the chordwise distributions and section coefficients at each angle of attack, three sets of polars are output. The first set is for linearized theory with and without full leading-edge thrust, the second set includes nonlinear corrections, and the third includes estimates of attainable leading-edge thrust and vortex increments along with the nonlinear corrections. The LTSTAR program is written in FORTRAN IV for batch execution and has been implemented on a CDC 6000 series computer with a central memory requirement of approximately 150K (octal) of 60 bit words. The LTSTAR program was developed in 1980.

Description: The nozzle afterbody is one of the main drag-producing components of an aircraft propulsion system. Thus, considerable effort has been devoted to developing techniques for predicting the afterbody flow field and drag. The RAXJET computer program was developed to predict the transonic, axisymmetric flow over nozzle afterbodies with supersonic jet exhausts and includes the effects of boundary-layer displacement, separation, jet entrainment, and inviscid jet plume blockage. RAXJET iteratively combines the South-Jameson relaxation procedure, the Reshotko-Tucker boundary-layer solution, the Presz separation model, the Dash-Pergament mixing model, and the Dash-Thorpe inviscid plume model into a single, comprehensive model. The approach taken in the RAXJET program requires considerably less computational time than the Navier-Stokes solutions and generally yields results of comparable accuracy. In RAXJET, the viscous-inviscid interaction model is constructed by dividing the afterbody flow field into six separate computational regions: (1) The inviscid external flow solution is based on the relaxation procedure of South and Jameson for solving the exact nonlinear potential flow equation in nonconservative form. (2) The flow field in the inviscid jet exhaust is solved by explicit spatial marching of the conservative finite-difference form of the inviscid flow equations for a uniform composition gas mixture. (3) The properties in the attached boundary-layer region are solved by a modified version of the Reshotko-Tucker integral method for turbulent flows. (4) The analysis of the separated flow region consists of predicting the separation location and calculating the discriminating streamline shape. (5) The jet wake region is determined by either a simple extrapolation model or by an integral method that accounts for entrainment effects. (6) The displacement-thickness distribution arising from entrainment into the jet mixing layer is calculated by the overlaid mixing model. The inviscid external flow solution and inviscid jet exhaust solution provide the necessary flow conditions to calculate the flow in the viscous regions. The viscous and inviscid flow fields are iteratively solved until a final solution is obtained. Input to the RAXJET program consists of body geometry data, free-stream conditions, main logic control parameters, and condition and control parameters for each of the six computational flow regions. Output from RAXJET includes detailed flow results and aerodynamic coefficients. The RAXJET program is written in FORTRAN IV for batch execution and has been implemented on a CDC CYBER 170 series computer with a central memory requirement of approximately 60K(octal) of 60 bit words. The RAXJET program was developed in 1982.

Description: Current investigations of the hydrogen-fueled supersonic combustion ramjet engine have delineated several technological problem areas. One area, the analysis of the injection, turbulent mixing, and combusiton of hydrogen, requires the accurate calculation of the supersonic combustion flow fields. This calculation has proven difficult because of an interesting phenomena which makes possible the transition from supersonic to subsonic flow in the combustion field, due to the temperature transitions which occur in the flow field. This computer program was developed to use viscous characteristics theory to analyze supersonic combustion flow fields with imbedded subsonic regions. Intended to be used as a practical design tool for two-dimensional and axisymmetric supersonic combustor development, this program has proven useful in the analysis of such problems as determining the flow field of a single underexpanded hydrogen jet, the internal flow of a gas sampling probe, the effects of fuel-injector strut shape, and the effects of changes in combustor configuration. Both combustion and diffusive effects can significantly alter the wave pattern in a supersonic field and generate significant pressure gradients in both the axial and radial directions. The induced pressure, in turn, substantially influences the ignition delay and reaction times as well as the velocity distribution. To accurately analyze the flow fields, the effects of finite rate chemistry, mixing, and wave propagation must be properly linked to one another. The viscous characteristics theory has been used in the past to describe flows that are purely supersonic; however, the interacting pressure effects in the combustor often allow for the development of shock waves and imbedded subsonic regions. Numerical investigation of these transonic situations has required the development of a new viscous characteristics procedure which is valid within the subsonic region and can be coupled with the standard viscous characteristics procedure in the supersonic region. The basic governing equations used are the 'viscous-inviscid' equations, similar to those employed in higher-order boundary layer analyses, with finite rate chemistry terms included. In addition, the Rankine-Hugoniot and Prandtl-Meyer relations are used to compute shock and expansion conditions. The program can handle up to 20 simultaneous shock waves. Chemistry terms are computed for a 7-species 8-mechanism hydrogen-air reaction scheme. The user input consists of a physical description of the combustor and flow determination parameters. Output includes detail flow parameter values at selected points within the flow field. This computer program is written in FORTRAN IV for batch execution and has been implemented on a CDC CYBER 175 with a central memory requirement of approximately 114K (octal) of 60 bit words. The program was developed in 1978.

Description: The computer program SALLY was developed to compute the incompressible linear stability characteristics and integrate the amplification rates of boundary layer disturbances on swept and tapered wings. For some wing designs, boundary layer disturbance can significantly alter the wing performance characteristics. This is particularly true for swept and tapered laminar flow control wings which incorporate suction to prevent boundary layer separation. SALLY should prove to be a useful tool in the analysis of these wing performance characteristics. The first step in calculating the disturbance amplification rates is to numerically solve the compressible laminar boundary-layer equation with suction for the swept and tapered wing. A two-point finite-difference method is used to solve the governing continuity, momentum, and energy equations. A similarity transformation is used to remove the wall normal velocity as a boundary condition and place it into the governing equations as a parameter. Thus the awkward nonlinear boundary condition is avoided. The resulting compressible boundary layer data is used by SALLY to compute the incompressible linear stability characteristics. The local disturbance growth is obtained from temporal stability theory and converted into a local growth rate for integration. The direction of the local group velocity is taken as the direction of integration. The amplification rate, or logarithmic disturbance amplitude ratio, is obtained by integration of the local disturbance growth over distance. The amplification rate serves as a measure of the growth of linear disturbances within the boundary layer and can serve as a guide in transition prediction. This program is written in FORTRAN IV and ASSEMBLER for batch execution and has been implemented on a CDC CYBER 70 series computer with a central memory requirement of approximately 67K (octal) of 60 bit words. SALLY was developed in 1979.

Description: SINDA, the Systems Improved Numerical Differencing Analyzer, is a software system for solving lumped parameter representations of physical problems governed by diffusion-type equations. SINDA was originally designed for analyzing thermal systems represented in electrical analog, lumped parameter form, although its use may be extended to include other classes of physical systems which can be modeled in this form. As a thermal analyzer, SINDA can handle such interrelated phenomena as sublimation, diffuse radiation within enclosures, transport delay effects, and sensitivity analysis. FLUINT, the FLUid INTegrator, is an advanced one-dimensional fluid analysis program that solves arbitrary fluid flow networks. The working fluids can be single phase vapor, single phase liquid, or two phase. The SINDA'85/FLUINT system permits the mutual influences of thermal and fluid problems to be analyzed. The SINDA system consists of a programming language, a preprocessor, and a subroutine library. The SINDA language is designed for working with lumped parameter representations and finite difference solution techniques. The preprocessor accepts programs written in the SINDA language and converts them into standard FORTRAN. The SINDA library consists of a large number of FORTRAN subroutines that perform a variety of commonly needed actions. The use of these subroutines can greatly reduce the programming effort required to solve many problems. A complete run of a SINDA'85/FLUINT model is a four step process. First, the user's desired model is run through the preprocessor which writes out data files for the processor to read and translates the user's program code. Second, the translated code is compiled. The third step requires linking the user's code with the processor library. Finally, the processor is executed. SINDA'85/FLUINT program features include 20,000 nodes, 100,000 conductors, 100 thermal submodels, and 10 fluid submodels. SINDA'85/FLUINT can also model two phase flow, capillary devices, user defined fluids, gravity and acceleration body forces on a fluid, and variable volumes. SINDA'85/FLUINT offers the following numerical solution techniques. The Finite difference formulation of the explicit method is the Forward-difference explicit approximation. The formulation of the implicit method is the Crank-Nicolson approximation. The program allows simulation of non-uniform heating and facilitates modeling thin-walled heat exchangers. The ability to model non-equilibrium behavior within two-phase volumes is included. Recent improvements to the program were made in modeling real evaporator-pumps and other capillary-assist evaporators. SINDA'85/FLUINT is available by license for a period of ten (10) years to approved licensees. The licensed program product includes the source code and one copy of the supporting documentation. Additional copies of the documentation may be purchased separately at any time. SINDA'85/FLUINT is written in FORTRAN 77. Version 2.3 has been implemented on Cray series computers running UNICOS, CONVEX computers running CONVEX OS, and DEC RISC computers running ULTRIX. Binaries are included with the Cray version only. The Cray version of SINDA'85/FLUINT also contains SINGE, an additional graphics program developed at Johnson Space Flight Center. Both source and executable code are provided for SINGE. Users wishing to create their own SINGE executable will also need the NASA Device Independent Graphics Library (NASADIG, previously known as SMDDIG; UNIX version, MSC-22001). The Cray and CONVEX versions of SINDA'85/FLUINT are available on 9-track 1600 BPI UNIX tar format magnetic tapes. The CONVEX version is also available on a .25 inch streaming magnetic tape cartridge in UNIX tar format. The DEC RISC ULTRIX version is available on a TK50 magnetic tape cartridge in UNIX tar format. SINDA was developed in 1971, and first had fluid capability added in 1975. SINDA'85/FLUINT version 2.3 was released in 1990.

Description: The Program for Solving the General-Frequency Unsteady Two-Dimensional Transonic Small-Disturbance Equation, XTRAN2L, is used to calculate time-accurate, finite-difference solutions of the nonlinear, small-disturbance potential equation for two- dimensional transonic flow about airfoils. The code can treat forced harmonic, pulse, or aeroelastic transient type motions. XTRAN2L uses a transonic small-disturbance equation that incorporates a time accurate finite-difference scheme. Airfoil flow tangency boundary conditions are defined to include airfoil contour, chord deformation, nondimensional plunge displacement, pitch, and trailing edge control surface deflection. Forced harmonic motion can be based on: 1) coefficients of harmonics based on information from each quarter period of the last cycle of harmonic motion; or 2) Fourier analyses of the last cycle of motion. Pulse motion (an alternate to forced harmonic motion) in which the airfoil is given a small prescribed pulse in a given mode of motion, and the aerodynamic transients are calculated. An aeroelastic transient capability is available within XTRAN2L, wherein the structural equations of motion are coupled with the aerodynamic solution procedure for simultaneous time-integration. The wake is represented as a slit downstream of the airfoil trailing edge. XTRAN2L includes nonreflecting farfield boundary conditions. XTRAN2L was developed on a CDC CYBER mainframe running under NOS 2.4. It is written in FORTRAN 5 and uses overlays to minimize storage requirements. The program requires 120K of memory in overlayed form. XTRAN2L was developed in 1987.

Description: The transient response of a thermal protection material to heat applied to the surface can be calculated using the CHAP III computer program. CHAP III can be used to analyze pyrolysis gas chemical kinetics in detail and examine pyrolysis reactions-indepth. The analysis includes the deposition of solid products produced by chemical reactions in the gas phase. CHAP III uses a modelling technique which can approximate a wide range of ablation problems. The energy equation used in CHAP III incorporates pyrolysis (both solid and gas reactions), convection, conduction, storage, work, kinetic energy, and viscous dissipation. The chemically reacting components of the solid are allowed to vary as a function of position and time. CHAP III employs a finite difference method to approximate the energy equations. Input values include specific heat, thermal conductivity, thermocouple locations, enthalpy, heating rates, and a description of the chemical reactions expected. The output tabulates the temperature at locations throughout the ablator, gas flow within the solid, density of the solid, weight of pyrolysis gases, and rate of carbon deposition. A sample case is included, which analyzes an ablator material containing several pyrolysis reactions subjected to an environment typical of entry at lunar return velocity. CHAP III is written in FORTRAN IV for batch execution and has been implemented on a CDC CYBER 170 series computer operating under NOS with a central memory requirement of approximately 102K (octal) of 60 bit words. This program was developed in 1985.

Description: BUCKO is a computer program developed to predict the buckling load of a rectangular compression-loaded orthotropic plate with a centrally located cutout. The plate is assumed to be a balanced, symmetric laminate of uniform thickness. The cutout shape can be elliptical, circular, rectangular, or square. The BUCKO package includes sample data that demonstrates the essence of the program and its ease of usage. BUCKO uses an approximate one-dimensional formulation of the classical two-dimensional buckling problem following the Kantorovich method. The boundary conditions are considered to be simply supported unloaded edges and either clamped or simply supported loaded edges. The plate is loaded in uniaxial compression by either uniformly displacing or uniformly stressing two opposite edges of the plate. The BUCKO analysis consists of two parts: calculation of the inplane stress distribution prior to buckling, and calculation of the plate axial load and displacement at buckling. User input includes plate planform and cutout geometry, plate membrane and bending stiffnesses, finite difference parameters, boundary condition data, and loading data. Results generated by BUCKO are the prebuckling strain energy, inplane stress resultants, buckling mode shape, critical end shortening, and average axial and transverse strains at buckling. BUCKO is written in FORTRAN V for batch execution and has been implemented on a CDC CYBER 170 series computer operating under NOS with a central memory requirement of approximately 343K of 60 bit words. This program was developed in 1984 and was last updated in 1990.

Description: The NASCRIN program was developed for analyzing two-dimensional flow fields in supersonic combustion ramjet (scramjet) inlets. NASCRIN solves the two-dimensional Euler or Navier-Stokes equations in conservative form by an unsplit, explicit, two-step finite-difference method. A more recent explicit-implicit, two-step scheme has also been incorporated in the code for viscous flow analysis. An algebraic, two-layer eddy-viscosity model is used for the turbulent flow calculations. NASCRIN can analyze both inviscid and viscous flows with no struts, one strut, or multiple struts embedded in the flow field. NASCRIN can be used in a quasi-three-dimensional sense for some scramjet inlets under certain simplifying assumptions. Although developed for supersonic internal flow, NASCRIN may be adapted to a variety of other flow problems. In particular, it should be readily adaptable to subsonic inflow with supersonic outflow, supersonic inflow with subsonic outflow, or fully subsonic flow. The NASCRIN program is available for batch execution on the CDC CYBER 203. The vectorized FORTRAN version was developed in 1983. NASCRIN has a central memory requirement of approximately 300K words for a grid size of about 3,000 points.

Description: The problem of axisymmetric transonic flow is of interest not only because of the practical application to missile and launch vehicle aerodynamics, but also because of its relation to fully three-dimensional flow in terms of the area rule. The RAXBOD computer program was developed for the analysis of steady, inviscid, irrotational, transonic flow over axisymmetric bodies in free air. RAXBOD uses a finite-difference relaxation method to numerically solve the exact formulation of the disturbance velocity potential with exact surface boundary conditions. Agreement with available experimental results has been good in cases where viscous effects and wind-tunnel wall interference are not important. The governing second-order partial differential equation describing the flow potential is replaced by a system of finite difference equations, including Jameson's "rotated" difference scheme at supersonic points. A stretching is applied to both the normal and tangential coordinates such that the infinite physical space is mapped onto a finite computational space. The boundary condition at infinity can be applied directly and there is no need for an asymptotic far-field solution. The system of finite difference equations is solved by a column relaxation method. In order to obtain both rapid convergence and any desired resolution, the relaxation is performed iteratively on successively refined grids. Input to RAXBOD consists of a description of the body geometry, the free stream conditions, and the desired resolution control parameters. Output from RAXBOD includes computed geometric parameters in the normal and tangential directions, iteration history information, drag coefficients, flow field data in the computational plane, and coordinates of the sonic line. This program is written in FORTRAN IV for batch execution and has been implemented on a CDC 6600 computer with an overlayed central memory requirement of approximately 40K (octal) of 60 bit words. Optional plotted output can be generated for the Calcomp plotting system. The RAXBOD program was developed in 1976.

Description: This program performs a one-dimensional numerical analysis of the transient thermal response of multi-layer insulative systems. The analysis can determine the temperature distribution through a system consisting of from one to four layers, one of which can be an air gap. Concentrated heat sinks at any interface can be included. The computer program based on the analysis will determine the thickness of a specified layer that will satisfy a temperature limit criterion at any point in the insulative system. The program will also automatically calculate the thickness at several points on a system and determine the total system mass. This program was developed as a tool for designing thermal protection systems for high-speed aerospace vehicles but could be adapted to many areas of industry involved in thermal insulation systems. In this package, the equations describing the transient thermal response of a system are developed. The governing differential equation for each layer and boundary condition are put in finite-difference form using a Taylor's series expansion. These equations yield an essentially tridiagonal matrix of unknown temperatures. A procedure based on Gauss' elimination method is used to solve the matrix. This program is written in FORTRAN IV for the CDC RUN compiler and has been implemented on a CDC 6000 series machine operating under SCOPE 3.0. This program requires a minimum of 44K (octal) of 60 bit words of memory.

Description: A computer program has been developed to analyze the transient response of an ablating axisymmetric body, including the effect of shape change. The governing differential equation, the boundary conditions for the analysis on which the computer program is based, and the method of solution of the resulting finite-difference equations are discussed in the documentation. Some of the features of the analysis and the associated program are (1) the ablation material is considered to be orthotropic with temperature-dependent thermal properties; (2) the thermal response of the entire body is considered simultaneously; (3) the heat transfer and pressure distribution over the body are adjusted to the new geometry as ablation occurs; (4) the governing equations and several boundary-condition options are formulated in terms of generalized orthogonal coordinates for fixed points in a moving coordinate system; (5) the finite-difference equations are solved implicitly; and (6) other instantaneous body shapes can be displayed with a user-supplied plotting routine. The physical problem to be modeled with the analysis is described by FORTRAN input variables. For example, the external body geometry is described in the W, Z coordinates; material density is given; and the stagnation cold-wall heating rate is given in a time-dependent array. Other input variables are required which control the solution, specify boundary conditions, and determine output from the program. The equations have been programmed so that either the International System of Units or the U. S. Customary Units may be used. This program is written in FORTRAN IV for batch execution and has been implemented on a CDC 6000 Series computer. This program was developed in 1972.

Description: Two separate and distinct theories are incorporated in this computer program to estimate the lift-induced pressures existent on a wing-body combination. These are (1) the second-order shock-expansion theory, which is used to obtain the lifting pressures on the body alone at small angles of attack, and (2) the linear-theory integral equations, which is used to evaluate the lifting pressures induced by the wing. These equations relate the local surface slope at a point on the lifting surface to the pressure differential at the point and the influence of the pressures upstream of the point. The numerical solution of these equations is effected by treating the wing-planform as a composite of elemental rectangles and applying summation techniques to satisfy the necessary integral relations. Most of the input required by this program is involved with the description of the missile planform geometry. The output consists of the computed value of the lifting pressure slope (the differential pressure coefficient per degree angle of attack) for each of the elements in the planform array. A force and moment summary is presented for the configuration under consideration.

Description: A modified strip analysis has been developed for rapidly predicting flutter of finite-span, swept or unswept wings at subsonic to hypersonic speeds. The method employs distributions of aerodynamic parameters which may be evaluated from any suitable linear or nonlinear steady-flow theory or from measured steady-flow load distributions for the underformed wing. The method has been shown to give good flutter results for a broad range of wings at Mach number from 0 to as high as 15.3. The principles of the modified strip analysis may be summarized as follows: Variable section lift-curve slope and aerodynamic center are substituted respectively, for the two-dimensional incompressible-flow values of 2 pi and quarter chord which were employed by Barmby, Cunningham, and Garrick. Spanwise distributions of these steady-flow section aerodynamic parameters, which are pertinent to the desired planform and Mach number, are used. Appropriate values of Mach number-dependent circulation functions are obtained from two-dimensional unsteady compressible-flow theory. Use of the modified strip analysis avoids the necessity of reevaluating a number of loading parameters for each value of reduced frequency, since only the modified circulation functions, and of course the reduced frequency itself, vary with frequency. It is therefore practical to include in the digital computing program a very brief logical subroutine, which automatically selects reduced-frequency values that converge on a flutter solution. The problem of guessing suitable reduced-frequency values is thus eliminated, so that a large number of flutter points can be completely determined in a single brief run on the computing machine. If necessary, it is also practical to perform the calculations manually. Flutter characteristics have been calculated by the modified strip analysis and compared with results of other calculations and with experiments for Mach numbers up to 15.3 and for wings with sweep angles from 0 degrees to 52.5 degrees, aspect ratios from 2.0 to 7.4, taper ratios from 0.2 to 1.0, and center-of-gravity positions between 34% chord and 59% chord. These ranges probably cover the great majority of wings that are of practical interest with the exception of very low-aspect-ratio surfaces such as delta wings and missile fins. This program has been implemented on the IBM 7094.