ALBERT

Description: Author: Jake Yeston

Description: Author: Jake Yeston

Description: Polynitrogens have the potential for ultrahigh-performing explosives or propellants because singly or doubly bonded polynitrogens can decompose to triply bonded dinitrogen (N2) with an extraordinarily large energy release. The large energy content and relatively low activation energy toward decomposition makes the synthesis of a stable polynitrogen allotrope an extraordinary challenge. Many elements exist in different forms (allotropes)—for example, carbon can exist as graphite, diamond, buckyballs, or graphene. However, no stable neutral allotropes are known for nitrogen, and only two stable homonuclear polynitrogen ions had been isolated until now—namely, the N3− anion (1) and the N5+ cation (2). On page 374 of this issue, Zhang et al. (3) report the synthesis and characterization of the first stable salt of the cyclo-N5− anion, only the third stable homonuclear polynitrogen ion ever isolated. Author: Karl O. Christe

Description: Author: Jake Yeston

Description: Author: Jake Yeston

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 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: 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 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: 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: 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: 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: 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: 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: 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 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 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: 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: 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 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 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: 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: 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: 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: 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: 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 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: 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: The GTRAN program was developed to solve transient, as well as steady state, problems for gas piping systems. GTRAN capabilities allow for the analysis of a variety of system configurations and components. These include: multiple pipe junctions; valves that change position with time; fixed restrictions (orifices, manual valves, filters, etc.); relief valves; constant pressure sources; and heat transfer for insulated piping and piping subjected to free or forced convection. In addition, boundary conditions can be incorporated to simulate specific components. The governing equations of GTRAN are the one-dimensional transient gas dynamic equations. The three equations for pressure, velocity, and density are reduced to numerical equations using an implicit Crank-Nicholson finite difference technique. Input to GTRAN includes a description of the piping network, the initial conditions, and any events (e.g. valve closings) occuring during the period of analysis. Output includes pressure, velocity, and density versus time. GTRAN is written in FORTRAN 77 for batch execution and has been implemented on a DEC VAX series computer. GTRAN was developed in 1983.

Description: Results are presented for the numerical simulation of unsteady viscous incompressible flow past thick airfoils. Specifically, flow past a NACA 4424 at an angle of attack of 2.5 deg and Reynolds numbers in the range of 1700-4000 has been simulated using the spectral element method. At these conditions the flow is separatedd and an unsteady wake is formed. Application of the method of empirical eigenfunction reveals the structure of the most energetic components of the flow. These are found to occur in pairs that, through phase exchange, are responsible for the vortex shedding. A set of ordinary differential equations is obtained for the amplitudes of these eigenfunctions by a Galerkin projection of the Navier-Stokes equations. The solutions of the model system are compared with the full simulation. The work is of relevance to the transition process and observed routes to chaos in airfoil wakes.

Description: The objective of this study is to develop a reduced mechanism for ethylene oxidation. The authors are interested in a model with a minimum number of species and reactions that still models the chemistry with reasonable accuracy for the expected combustor conditions. The model will be validated by comparing the results to those calculated with a detailed kinetic model that has been validated against the experimental data.

Description: Artificial viscosity is added either implicity or explicitly in practically every numerical scheme for suppressing spurious oscillations in the solution of fluid-dynamics equations. In the present central-difference scheme, artificial viscosity is added explicitly for suppressing high-frequency oscillations and achieving good convergence properties. The amount of artificial viscosity added is controlled through the use of preselected coefficients. In the standard scheme, scalar coefficients based on the spectral radii of the Jacobian of the convective fluxes are used. However, this can add too much viscosity to the slower waves. Hence, the use of matrix-valued coefficients, which give appropriate viscosity for each wave component, is suggested. With the matrix-valued coefficients, the central-difference scheme produces more accurate solutions on a given grid, particularly in the vicinity of shocks and boundary layers, while still maintaining good convergence properties.

Description: A potential flow based three-dimensional panel method was modified to treat time-dependent conditions in which several submerged bodies can move within the fluid along different trajectories. This modification was accomplished by formulating the momentary solution in an inertial frame of reference, attached to the undisturbed stationary fluid. Consequently, the numerical interpretation of the multiple-body, solid-surface boundary condition and the viscous wake rollup was considerably simplified. The usteady capability of this code was calibrated and validated by comparing computed results with closed-form analytical results available for an airfoil, which was impulsively set into a constant speed forward motion. To demonstrate the multicomponent capability, computations were made for two wings following closely intersecting paths (i.e., simulations aimed at avoiding mid-air collisions) and for a flowfield with relative rotation (i.e., the case of a helicopter rotor rotating relative to the fuselage). Computed results for the cases were compared to experimental data, when such data was available.

Description: A thermoviscoplastic finite element method employing the Bodner-Partom constitutve model is used to investigate the response of simplified thermal-structural models to intense local heating. The computational method formulates the problem in rate and advances the solution in time by numerical integration. The thermoviscoplastic response of simplified structures with prescribed temperatures is investigated. With rapid rises of temperature, the nickel alloy structures display initially higher yield stresses due to strain rate effects. As temperatures approach elevated values, yield stress and stiffness degrade rapidly and pronounced plastic deformation occurs.

Description: A three-bay, space, cantilever truss is probabilistically evaluated to describe progressive buckling and truss collapse in view of the numerous uncertainties associated with the structural, material, and load variables that describe the truss. Initially, the truss is deterministically analyzed for member forces, and members in which the axial force exceeds the Euler buckling load are identified. These members are then discretized with several intermediate nodes, and a probabilistic buckling analysis is performed on the truss to obtain its probabilistic buckling loads and the respective mode shapes. Furthermore, sensitivities associated with the uncertainties in the primitive variables are investigated, margin of safety values for the truss are determined, and truss end node displacements are noted. These steps are repeated by sequentially removing buckled members until onset of truss collapse is reached. Results show that this procedure yields an optimum truss configuration for a given loading and for a specified reliability.

Description: An approximate method for calculating heating rates on three-dimensional vehicles at angle of attack is presented. The method is based on the axisymmetric analog for three-dimensional boundary layers and uses a generalized body-fitted coordinate system. Edge conditions for the boundary-layer solution are obtained from an inviscid flowfield solution, and because of the coordinate system used, the method is applicable to any blunt body geometry for which an inviscid flowfield solution can be obtained. The method is validated by comparing with experimental heating data and with thin-layer Navier-Stokes calculations on the shuttle orbiter at both wind-tunnel and flight conditions and with thin-layer Navier-Stokes calculations on the HL-20 at wind-tunnel conditions.

Description: This paper presents the results of a study on the variation of the critical Marangoni number (Ma(sub c)) for the onset of Benard convection in a finite liquid layer bounded horizontally as well as from below. A direct-numerical-simulation procedure is devised to determine the Ma(sub c) for aspect ratios (Ar) ranging from 0.8 to 10. The results predict a strong increase of Ma(sub c) as Ar decreases to below 2. A dip of Ma(sub c) occurs between Ar = 1.45 and 1.3, which is accompanied by a pattern transition from a two-cell convection to a unicellular flow. For Ar above 4, the calculated Ma(sub c) shows little change and asymptotically approach a value of 116.15, with Biot number (Bi) equal to 1.

Description: An experimental investigation carried out to determine aerodynamic and acoustic characteristics of a low area ratio rectangular jet ejector is reported. A supersonic primary jet issuing from a rectangular convergent-divergent nozzle of aspect ratio 4, into a rectangular duct of area ratio 3, was used. Improved performance was found when the ejector screech tone is most intense and appears to match the most unstable Strouhal number of the free rectangular jet. When the primary jet was operating at over and ideally expanded conditions, significant noise reduction was obtained with the ejector as compared to a corresponding free jet. Application of particle image velocimetry to high speed ejector flows was demonstrated through the measurement of instantaneous two dimensional velocity fields.

Description: Particle imaging velocimetry, which is rapidly becoming an essential tool for the synoptic measurement of two dimensional velocity fields, is addressed. This rapid development is carried out by a growing number of fluid mechanics experimentalists who recognize the unique capabilities of the technique to measure velocity fields in both space and time, and is supported by the concomitant development and availability of microcomputer hardware and software. Various applications of the technique to map different flow regimes are described and illustrated. Hardware implementations of the technique which utilize both conventional photography for image acquisition as well as digital 'online' methods for integrated image acquisition and processing are discussed. Recommendations to further enhance the technique and make it possible to map the three velocity components in three dimensional flow regions are given.

Description: One of the key technical elements in NASA's high speed research program is reducing the noise level to meet the federal noise regulation. The dominant noise source is associated with the supersonic jet discharged from the engine exhaust system. Whereas the turbulence mixing is largely responsible for the generation of the jet noise, a broadband shock-associated noise is also generated when the nozzle operates at conditions other than its design. For both mixing and shock noise components, because the source of the noise is embedded in the jet plume, one can expect that jet noise can be predicted from the jet flowfield computation. Mani et al. developed a unified aerodynamic/acoustic prediction scheme by applying an extension of Reichardt's aerodynamic model to compute turbulent shear stresses which are utilized in estimating the strength of the noise source. Although this method produces a fast and practical estimate of the jet noise, a modification by Khavaran et al. has led to an improvement in aerodynamic solution. The most notable feature in this work is that Reichardt's model is replaced with the computational fluid dynamics (CFD) solution of Reynolds-averaged Navier-Stokes equations. The major advantage of this work is that the essential, noise-related flow quantities such as turbulence intensity and shock strength can be better predicted. The predictions were limited to a shock-free design condition and the effect of shock structure on the jet mixing noise was not addressed. The present work is aimed at investigating this issue. Under imperfectly expanded conditions the existence of the shock cell structure and its interaction with the convecting turbulence structure may not only generate a broadband shock-associated noise but also change the turbulence structure, and thus the strength of the mixing noise source. Failure in capturing shock structures properly could lead to incorrect aeroacoustic predictions.

Description: A dimensionless group, called a pressure loss modulus (N(sub PL)), is introduced that, in conjunction with an appropriately defined Reynolds number, is of considerable engineering utility in correlating steady-state Delta p vs flow calibration data and subsequently as a predictor, using the same or a different fluid, in uniformly distributed pressure loss devices. It is particularly useful under operation in the transition regime. Applications of this simple bivariate correlation to three diverse devices of particular interest for small liquid rocket engine fluid systems are discussed: large L/D capillary tube restrictors, packed granular catalyst beds, and stacked vortex-loss disk restrictors.

Description: A numerical study of axisymmetric overexpanded nozzle is presented. The flow structure of the startup and throttle-down processes are examined. During the impulsive startup process, observed flow features include the Mach disk, separation shock, Mach stem, vortex core, contact surface, slip stream, initial shock front, and shocklet. Also the movement of the Mach disk is not monotonical in the downstream direction. For a range of pressure ratios, hysteresis phenomenon occurs; different solutions were obtained depending on different processes. Three types of flow structures were observed. The location of separation point and the lower end turning point of hysteresis are closely predicted. A high peak of pressure is associated with the nozzle flow reattachment. The reversed vortical structure and affects engine performance.

Description: Incompresible, thin sheet flows have been of research interest for many years. Those studies were mainly concerned with the stability of the flow in a surrounding gas. Squire was the first to carry out a linear, invicid stability analysis of sheet flow in air and compare the results with experiment. Dombrowski and Fraser did an experimental study of the disintegration of sheet flows using several viscous liquids. They also detected the formulation of holes in their sheet flows. Hagerty and Shea carried out an inviscid stability analysis and calculated growth rates with experimental values. They compared their calculated growth rates with experimental values. Taylor studied extensively the stability of thin liquid sheets both theoretically and experimentally. He showed that thin sheets in a vacuum are stable. Brown experimentally investigated thin liquid sheet flows as a method of application of thin films. Clark and Dumbrowski carried out second-order stability analysis for invicid sheet flows. Lin introduced viscosity into the linear stability analysis of thin sheet flows in a vacuum. Mansour and Chigier conducted an experimental study of the breakup of a sheet flow surrounded by high-speed air. Lin et al. did a linear stability analysis that included viscosity and a surrounding gas. Rangel and Sirignano carried out both a linear and nonlinear invisid stability analysis that applies for any density ratio between the sheet liquid and the surrounding gas. Now there is renewed interest in sheet flows because of their possible application as low mass radiating surfaces. The objective of this study is to investigate the fluid dynamics of sheet flows that are of interest for a space radiator system. Analytical expressions that govern the sheet geometry are compared with experimental results. Since a space radiator will operate in a vacuum, the analysis does not include any drag force on the sheet flow.

Description: The transition of an incompressible three-dimensional boundary layer with strong cross-flow is considered theoretically and computationally in the context of vortex/wave interactions. Specifically the work centers on two lower-branch Tollmien-Schlichting waves which mutually interact nonlinearly to induce a longitudinal vortex flow. The vortex motion in turn gives rise to significant wave modulation via wall-shear forcing. The characteristic Reynolds number is large and, as a consequence, the waves' and the vortex motion are governed primarily by triple deck theory. The nonlinear interaction is captured by a viscous partial-differential system for the vortex coupled with a pair of amplitude equations for each wave pressure. Following analysis and computation over a wide range of parameters, three distinct responses are found to emerge in the nonlinear behavior of the flow solution downstream: an algebraic finite-distance singularity, far-downstream saturation or far-downstream wave decay leaving pure vortex flow. These depend on the input conditions, the wave angles and the size of the cross flow.

Description: The small-crack effect, where small fatigue cracks grow faster and at lower stress-intensity factors than large cracks, has been found to be significant for many materials and loading conditions. In this paper, plasticity effects and crack-closure modelling of small fatigue cracks are reviewed. A crack-closure model with a cyclic-plastic-zone-corrected effective stress-intensity factor range (related to the cyclic J-integral) and microstructural data on crack-initiation sites were used to calculate small-crack growth rates and fatigue lives for unnotched and notched specimens made of two aluminum alloys. The crack-closure transient from the plastic wake was shown to be the dominant cause of the small-crack effect and plasticity effects on the cyclic-plastic-zone-corrected stress-intensity factor range were negligible except at extremely high stress levels. Small-crack growth rates and fatigue lives under both constant-amplitude and spectrum loading from tests and analyses agreed well.

Description: A computational method for predicting the two-phase transient fluid flow and heat transfer characteristics within a reservoir of the capillary-pumped-loop, intended to be used for spacecraft thermal management, has been developed. The model is based on the enthalpy formulation in an axisymmetric configuration. The reservoir operates under a constant thermodynamic pressure by allowing mass exchange between the reservoir and the outside loop. Both 1 g and 0 g environments have been considered to assess the effects of gravity on the reservoir performance. Depending on the gravity level, the power input and the reservoir orientation, three different convection modes have been identified, namely, the thermocapillary mode, the buoyancy mode, and the rapid-expansion mode (caused by interface movement). The impact of these modes on the performance of the reservoir and the associated physical phenomena have been discussed.

Description: A study is made of the thermomechanical buckling of flat unstiffened composite panels with central circular cutouts. The panels are subjected to combined temperature changes and applied edge loading (or edge displacements). The analysis is based on a first-order shear deformation plate theory. A mixed formulation is used with the fundamental unknowns consisting of the generalized displacements and the stress resultants of the plate. Both the stability boundary and the sensitivity coefficients are evaluated. The sensitivity coefficients measure the sensitivity of the buckling response to variations in the different lamination and material parameters of the panel. Numerical results are presented showing the effects of the variations in the hole diameter, laminate stacking sequence, fiber orientation, and aspect ratio of the panel on the thermomechanical buckling response and its sensitivity coefficients.

Description: A preconditioned Krylov subspace method (GMRES) is used to solve the linear systems of equations formed at each time-integration step of the unsteady, two-dimensional, compressible Navier-Stokes equations of fluid flow. The Navier-Stokes equations are cast in an implicit, upwind finite-volume, flux-split formulation. Several preconditioning techniques are investigated to enhance the efficiency and convergence rate of the implicit solver based on the GMRES algorithm. The superiority of the new solver is established by comparisons with a conventional implicit solver, namely line Gauss-Seidel relaxation (LGSR). Computational test results for low-speed (incompressible flow over a backward-facing step at Mach 0.1), transonic flow (trailing edge flow in a transonic turbine cascade), and hypersonic flow (shock-on-shock interactions on a cylindrical leading edge at Mach 6.0) are presented. For the Mach 0.1 case, overall speedup factors of up to 17 (in terms of time-steps) and 15 (in terms of CPU time on a CRAY-YMP/8) are found in favor of the preconditioned GMRES solver, when compared with the LGSR solver. The corresponding speedup factors for the transonic flow case are 17 and 23, respectively. The hypersonic flow case shows slightly lower speedup factors of 9 and 13, respectively. The study of preconditioners conducted in this research reveals that a new LUSGS-type preconditioner is much more efficient than a conventional incomplete LU-type preconditioner.

Description: A grid generation and flow solution algorithm for the Euler equations on unstructured grids is presented. The grid generation scheme utilizes Delaunay triangulation and self-generates the field points for the mesh based on cell aspect ratios and allows for clustering near solid surfaces. The flow solution method is an implicit algorithm in which the linear set of equations arising at each time step is solved using a Gauss Seidel procedure which is completely vectorizable. In addition, a study is conducted to examine the number of subiterations required for good convergence of the overall algorithm. Grid generation results are shown in two dimensions for a National Advisory Committee for Aeronautics (NACA) 0012 airfoil as well as a two-element configuration. Flow solution results are shown for two-dimensional flow over the NACA 0012 airfoil and for a two-element configuration in which the solution has been obtained through an adaptation procedure and compared to an exact solution. Preliminary three-dimensional results are also shown in which subsonic flow over a business jet is computed.

Description: PATSTAGS translates PATRAN finite model data into STAGS (Structural Analysis of General Shells) input records to be used for engineering analysis. The program reads data from a PATRAN neutral file and writes STAGS input records into a STAGS input file and a UPRESS data file. It is able to support translations of nodal constraints, nodal, element, force and pressure data. PATSTAGS uses three files: the PATRAN neutral file to be translated, a STAGS input file and a STAGS pressure data file. The user provides the names for the neutral file and the desired names of the STAGS files to be created. The pressure data file contains the element live pressure data used in the STAGS subroutine UPRESS. PATSTAGS is written in FORTRAN 77 for DEC VAX series computers running VMS. The main memory requirement for execution is approximately 790K of virtual memory. Output blocks can be modified to output the data in any format desired, allowing the program to be used to translate model data to analysis codes other than STAGSC-1 (HQN-10967). This program is available in DEC VAX BACKUP format on a 9-track magnetic tape or TK50 tape cartridge. Documentation is included in the price of the program. PATSTAGS was developed in 1990. DEC, VAX, TK50 and VMS are trademarks of Digital Equipment Corporation.

Description: The Composite Structure Preliminary Sizing program, COMPSIZE, is an analytical tool which structural designers can use when doing approximate stress analysis to select or verify preliminary sizing choices for composite structural members. It is useful in the beginning stages of design concept definition, when it is helpful to have quick and convenient approximate stress analysis tools available so that a wide variety of structural configurations can be sketched out and checked for feasibility. At this stage of the design process the stress/strain analysis does not need to be particularly accurate because any configurations tentatively defined as feasible will later be analyzed in detail by stress analysis specialists. The emphasis is on fast, user-friendly methods so that rough but technically sound evaluation of a broad variety of conceptual designs can be accomplished. Analysis equations used are, in most cases, widely known basic structural analysis methods. All the equations used in this program assume elastic deformation only. The default material selection is intermediate strength graphite/epoxy laid up in a quasi-isotropic laminate. A general flat laminate analysis subroutine is included for analyzing arbitrary laminates. However, COMPSIZE should be sufficient for most users to presume a quasi-isotropic layup and use the familiar basic structural analysis methods for isotropic materials, after estimating an appropriate elastic modulus. Homogeneous materials can be analyzed as simplified cases. The COMPSIZE program is written in IBM BASICA. The program format is interactive. It was designed on an IBM Personal Computer operating under DOS with a central memory requirement of approximately 128K. It has been implemented on an IBM compatible with GW-BASIC under DOS 3.2. COMPSIZE was developed in 1985.

Description: ELM is a simple computational tool for modeling the steady-state thermal-hydraulics of propellant flow through fuel element coolant channels in nuclear thermal rockets. Written for the nuclear propulsion project of the Space Exploration Initiative, ELM evaluates the various heat transfer coefficient and friction factor correlations available for turbulent pipe flow with heat addition. In the past, these correlations were found in different reactor analysis codes, but now comparisons are possible within one program. The logic of ELM is based on the one-dimensional conservation of energy in combination with Newton's Law of Cooling to determine the bulk flow temperature and the wall temperature across a control volume. Since the control volume is an incremental length of tube, the corresponding pressure drop is determined by application of the Law of Conservation of Momentum. The size, speed, and accuracy of ELM make it a simple tool for use in fuel element parametric studies. ELM is a machine independent program written in FORTRAN 77. It has been successfully compiled on an IBM PC compatible running MS-DOS using Lahey FORTRAN 77, a DEC VAX series computer running VMS, and a Sun4 series computer running SunOS UNIX. ELM requires 565K of RAM under SunOS 4.1, 360K of RAM under VMS 5.4, and 406K of RAM under MS-DOS. Because this program is machine independent, no executable is provided on the distribution media. The standard distribution medium for ELM is one 5.25 inch 360K MS-DOS format diskette. ELM was developed in 1991. DEC, VAX, and VMS are trademarks of Digital Equipment Corporation. Sun4 and SunOS are trademarks of Sun Microsystems, Inc. IBM PC is a registered trademark of International Business Machines. MS-DOS is a registered trademark of Microsoft Corporation.

Description: The Jet Propulsion Laboratory Contamination Analysis Program (CAP) is a generalized transient executive analysis computer code which solves realistic mass transport problems in the free molecular flow environment. These transport problems involve mass flux from surface source emission and re-emission, venting, and engine emission. CAP solution capability allows for one-bounce mass reflections if required. CAP was developed to solve thin-film contamination problems in the free molecular flow environment, the intent being to provide a powerful analytic tool for evaluating spacecraft contamination problems. The solution procedure uses an enclosure method based on a lumped-parameter multinodal approach with mass exchange between nodes. Transient solutions are computed by the finite difference Euler method. First-order rate theory is used to represent surface emission and reemission (user care must be taken to insure the problem is appropriate for such behavior), and all surface emission and reflections are assumed diffuse. CAP does not include the effects of post-deposition chemistry or interaction with the ambient atmosphere. CAP reads in a model represented by a multiple-block data stream. CAP allows the user to edit the input data stream and stack sequential editing operations (or cases) in order to make complex changes in behavior (surface temperatures, engine start-up and shut-down, etc.) in a single run if desired. The eight data blocks which make up the input data stream consist of problem control parameters, nodal data (area, temperature, mass, etc.), engine or vent distribution factors (based upon plume definitions), geometric configuration factors (diffuse surface emission), surface capture coefficient tables, source emission rate constant tables, reemission rate constant tables, and partial node to body collapse capability (for deposition rates only). The user must generate this data stream, since neither the problem-specific geometric relationships, the constituents involved, nor plume distribution functions are a part of CAP. Instead, these are used to generate the data stream model CAP solves. Outputs vary from individual deposition rates of exchange, on an internodal basis and on a constituent basis as a function of time, to deposition on each surface on a constituent basis as a function of time. The type of outputs may be user-specified by control parameters. CAP allows the user to select output intervals within the solution interval and to generate restart nodal data blocks. CAP is composed of several FORTRAN subroutines which serve specific functions and can be easily edited. The code is relatively small (2152 statements), and contains comment statements for all operations. It is written in relatively generic FORTRAN to be adaptable to a variety of computers. CAP was implemented on a DEC VAX 11/780 computer, and is distributed on a 9-track DEC VAX BACKUP format magnetic tape. Virtual memory required is 4.6 MB, which corresponds to a 900 node model capacity. CAP was originally developed under contract for NASA/Goddard Space Flight Center in 1979 by JPL, and was subsequently modified as required for project support at JPL. CAP is a copyrighted work with all copyright vested in NASA.

Description: A robust and efficient technique is discussed for surface-grid generation on a general curvilinear surface. This technique is based on a nonuniform parameter space and allows for the generation of surface grids on highly skewed and nonuniform spaced background surface-grids. This method has been successfully integrated into the GRIDGEN software system.

Description: JTMIX was written to allow the prediction of both ideal and realistic properties of mixed gases in the 65-80K temperature range. It allows mixed gas J-T analysis for any fluid combination of neon, nitrogen, various hydrocarbons, argon, oxygen, carbon monoxide, carbon dioxide, and hydrogen sulfide. When used in conjunction with the NIST computer program DDMIX, JTMIX has accurately predicted order-of-magnitude increases in J-T cooling capacities when various hydrocarbons are added to nitrogen, and it predicts nitrogen normal boiling point depressions to as low as 60K when neon is added. JTMIX searches for heat exchanger "pinch points" that can result from insolubility of various components in each other. These points result in numerical solutions that cannot exist. The length of the heat exchanger is searched for such points and, if they exist, the user is warned and the temperatures and heat exchanger effectiveness are corrected to provide a real solution. JTMIX gives very good correlation (within data accuracy) to mixed gas data published by the USSR and data taken by APD for the U.S. Naval Weapons Lab. Data taken at JPL also confirms JTMIX for all cases tested. JTMIX is written in Turbo C for IBM PC compatible computers running MS-DOS. The National Institute of Standards and Technology's (NIST, Gaithersburg, MD, 301-975-2208) computer code DDMIX is required to provide mixed-fluid enthalpy data which is input into JTMIX. The standard distribution medium for this program is a 5.25 inch 360K MS-DOS format diskette. JTMIX was developed in 1991 and is a copyrighted work with all copyright vested in NASA.

Description: This paper introduces a study on an Electromagnetically Levitated Vibration Isolation System (ELVIS) for isolation control of large-scale vibration. This system features no mechanical contact between the isolation table and the installation floor, using a total of four electromagnetic actuators which generate magnetic levitation force in the vertical and horizontal directions. The configuration of the magnet for the vertical direction is designed to prevent any generation of restoring vibratory force in the horizontal direction. The isolation system is set so that vibration control effects due to small earthquakes can be regulated to below 5(gal) versus horizontal vibration levels of the installation floor of up t 25(gal), and those in the horizontal relative displacement of up to 30 (mm) between the floor and levitated isolation table. In particular, studies on the relative displacement between the installation floor and the levitated isolation table have been made for vibration control in the horizontal direction. In case of small-scale earthquakes (Taft wave scaled: max. 25 gal), the present system has been confirmed to achieve a vibration isolation to a level below 5 gal. The vibration transmission ratio of below 1/10 has been achieved versus continuous micro-vibration (approx. one gal) in the horizontal direction on the installation floor.

Description: The objective of this study is to experimentally determine an empirical model of the vibrational dynamics of the Spacecraft COntrol Laboratory Experiment (SCOLE) facility. The first two flexible modes of this test article are identified using a linear least-square identification procedure and the data utilized for this procedure are obtained by exciting the structure from a quiescent state with torque wheels. The time history data of rate gyro sensors and accelerometers due to excitation and after excitation in terms of free-decay are used in the parameter estimation of the vibrational model. The free-decay portion of the data is analyzed using the Discrete Fourier transform to determine the optimal model order to use in modelling the response. Linear least-square analysis is then used to select the parameters that best fit the output of an Autoregressive (AR) model to the data. The control effectiveness of the torque wheels is then determined using the excitation portion of the test data, again using linear least squares.

Description: A new approach to find homogeneous models for beam-like repeated flexible structures is proposed which conceptually involves two steps. The first step involves the approximation of 3-D non-homogeneous model by a 1-D periodic beam model. The structure is modeled as a 3-D non-homogeneous continuum. The displacement field is approximated by Taylor series expansion. Then, the cross sectional mass and stiffness matrices are obtained by energy equivalence using their additive properties. Due to the repeated nature of the flexible bodies, the mass, and stiffness matrices are also periodic. This procedure is systematic and requires less dynamics detail. The first step involves the homogenization from a 1-D periodic beam model to a 1-D homogeneous beam model. The periodic beam model is homogenized into an equivalent homogeneous beam model using the additive property of compliance along the generic axis. The major departure from previous approaches in literature is using compliance instead of stiffness in homogenization. An obvious justification is that the stiffness is additive at each cross section but not along the generic axis. The homogenized model preserves many properties of the original periodic model.

Description: This report will discuss the design of a liquid hydrogen fuel tank constructed from composite materials. The focus of this report is to recommend a design for a fuel tank which will be able to withstand all static and dynamic forces during manned flight. Areas of study for the design include material selection, material structural analysis, heat transfer, thermal expansion, and liquid hydrogen diffusion. A structural analysis FORTRAN program was developed for analyzing the buckling and yield characteristics of the tank. A thermal analysis Excel spreadsheet was created to determine a specific material thickness which will minimize heat transfer through the wall of the tank. The total mass of the tank was determined by the combination of both structural and thermal analyses. The report concludes with the recommendation of a layered material tank construction. The designed system will include exterior insulation, combination of metal and organize composite matrices and honeycomb.

Description: The Hydroblast Research Cell at Marshall Space Flight Center is used to investigate the use of high pressure waterjets to strip paint, grease, adhesive and thermal spray coatings from various substrates. Current methods of cleaning often use ozone depleting chemicals (ODC) such as chlorinated solvents. High pressure waterjet cleaning has proven to be a viable alternative to the use of solvents. A popular method of waterjet cleaning involves the use of a rotating, multijet, high pressure water nozzle which is robotically controlled. This method enables rapid cleaning of a large area, but problems such as incomplete coverage and damage to the substrate from the waterjet have been observed. This report summarizes research consisting of identifying and investigating the basic properties of rotating, multijet, high pressure water nozzles, and how particular designs and modes of operation affect such things as stripping rate, standoff distance and completeness of coverage. The study involved computer simulations, an extensive literature review, and experimental studies of different nozzle designs.

Description: The purpose of this project was to develop a simple model which could be used to study residual stress. The mechanism that results in residual stresses in the welding process starts with the deposition of molten weld metal which heats the immediately adjacent material. After solidification of weld material, normal thermal shrinkage is resisted by the adjacent, cooler material. When the thermal strain exceeds the elastic strain corresponding to the yield point stress, the stress level is limited by this value, which decreases with increasing temperature. Cooling then causes elastic unloading which is restrained by the adjoining material. Permanent plastic strain occurs, and tension is caused in the region immediately adjacent to the weld material. Compression arises in the metal farther from the weld in order to maintain overall static equilibrium. Subsequent repair welds may add to the level of residual stresses. The level of residual stress is related to the onset of fracture during welding. Thus, it is of great importance to be able to predict the level of residual stresses remaining after a weld procedure, and to determine the factors, such as weld speed, temperature, direction, and number of passes, which may affect the magnitude of remaining residual stress. It was hoped to use traditional analytical modeling techniques so that it would be easier to comprehend the effect of these variables on the resulting stress. This approach was chosen in place of finite element methods so as to facilitate the understanding of the physical processes. The accuracy of the results was checked with some existing experimental studies giving residual stress levels found from x-ray diffraction measurements.

Description: All long-duration spacecraft in low-earth-orbit are subject to high speed impacts by meteoroids and pieces of orbital debris. The threat of damage from such impacts is a significant design consideration in the development of long duration earth-orbiting spacecraft. This report presents the results of a study whose objective was to develop an empirical model to predict the magnitude of the various cracking and through-hole creation phenomena accompanying a habitable module penetration. The significance of the work performed is that the model predictions can be fed directly into a survivability analysis to determine whether or not module unzipping would occur under a specific set of impact conditions. The likelihood of module unzipping over a structure's lifetime can also be determined in such an analysis. In addition, effective hole size predictions can be used as part of a survivability analysis to determine the time available for module evacuation prior to the onset of incapacitation due to air loss. Some of the phenomena considered include maximum petal length, maximum tip-to-tip crack distance, depth of petal deformation, number of cracks formed, orientation of the maximum tip-to-tip distance with respect to the inner wall grain direction, and the effective inner wall hole diameter.

Description: A new model for the two-point vector stream function correlation has been developed using tensor invariant arguments and evaluated by the comparison of model predictions with DNS data for incompressible homogeneous turbulent shear flow. This two-point vector stream function model correlation can then be used to calculate the two-point velocity correlation function and other quantities useful in turbulence modeling. The model assumes that the two-point vector stream function correlation can be written in terms of the separation vector and a new tensor function that depends only on the magnitude of the separation vector. The model has a single free model coefficient, which has been chosen by comparison with the DNS data. The relative error of the model predictions of the two-point vector stream function correlation is only a few percent for a broad range of the model coefficient. Predictions of the derivatives of this correlation, which are of interest in turbulence modeling, may not be this accurate.

Description: A non-local model is presented for approximating the pressure diffusion in calculations of turbulent free shear and boundary layer flows. It is based on the solution of an elliptic relaxation equation which enables local diffusion sources to be distributed over lengths of the order of the integral scale. The pressure diffusion model was implemented in a boundary layer code within the framework of turbulence models based on both the kappa-epsilon-(bar)upsilon(exp 2) system of equations and the full Reynolds stress equations. Model computations were performed for mixing layers and boundary layer flows. In each case, the pressure diffusion model enabled the well-known free-stream edge singularity problem to be eliminated. There was little effect on near-wall properties. Computed results agreed very well with experimental and DNS data for the mean flow velocity, the turbulent kinetic energy, and the skin-friction coefficient.

Description: The structure of the separated flow region over a 20 degree compression corner at a free-stream Mach number of 2.84 is investigated computationally using a Reynolds averaged Navier Stokes (R.A.N.S.) solver and kappa-epsilon model. At this Mach number and ramp angle, a steady-state recirculation region of order delta(sub o) is observed, with onset of a 'plateau' in the wall pressure distribution near the corner. At lower ramp angles, separation is negligible, while at an angle of 24 degrees, separation regions of length 2 delta(sub o) are expected. Of interest here is the response of the mathematical model to inclusion of the pressure dilatation term for turbulent kinetic energy. Compared with the experimental data of Smits and Muck (1987), steady-state computations show improvement when the pressure dilatation term is included. Unsteady computations, using both unforced and then forced inlet conditions, did not predict the oscillatory motion of the separation bubble as observed in laboratory experiments. An analysis of the separation bubble oscillation and the turbulent boundary layer (T.B.L.) frequencies for this flow suggests that the bubble oscillations are of nearly the same order as the turbulent frequencies, and therefore difficult for the model to separate and resolve.

Description: A direct numerical approach has been developed to simulate supersonic turbulent boundary layers. The mean flow quantities are obtained by solving the parabolized Reynolds-averaged Navier-Stokes equations (globally). Fluctuating quantities are computed locally with a temporal direct numerical simulation approach, in which nonparallel effects of boundary layers are partially modeled. Preliminary numerical results obtained at the free-stream Mach numbers 3, 4.5, and 6 with hot-wall conditions are presented. Approximately 5 million grid points are used in all three cases. The numerical results indicate that compressibility effects on turbulent kinetic energy, in terms of dilatational dissipation and pressure-dilatation correlation, are small. Due to the hot-wall conditions the results show significant low Reynolds number effects and large streamwise streaks. Further simulations with a bigger computational box or a cold-wall condition are desirable.

Description: A major problem in modeling of turbulent supersonic flows is the correct assessment of viscous-inviscid interaction problems. Of particular interest is the interaction of boundary layers with shocks. Present turbulence models give in most cases unsatisfactory results in the region of rapid distortion and in the separation region (if one is present) in particular with regard to mean flow profiles and turbulence quantities. The objective of the present work is the direct numerical simulation of shock boundary layer interaction. This report summarizes the first phase during which a numerical method suitable for this problem has been developed and a computer code has been written and tested.

Description: Since high resolution numerical simulations of turbulent flows, or experiments at high Reynolds numbers, represent a substantial investment in resources, the estimation of the minimum resolution required for the study of a given property has been the subject of continued interest. Different properties require, in general, different resolutions, and the present paper is dedicated to the requirements for the measurement of the probability distribution functions of the velocity gradients and, in particular, of their low order moments. The deviation of these quantities from the values corresponding to a Gaussian distribution was one of the first indications of the presence of Reynolds number-dependent intermittency and has been the object of recent interest as numerical simulations have become able to explore the distribution of gradients in the low Reynolds number range, while new experiments have extended the range to increasingly high Reynolds numbers. We will use progressive filtering of the results of numerical simulations of isotropic turbulence as a model for the effect of a sensor of finite size. The numerical issues will be addressed first to ensure that the simulations are fully resolved from the point of view of the velocity gradients. This will also give us an estimate for the numerical resolution required for the different quantities.

Description: The MSU compact four-wire transverse vorticity probe permits omega(sub z)(t) measurements in a nominally 1 sq mm domain. Note that a conventional coordinate system is used with x and y in the streamwise and normal directions respectively. The purpose of this investigation was to acquire time series data in the same access port at the ceiling of the 80 ft x 120 ft wind tunnel (NASA Ames Research Center) as earlier used by the Wallace group from the University of Maryland and to compare the present results with those of the three-component vorticity probe used in that earlier study.

Description: Boundary layer flow transition has long been suggested as a potential noise source in both marine (sonar-dome self noise) and aeronautical (aircraft cabin noise) applications, owing to the highly transient nature of process. The design of effective noise control strategies relies upon a clear understanding of the source mechanisms associated with the unsteady flow dynamics during transition. Due to formidable mathematical difficulties, theoretical predictions either are limited to early linear and weakly nonlinear stages of transition, or employ acoustic analogy theories based on approximate source field data, often in the form of empirical correlation. In the present work, an approach which combines direct numerical simulation of the source field with the Lighthill acoustic analogy is utilized. This approach takes advantage of the recent advancement in computational capabilities to obtain detailed information about the flow-induced acoustic sources. The transitional boundary layer flow is computed by solving the incompressible Navier-Stokes equations without model assumptions, thus allowing a direct evaluation of the pseudosound as well as source functions, including the Lighthill stress tensor and the wall shear stress. The latter are used for calculating the radiated pressure field based on the Curle-Powell solution of the Lighthill equation. This procedure allows a quantitative assessment of noise source mechanisms and the associated radiation characteristics during transition from primary instability up to the laminar breakdown stage. In particular, one is interested in comparing the roles played by the fluctuating volume Reynolds stress and the wall-shear-stresses, and in identifying specific flow processes and structures that are effective noise generators.

Description: Turbulence modeling continues to be one of the most difficult problems in fluid mechanics. Existing prediction methods are well developed for certain classes of simple equilibrium flows, but are still not entirely satisfactory for a large category of complex non-equilibrium flows found in engineering practice. Direct and large-eddy simulation (LES) approaches have long been believed to have great potential for the accurate prediction of difficult turbulent flows, but the associated computational cost has been prohibitive for practical problems. This remains true for direct simulation but is no longer clear for large-eddy simulation. Advances in computer hardware, numerical methods, and subgrid-scale modeling have made it possible to conduct LES for flows or practical interest at Reynolds numbers in the range of laboratory experiments. The objective of this work is to apply ES and the dynamic subgrid-scale model to the flow of a boundary layer over a concave surface.