MSC 1993 World Users' Conference Proceedings

The conference proceedings for the 1993 World Users' Conference are now available on-line in Adobe Acrobat PDF format. (The Adobe Acrobat Reader software is available for free download from Adobe's web site at www.adobe.com). When printed, these PDF files will produce a better quality image than the one shown on your computer screen. All papers have been categorized by topic.

Categories

• Analysis Methods
• Applications
• Automobile Applications
• Dynamics
• Electromagnetics
• Nonlinear Applications
• Nonlinear Methods
• Optimization
• Pre- and Postprocessing
• Shape Optimization
• Superelements
• Testing and Applications

Analysis Methods

COUPLING OF MSC/NASTRAN AND BEM STRUCTURAL MATRICES (Acrobat 702KB) #0493

M. J. MeNamee -- United Technologies Corporation

K. L. Leung -- United Technologies Corporation

P. B. Zavareh -- United Technologies Corporation

ABSTRACT: Accurate stress analysis using a combination of the Finite Element Method (FEM) and the Boundary Element Method (BEM) is achieved by coupling MSC/NASTRAN with symmetric BEM structural matrices through the use of the external superelement technique. A DMAP procedure is developed to incorporate the symmetric stiffness matrix derived by the BEM into the global stiffness matrix generated by MSC/NASTRAN. This coupled MSC/NASTRAN BEM analysis procedure provides a way to exploit the versatility of MSC/NASTRAN in handling wide classes of engineering problems, including those involving nonlinearity and inhomogeneity, while maintaining the accuracy of the BEM in areas of crack and stress concentration. Several analysis examples are given to illustrate the usefulness of the proposed technique.

EXACT CALCULATION OF MINIMUM MARGIN OF SAFETY FOR FREQUENCY RESPONSE ANALYSIS STRESS RESULTS USING YIELDING OR FAILURE THEORIES (Acrobat 465KB) #0593

Françis Charron -- Université de Sherbrooke

ABSTRACT: In static analysis, the calculation of minimum margins of safety using yielding (Von Mises,...) or failure theories (instability for honeycomb structure,...) requires all stress components (3D case : sigma xx, sigma yy, sigma zz, tau xy, tau xz, tau yz ; both magnitude and sign) for a specific element. In frequency response analysis, the stress component magnitude and sign are functions of the reference phase angle and the phase angle of each of the various stress components. When the phase angle difference between the various stress components is almost equal to 0 or 180 degrees, the calculation of the minimum margin of safety is simple. However, in the general case, the minimum margin of safety will be dependent upon both the reference phase angle, as well as the phase angle of each of the various stress components. This paper describes a method used for the calculation of the exact minimum margin of safety for the general case. For the 2D and 1D elements, the exact minimum margin of safety is evaluated at the lower and upper fibers of the element where the flexural stress is maximum and the transverse shear stress contribution is equal to zero. The calculation of the exact minimum margin of safety is done by a general stress processor using the MSC/NASTRAN OUTPUT2 file.

MSC/NASTRAN SHELL AND SOLID ELEMENT MESH REQUIREMENTS IN THE VICINITY OF A CIRCULAR HOLE STRESS CONCENTRATION (Acrobat 1.04MB) #0393

Robert P. Thacker, Jr. -- Boeing Computer Support Services

ABSTRACT: This paper takes a look at the mesh density requirements when the geometry is a flat plate, has a circular hole, and is subjected to a uniform tension along two opposite edges. The program to be used is MSC/NASTRAN, Version 67, a finite element program from the MSC.Software Corporation. The elements used are the 8 node parabolic shell, 8 node linear solid, and the 20 node parabolic solid. The objective is to determine the number of elements, in the radial and angular directions, to achieve an accuracy in deflection and in stress of less than 2% error. The mesh density to achieve the 2% error criteria will be checked on all nodes along 3 straight lines, at angles 0, 45, and 90 degrees, from the hole to the edge of the plate and also all nodes around the circumference of the hole.

PARALLEL PROCESSING IN MSC/NASTRAN (Acrobat 500KB) #0293

Dr. Louis Komzsik -- The MSC.Software Corporation

ABSTRACT: The MSC.Software Corporation has been researching parallel computational methods and evaluating their applicability to its products since 1985. Limited parallelism has been offered in various MSC/NASTRAN products, mainly on supercomputers, since 1987. Presently, there are 5 major computer platforms where shared memory parallel execution is supported.

The paper will discuss some of the technical details of the shared memory parallel methodology and explore the limitations of parallel speedup on the current parallel environments using MSC/NASTRAN. Results of the leading parallel applications will be shown on a moderate number of processors. Specifically, a 4 fold parallel speedup on 16 CPUs analyzing a large automobile industry job will be demonstrated.

The results of investigations on using a distributed memory methodology (applied in the massively parallel computers) will also be discussed. The main strategic aspect of Lagrange multiplier-based solution sequences and the topic of supermodules will be briefly addressed.

Applications

COMPUTATION OF STRESS INTENSITY FACTORS USING MSC/PROBE VERSION 5 (Acrobat 1.16MB) #7693

John E. Schiermeier -- The MSC.Software Corporation

ABSTRACT: In fracture mechanics, the stress intensity factor is used to determine whether a crack will run, possibly causing catastrophic failure or arrest. Typically, this value can be computed from the stress or displacement fields around the crack tip, either by hand or by numerical methods, and then compared with empirical data.

MSC/PROBE-PLANAR has long had two methods of computing Mode l and II stress intensity factors which take advantage of the p-version, the contour integral and cutoff function methods, as well as the standard energy release rate. In Version 5, MSC/PROBE-SOLID has incorporated singularity elements to model exactly the displacement field for a closed crack, and the crack-opening displacement (COD) method to automatically compute Mode I, II, and III stress intensity factors. Combined with the automatic p-adaptivity, accurate and reliable factors may be computed for fully three-dimensional problems in an efficient manner.

This paper provides an explanation of the stress intensity factors and the methods used to compute them. Sample problems are run, using MSC/XL V3B as the pre- and postprocessor, and the computed stress intensity factors are compared with theoretical results where available.

DESIGN AND ANALYSIS OF CORIOLIS MASS FLOWMETERS USING MSC/NASTRAN (Acrobat 833KB) #5493

C. P. Stack -- Micro Motion, Inc.

T. J. Cunningham -- Micro Motion, Inc.

ABSTRACT: A Coriolis mass flowmeter measures mass flow directly by vibrating a fluid-conveying pipe at resonance. Design of these meters was facilitated by solving the equations of motion via the finite element method. A complex eigenvalue analysis (SOL107) was necessary due to the presence of the Coriolis force terms in the damping matrix. Performance parameters of the meter can be predicted from the complex eigenvectors and were found to match experimentally determined values to within five percent.

Since vibration occurs during Coriolis meter operation, low stress designs must be used to preclude fatigue failure. This analysis was also performed using MSC/NASTRAN.

DETERMINING TUBE STRESS FROM CBEND ELEMENT FORCES AND MOMENTS (Acrobat 740KB) #6493

Raymond N. Frick, Ph.D., P.E. -- Pratt & Whitney

ABSTRACT: MSC/NASTRAN is used extensively in the design of external tubing for turbo-fan aircraft engines at Pratt & Whitney. It accurately calculates the stress of tubes under pressure, thermal, and case displacements and also natural frequencies. Many of the external tubes are small diameter (under 3/4 inch) and are part of a complex tube system. The most effective element type for these tubes is a "beam" element such as the CBEND. A complex small diameter tube system modeled with CBEND elements is very efficient compared with the same system modeled with CQUAD4 plate elements. However, while the MSC/NASTRAN CBEND element uses the ASME Code equations to account for the ovalization of the tube in the bends, the stress output is not complete. The in-plane and out-of-plane bending moments are not combined while the torque stress and the hoop stress are ignored. Therefore, the correct principal stresses are not determined. Pratt & Whitney developed a CBEND post-processor which uses the ASME Code equations to determine the complete stress field from the MSC/NASTRAN calculated forces and moments. This paper presents the ASME Code equations used by the CBEND post-processor and compares the results to equivalent plate models. Based upon these comparisons, the use of the MSC/NASTRAN CBEND element has been implemented in the design of small diameter tubes.

DYNAMIC MODELING AND ANALYSIS OF SPINNING POLYGON ASSEMBLIES USING MSC/NASTRAN (Acrobat 438KB) #6693

Bill Nowak -- Xerox Corporation

Courtney James -- Xerox Corporation

ABSTRACT: This paper describes the application of MSC/NASTRAN to calculate the dynamic response of a spinning polygon, motor, and motor housing used in xerographic printers. Initially, different levels of model sophistication were evaluated showing convergence to a representative model. Using the converged model, the dynamic response of the polygon mirror was evaluated from the effects of base excitation. Frequency correlation in the 0 to 1000 hertz range was demonstrated to be better than 10% when compared to a dynamic modal test. Mechanical gain correlation at the fundamental rotor resonance was of the same order as the modal test with amplitude variations attributed to the assumed damping of the model and differences between empirical and analytical response locations. Conclusions and recommendations on future work are also cited.

ELECTRO-MECHANICAL RESPONSE SIMULATION OF ELECTROSTATIC VOLTMETERS USING MSC/NASTRAN (Acrobat 469KB) #6593

Bill Nowak -- Xerox Corporation

ABSTRACT: This paper describes the application of MSC/NASTRAN to simulate the electro-mechanical behavior of in situ electrostatic voltmeters. These "tuning fork" style transducers are used to continuously monitor photoconductor voltage during xerographic copier and printer operation. Voltmeter theory of operation is discussed, and finite element model development is detailed showing a 6% correlation of natural frequency with empirical results. From an existence proof, model boundary conditions were adjusted to show correct voltmeter dynamic response. Based on this empirical behavior, structural design modifications were made to the model until similar dynamic response was analytically achieved. These modifications were then applied to the hardware and correct performance was empirically verified. Conclusions and recommendations on future work are also cited.

EXTENDED DETAILED FINITE ELEMENT ANALYSIS OF A 9.6 METRE AUTOGENOUS SAG MILL (Acrobat 1.52MB) #5393

John Hadaway -- ANI PRODUCTS

Eric Hecht -- COMPUMOD Pty. Ltd.

ABSTRACT: Finite element analysis is becoming an integral part of the design and manufacturing process of heavy engineering machinery for the mining industry in Australia. The Grinding Mill division of Australian National Industries (ANI) which is the division of ANI responsible for design, manufacture and supply of ore grinding mills to mines throughout Australia and overseas have been using the MSC/NASTRAN finite element analysis code extensively in all phases in the supply of their mills.

The finite element analyses of a 9600 mm diameter (5640 mm length) Autogenous Grinding Mill is used to describe how today's available technology is being used by one company to design and manufacture machinery for the mining industry.

FINITE ELEMENT ANALYSIS OF THE RISER COLUMN CHAINTABLE AND CHAINHAWSE STRUCTURES FOR A FLOATING PRODUCTION STORAGE AND OFFLOADI NG FACILITY USING MSC/NASTRAN (Acrobat 1.92MB) #4393

D. J. Twine -- BHP Engineering Pty. Ltd.
Dr. S. Loganathan -- BHP Engineering Pty Ltd

ABSTRACT: An overview is given of a recently completed finite element stress analysis of the Riser Column Chaintable and Chainhawse Structures for an offshore Floating Production Storage and Offloading (FPSO) oil and gas facility.

The project involved finite element modeling of a symmetric half model of the chaintable and analysis of the model under a series of Unit Load Cases and Load Combinations. The primary objective of the analysis was to quantify stress levels and stress combinations throughout the structure to enable strength and fatigue capacity of the chaintable and chainhawse structures to be confirmed.

The model involved approximately 60,000 DOF and analysis was carried out using a CRAY-YMP Supercomputer.

The paper will present a summary description of the problem and objectives of the analysis, finite element discretisation of the structure, analysis approach, and quality assurance checking procedures applied to verify the results.

FINITE ELEMENT ANALYSIS (FEA) OF THE STATIC AND DYNAMIC RESPONSE OF THE SHEET MOLD COMPOUNDING (SMC) STRUCTURE USING THE ENHANCED MODELING TECHNIQUE OF THE ADHESIVE JOINING REGION (Acrobat 589KB) #6393

Joon B. Park -- Automated Analysis Corporation

Randal H. Visintainer -- Ford Motor Company

ABSTRACT: An FEA study was conducted to examine the static and dynamic response of an SMC structure fabricated by adhesive joining method using MSC/NASTRAN. The appropriate modeling of the adhesive joining region is essential in precise FEA predictions of static and dynamic response of the SMC structure presented. An enhanced modeling technique for the adhesive joining region is presented in this paper. A composite mechanics approach was used to define the mechanical property of finite elements along adhesive joining region. Static and dynamic predictions of the enhanced modeling technique were compared with predictions of the conventional linkage modeling method in structure analysis.

A FINITE ELEMENT METHODOLOGY FOR PREDICTING RELATIVE MOTION OF AVIONICS MODULE CONNECTOR CONTACTS (Acrobat 99KB) #4193

Mark H. Morton -- Lockheed Aeronautical Systems Company

David S. Layton -- Lockheed Aeronautical Systems Company

ABSTRACT: This paper presents a methodology for modeling and predicting electronic module connector displacements in an advanced avionics system. The system is modeled using finite element theory and the equations of motion solved using MSC/NASTRAN. Emphasis in this paper is placed on the finite element modeling (FEM) approach and reduction scheme. In order to achieve the desired accuracy in analysis, the initial FEM contained over 20,000 degrees of freedom. A FEM of this size is typically cost prohibitive to use and sensitive to numerical instabilities, particularly when the frequency range of interest may be as low as 10 Hertz and exceed 1000 Hertz. Several model reduction schemes and various superelement approaches are examined in an effort to reduce model size and improve numerical stability. Results include guidelines for model reduction of similar type structures, as well as selection of the appropriate eigenvalue solver and associated parameters when using MSC/NASTRAN. Of particular interest to the authors was the random response of the connector. To this end, modal analysis is employed to identify natural modes of vibration. Forced frequency response analysis using the large mass method provides transfer functions between the source excitation and the response at the connector. The transfer functions, along with a user specified input power spectral density function, are used as input to a random analysis. Results include rms displacements and frequencies of the connectors.

MANEUVERING OF THE SPACE STATION/ORBITER DURING AN ASSEMBLY FLIGHT (Acrobat 1.54MB) #3093

Paul A. Cooper -- NASA Langley Research Center

Alan E Stockwell -- Lockheed Engineering & Sciences Company

Shih-Chin Wu -- Lockheed Engineering & Sciences Company

ABSTRACT: A large-angle, multi-body, dynamic modeling capability was developed to help validate numerical simulations of the dynamic motion and control forces which occur while berthing Space Station Freedom to the Shuttle Orbiter during early assembly flights. The paper describes the dynamics and control of the station, the attached Shuttle Remote Manipulator System, and the Orbiter during a maneuver from a gravity-gradient attitude to a torque equilibrium attitude using the station reaction control jets. The influence of the elastic behavior of the station and of the remote manipulator system on the attitude control of the station/Orbiter system during the maneuver is investigated. The flexibility of the station and the arm had only a minor influence on the attitude control of the system during the maneuver.

MODELING AND ANALYSIS OF 4-STEP 3-D CARTESIAN BRAIDED COMPOSITES (Acrobat 1.04MB) #7593

Soheil Mohajerjasbi -- Boeing Defense & Space Group

ABSTRACT: The fiber architecture of the preform produced in a 4-Step (l x1) 3-D Cartesian braiding process is investigated based on a study of the movement of the fiber carriers on the machine bed. Distinctly different fiber architectures are identified for the interior, boundary, and corner regions of the preform and the composite. Since different fiber architectures will result in different deformation properties, the effective deformation behavior of the composite is expected to be the result of contributions from these different stiffness properties.

In contrast with these findings, some of the present analytical models consider a "unit cell" as a repeat unit for the braided composite, and attempt to model the mechanical behavior of the composite from the properties of this unit cell. This unit cell is in the form of a parallelpiped with yarns connecting between the opposite corners along the body diagonals.

In this paper, a finite element based method is proposed for modeling the structure of the 3-D braided composite, and determining the elastic constants and coefficients of thermal expansion. MSC/NASTRAN is used in modeling the thermoelastic properties of the composite. Estimates of elastic constants and coefficients of thermal expansion are developed as a function of "interior braiding angle".

Among advantages of this technique are simplicity, and the ability to model and study the response of complex shapes subject to complex loads applied at the boundary.

N-250 WINDSHIELD FINITE ELEMENT MODEL (Acrobat 338KB) #7893

Masduki Suwandi -- Nusantara Aircraft Industry Ltd.

Kurnia Witono -- Nusantara Aircraft Industry Ltd.

Nurdin -- Nusantara Aircraft Industry Ltd.

ABSTRACT: In analyzing the windshield of N-250, - Indonesian / IPTN's latest aircraft product - one major issue is how to model the windshield glass panel and the distribution of the load to its mounting structure. This paper concentrates on this issue.

NON-LINEAR ANALYSIS OF A PROPELLER BLADE RETENTION SYSTEM (Acrobat 480KB) #3293

John C. Lambert -- United Technologies Corporation

ABSTRACT: A primary problem in the design of aircraft propeller systems is the prediction of blade natural frequencies. A major problem in predicting these frequencies is determination of the blade retention stiffness. This stiffness is difficult to determine due to the many linear and non-linear variables involved in such systems. This paper discusses the manner in which MSC/NASTRAN V65B was used to advance Hamilton Standard's methodology in evaluating blade retention stiffness.

RADIATION HEAT TRANSFER WITH SPECTRAL SURFACE BEHAVIOR (Acrobat 199KB) #5593

Mike Chainyk -- The MSC.Software Corporation

ABSTRACT: Prior to Version 68, MSC/NASTRAN radiation exchange was restricted to ideal black/grey opaque surfaces exhibiting diffuse emission, absorption, and reflection characteristics. Most real surface behavior is considerably more complicated due to varying degrees of specular, spectral, and temperature dependent properties, as well as radiant transmission. Generalized numerical treatment of all of these phenomena simultaneously is beyond the scope of the current Version 68 effort, however, most materials in a practical engineering sense may be characterized by one or two dominant surface conditions. In particular, most solar collection device materials as well as high temperature metals can be satisfactorily described in terms of their spectral and temperature dependent surface properties. This problem class is addressed with a method known as the radiation energy-band approximation.

A REFINED METHOD FOR LIVE-LOAD DISTRIBUTION PREDICTION OF BRIDGES AND COMPARATIVE STUDY (Acrobat 678KB) #5293

Yohchia Chen, Ph.D., P.E. -- The Pennsylvania State University

ABSTRACT: A refined analysis method is proposed for predicting the distribution of vehicle live loads on bridge girders. An effective and efficient iteration scheme is used to solve the nonlinear equations. Two representative bridge systems are investigated. The obtained results from the proposed method are compared to experimental data and those obtained from other analysis methods. The prediction method for live-load distribution implemented in the current bridge design code is carefully examined. The paper concludes with a number of actual bridge examples and recommendations.

RESPONSE OF PROTECTIVE STRUCTURES TO INTERNAL EXPLOSIONS WITH BLAST VENTING (Acrobat 1.77MB) #4293

Y. Kivity -- The MSC.Software Company

C. Florie -- The MSC.Software Company

H. Lenselink -- The MSC.Software Company

ABSTRACT: This paper presents a computational study of the response of generic protective structures to internal blast waves from high explosive charges. The computations are carried out with the three-dimensional program MSC/DYTRAN, with explicit treatment of the fluid-structure interactions inherent to the problem. The modeled generic structures include frangible panels for blast venting and internal partitions for blast wave deflection. The structural description includes both a thin shell approach for thin walled containers and a solid finite element representation for concrete type structures. The flow of the detonation products and the ambient air is described employing an Arbitrary-Lagrange-Euler (ALE) approach. This approach also allows internal partitions to be attached to the ALE mesh without degrading the computational efficiency.

STRUCTURAL ANALYSIS OF AN 85 FT. FREE STANDING TOWER (Acrobat 946KB) #3393

Paul F. Martin -- Massachusetts Institute of Technology

ABSTRACT: An 85' high rectangular truss type antenna support tower is currently in use. The tower bracing scheme is typical of and similar to other towers used for this application. Tower members were designed using conventional analytical techniques as presented in design codes such as AISC, ANSI/A58.1 and ANSI/EIA-222. Resulting stress analysis showed factors of safety for all members to be satisfactory and capable of sustaining design loads. Since a conventional tower bracing scheme was used, the overall buckling capacity of this tower was assumed to be adequate. Recently a new antenna configuration proposed for use required additional structural analysis of this tower. The resulting analysis included an overall buckling analysis utilizing the MSC/NASTRAN program with the buckling solution sequence. The buckling analysis revealed that the tower was incapable of withstanding the design loads for either the original or the proposed antenna configurations. In conclusion, it should also be noted that the results of this study suggest that other towers currently in use may also be inadequately designed and subject to potential failures.

THERMOMECHANICAL FINITE ELEMENT ANALYSIS OF STIFFENED, UNSYMMETRIC COMPOSITE PANELS WITH TWO DIMENSIONAL MODELS (Acrobat 1.53MB) #3193

Craig S. Collier, P.E. -- Lockheed Engineering and Sciences Co., NASA Langley Research Center

Kevin A. Spoth -- Lockheed Engineering and Sciences Co., NASA Langley Research Center

Glenn C. Grassi -- The MSC.Software Corporation

ABSTRACT: A method is presented for formulating stiffness terms and thermal coefficients of stiffened, fiber-reinforced composite stiffened panels for input to finite element analysis (FEA). The method is robust enough to handle panels with general cross sectional shapes, including those hich are unsymmetric or unbalanced. New thermal coefficients are introduced to quantify panel response from through-the-thickness temperature gradients. Equations are defined for stiffness, thermal expansion, and thermal bending that consider the full complement of membrane, bending, and membrane-bending coupling. A technique of implementing this capability with a single plane of shell finite elements using the MSC/NASTRAN² FEA program is revealed. Thermomechanical analyses of an unsymmetric, hat stiffened, metal matrix composite panel are shown to demonstrate the accuracy possible with planar, 2-D FEM's. 3-D FEA results are presented to verify the solutions. Ultimately, the significance of including this additional accuracy in smeared, equivalent plate 2-D models is proved with FEA of an aerospace plane.

VIRTUAL MASS OF FLUID IN EGG-SHAPED DIGESTERS (Acrobat 316KB) #7793

Atis A. Liepins -- Simpson Gumpertz & Heger Inc.

Hamid Nazemi -- Simpson Gumpertz & Heger Inc.

ABSTRACT: The MFLUID capability in MSC/NASTRAN-WS is used to calculate the virtual mass of fluid in an egg-shaped digester tank. In earthquake response calculations of this type of tank a finite element analysis of fluid/structure interaction is needed because simplified methods, such as for cylindrical tanks, are not available. Described are observations about the performance of the QUAD4 element for fluid/structure interaction and a verification of the virtual mass matrix. Useful enhancements are proposed.

Automobile Applications

THE APPLICATION OF FEM-EMA CORRELATION AND VALIDATION TECHNIQUES ON A BODY-IN-WHITE (Acrobat 1.05MB) #0693

Marc Brughmans -- LMS International

ABSTRACT: The paper reviews the application of FEM - EMA correlation and validation techniques to a body-in-white, namely the 1991 GM Saturn four door Sedan. The FEM model of this car consisted of 46830 dof's (half model). A multi-point experimental modal analysis (EMA) survey was executed for 360 response dof's. Classical techniques for correlation analysis such as MAC are applied. The paper introduces, as well, a variation of the MAC calculation that enables a better identification of regions of difference between FEM and EMA. Error localization methods have been applied to identify the regions of the FEM model causing most of the discrepancies between FEM and EMA. An FEM model updating procedure was executed to reduce the difference between FEM and EMA to acceptable limits.

BUMPER DESIGN USING COMPUTER SIMULATION (Acrobat 387KB) #1093

Hwa-Won Lee -- SsangYong Motor Company

Sung-Kuk Jang -- SsangYong Motor Company

ABSTRACT: In order to meet the current safety standards, it is necessary that a series of destructive tests for new vehicles including automotive bumpers to be performed. These tests are very expensive and time consuming. Therefore, the necessity of economical design and analysis using finite element method is increasing day by day.

This paper attempts to present such a design and analysis method, using relatively simple beam model and fine meshed shell model. The analysis has been performed for center - center pendulum and barrier hits of a bumper system of XENOY² 1102 newly developed by GEP (General Electric Plastics US) under the process of injection molding.

Finally, the results predicted from finite element method are compared with those of experimental tests to evaluate the analysis procedure.

DEVELOPMENT OF A METHODOLOGY TO PREDICT THE ROAD NOISE PERFORMANCE CHARACTERISTICS (Acrobat 329KB) #0993

Mark P. Voutyras -- Chrysler Corporation

William F. Resh -- Chrysler Corporation

Kevin R. Thomson -- Chrysler Corporation

ABSTRACT: This paper describes the development of a methodology for predicting the road noise performance characteristics of an automotive vehicle system. An MSC/NASTRAN finite element model of a complete suspension system was constructed and analyzed for dynamic response. In addition, a customized system simulation software package, GRADAM, was developed for combining the FEA and experimental results in order to assess vehicle structural sensitivity to noise and also to study the effect of suspension design modifications on the interior noise levels. In the model development stage, the FEM modal and frequency response results for all the relevant suspension components were validated by comparing with the corresponding experimental measurements. The interior noise levels were then obtained through the system simulation software. This customized software combines the MSC/NASTRAN force output at suspension-to-body attachment points with the corresponding pressure/force (P/F) experimental data in order to predict the interior noise levels. The methodology developed herein permits the noise pressure levels to be determined for desired frequency domains. Furthermore, this methodology allows design engineers to answer "what if" questions in order to evaluate the effect of suspension design changes on the interior noise levels. It is anticipated that the methodology presented herein will be instrumental in optimizing the noise and vibration (NVH) characteristics of future car lines.

FLEXIBLE VEHICLE SIMULATION OR MODELING VEHICLE SUSPENSION COMPLIANCE AT FORD MOTOR CO. USING A COUPLING OF ADAMS AND MSC/NASTRAN (Acrobat 712KB) #0793

Shawn P. McGuan -- Mechanical Dynamics, Inc.

Steve Pintar -- Ford Motor Company

ABSTRACT: The Core Development Technology group at the Ford Motor Company is actively involved in correlating analytical (ADAMS) vehicle models with objective (telemetric) data from instrumented vehicles. In the past, vehicle compliances were introduced into the ADAMS model by using the ADAMS elements; TIREs, nonlinear BUSHINGs and BEAMs. With the advent of ADAMS/FEA^TM, a data translator which provides a two-way interface between ADAMS and MSC/NASTRAN, vehicle models may now include the effects of geometrically complex component stiffness and total body compliance in the ADAMS full vehicle simulation. This paper examines the effects of these added compliances on an ADAMS vehicle model by comparing the dynamic toe and camber angles of a vehicle with rigid upper and lower control arms to a vehicle with flexible upper and lower control arms built from MSC/NASTRAN data. The results demonstrate that complex problems can be very efficiently modeled and simulated by combining finite element analysis (MSC/NASTRAN) and mechanical system simulation (ADAMS) technologies.

USING MSC/NASTRAN FOR THE CORRELATION OF EXPERIMENTAL MODAL MODELS FOR AUTOMOTIVE POWERTRAIN STRUCTURES (Acrobat 879KB) #0893

L.D. Isley -- Borg Warner Automotive

W.R. Kelley -- Borg Warner Automotive

ABSTRACT: This paper describes the methods and analyses used in the correlation of experimentally determined modal models to analytical solutions. The paper illustrates the significant benefits of Component Mode Synthesis and Design Sensitivity Analysis options available in MSC/NASTRAN in the correlation process. Neutral File Interface programs and remote orthogonality calculation methods that simplify the communication between the FEA solution and the laboratory results are also discussed. Three (3) specific experimental procedures formulated the basis for the paper. The first, a simple free-free beam case is presented. The second and third are examples of more complex automotive transmission assemblies with special boundary condition issues and internal component influences. Specific recommendations for improving correlation potential both in laboratory test methods and in the finite element modeling task have been provided.

Dynamics

APPLICATION OF APPROXIMATE TECHNIQUES IN THE ESTIMATION OF EIGENVALUE QUALITY (Acrobat 582KB) #1193

B. P. Wang -- The University of Texas at Arlington

S. P. Caldwell -- McDonnell Douglas Corporation

ABSTRACT: An eigenvalue quality estimate has been implemented in MSC/NASTRAN. The quality estimate is based on the eigenvalue difference from a lumped and consistent mass matrix formulation. This difference represents the error associated with the discretization of the finite element model. Normally two eigensolutions are required to compute the error estimate. However, several approximate solution techniques have been provided to efficiently compute the consistent mass matrix eigenvalues. The eigenvalue quality estimator has been implemented as a set of Direct Matrix Abstraction Programming (DMAP) alters to SOL 103 (SEMODES) of MSC/NASTRAN Version 67. Several numerical examples are provided to demonstrate the method.

APPLICATION OF MSC/DYNA IN HEAD IMPACT ANALYSIS (Acrobat 205KB) #7093

Vasudeva Murthy -- Johnson Controls

Han Sankara -- Johnson Controls

ABSTRACT: The Federal Motor Vehicle Safety Standard 201 (FMVSS 201) specifies requirements to afford impact protection for occupants. It provides (see reference 1) that "when that area of the seat back that is impacted in accordance with S3.22 by a 15 pound, 6.5 inches diameter head-form at a relative velocity of 15 mph the deceleration of the head-form shall not exceed 80 g's for more than 3 ms". This paper presents the results of a head impact simulation on an automotive seat using MSC/DYNA. The results were correlated with those from an actual laboratory test. Good correlation was achieved in the deceleration time history as well as deformation of the structure. Though simulations of the crash type have been performed on entire vehicles few have been focussed on the seat to the level of detail here.

DEVELOPMENT OF A CALCULATION PROCEDURE INCLUDING FLUID STRUCTURE COUPLING TO ASSESS CAVITATION EFFECTS (Acrobat 982KB) #6993

Gotthard Ph. Rainer -- AVL - Company for Internal Combustion Engines and Instrumentation Ltd.

Ernst Gschweitl -- AVL - Company for Internal Combustion Engines and Instrumentation Ltd.

Hans Christoph Raffel -- AVL - Company for Internal Combustion Engines and Instrumentation Ltd.

ABSTRACT: Between cylinder liner and parent bore of internal combustion engines cooling water flow is existing.

Due to the operating conditions cavitation effected by

• dynamic acoustic excitations and
• fluid flow

may occur.

AVL is investigating both effects using the MSC/NASTRAN acoustic element to consider the dynamic acoustic effected cavitation and the AVL developed CFD-software FIRE for the fluid flow effects.

This paper considers only the dynamic acoustic effect. Here the calculational and experimental work is explained, which has been done till now to apply MSC/NASTRAN to cavitation calculation. In addition to this, the difficulties to per form measurements and to compare the results out of calculation and of measurements are described.

The MAC (Modal Accuracy Criterion) is used as a tool to compare measurement and calculation. The calculations are carried out as eigenmode analysis and as forced vibrations.

AN EQUIVALENT LINEARIZATION SOLUTION SEQUENCE FOR MSC/NASTRAN (Acrobat 786KB) #3593

J.H. Robinson -- NASA Langley Research Center

C.K. Chiang -- Old Dominion University

ABSTRACT: A classical equivalent linearization solution procedure for the geometric nonlinear random response of structures is incorporated into MSC/NASTRAN by Direct Matrix Abstraction Programming (DMAP). The equivalent linearization solution sequence was derived from the existing Super Element Modal Frequency (SEMFREQ) response solution sequence. The definition of the equivalent linear stiffness matrix in terms of the MSC/NASTRAN differential stiffness for Gaussian random loads is presented. The required modification and inclusions to the SEMFREQ solution sequence are discussed. Results are presented for the nonlinear random response of a simple and a complex panel.

EULER BUCKLING (Acrobat 434KB) #6793

Kevin E. Arden -- Newport News Shipbuilding

ABSTRACT: This paper is intended to investigate the accuracy of MSC/NASTRAN's Solution 105 for use in calculating linear elastic (Euler) buckling modes.

Column buckling, panel buckling, and stiffened panel buckling is analyzed using Euler equations and Solution 105. Comparisons and modeling recommendations are made for each type of structure.

Solution 105 provides excellent results for Euler type buckling. Panel buckling requires the use of an adequate number of elements. A convergence plot shows that four QUAD4 elements per half sine wave are necessary for accurate results.

The opinions expressed herein are those of the author and do not necessarily reflect those of Newport News Shipbuilding and Dry Dock Company.

IDENTIFICATION OF CRITICAL SPEEDS OF ROTORS ATTACHED TO FLEXIBLE SUPPORTS (Acrobat 333KB) #3493

David Bella -- The MSC.Software Corporation

Hans Hartmueller -- BMW Rolls-Royce GmbH

Dr. Karsten Muehlenfeld -- BMW Rolls-Royce GmbH

Dr. Gabriel Tokar -- BMW Rolls-Royce GmbH

ABSTRACT: It is common practice to include the gyroscopic terms for rigid rotors into the equations of motion for the calculation of critical speeds. This procedure works well for the case in which the rotors are connected directly to ground by elastic and/or damping elements. All calculated eigenfrequencies are critical speeds of the rotor. If additional degrees of freedom are included to model the actual support structure, then not all eigenfrequencies are critical speeds. Many of the calculated eigenfrequencies are simply modes of the support structure. The procedure presented allows critical speeds to be filtered from the set eigenfrequencies calculated when the support structure is included in the analysis. The methodology is used to determine the critical speeds of the BRR 700 series aerojet.

PSDTOOL - A DMAP ENHANCEMENT TO HARMONIC/RANDOM RESPONSE ANALYSIS IN MSC/NASTRAN (Acrobat 1.12MB) #3693

Alwar Parthasarathy -- Advanced Technology Systems Group, Spar Aerospace Limited

Mohamed Elzeki -- Advanced Technology Systems Group, Spar Aerospace Limited

Vivianne Abramovici -- Advanced Technology Systems Group, Spar Aerospace Limited

ABSTRACT: This Paper presents a DMAP framework called PSDTOOL that significantly enhances the scope of harmonic/random response analysis in MSC/NASTRAN. PSDTOOL addresses the case where grid-point accelerations (and displacements) form the primary output of a dynamic analysis and locates response-maxima within and across subcases and frequencies. In the common aerospace application where the input-excitation is specified as a base-motion PSD-spectrum in the form of a scalar or a diagonal matrix, PSDTOOL outputs the RMS (and RSS) responses at all the grid-points of a finite-element model as a single OLOAD-type table and locates peak RMS responses over the whole system or over individual user-defined subsystems. For facilitating high-level system design and integration, PSDTOOL includes a novel capability for computing mass-weighted average- responses at 'centers-of-mass' of system and subsystems. The goal is to automate and condense output-generation and to produce design-data with speed and confidence. PSDTOOL offers some incidental conveniences such as a conventional relative-motion formulation of base-excitation analysis and output-screening in a common co-ordinate system. The DMAP scheme is implemented as an ALTER in a DBLOCATE-restart from a mode-extraction database and the various special features are driven by non-standard PARAMeters. Extension to element-output is being envisioned.

SOLVING MODAL EQUATIONS OF MOTION WITH INITIAL CONDITIONS USING MSC/NASTRAN DMAP, PART 1: IMPLEMENTING EXACT MODE SUPERPOSITION (Acrobat 533KB) #1293

Ayman A. Abdallah -- Analex Corporation

Alan R. Barnett -- Analex Corporation

Omar M. Ibrahim -- Analex Corporation

Richard T. Manella -- National Aeronautics and Space Administration

ABSTRACT: Within the MSC/NASTRAN DMAP module TRDl , solving physical (coupled) or modal (uncoupled) transient equations of motion is performed using the Newmark-Beta or mode superposition algorithms, respectively. For equations of motion with initial conditions, only the Newmark-Beta integration routine has been available in MSC/NASTRAN solution sequences for solving physical systems and in custom DMAP sequences or Alters for solving modal systems. In some cases, one difficulty with using the Newmark-Beta method is that the process of selecting suitable integration time steps for obtaining acceptable results is lengthy. In addition, when very small step sizes are required, a large amount of time can be spent integrating the equations of motion. For certain aerospace applications, a significant time savings can be realized when the equations of motion are solved using an exact integration routine instead of the Newmark-Beta numerical algorithm. In order to solve modal equations of motion with initial conditions and take advantage of efficiencies gained when using uncoupled solution algorithms (like that within TRD1), an exact mode superposition method using MSC/NASTRAN DMAP has been developed and successfully implemented as an enhancement to an existing coupled loads methodology at the NASA Lewis Research Center.

SOLVING MODAL EQUATIONS OF MOTION WITH INITIAL CONDITIONS USING MSC/NASTRAN DMAP, PART 2: COUPLED VS. UNCOUPLED INTEGRATION (Acrobat 581KB) #1393

Alan R. Barnett -- Analex Corporation

Omar M. Ibrahim -- Analex Corporation

Ayman A. Abdallah -- Analex Corporation

Timothy L. Sullivan -- National Aeronautics and Space Administration

ABSTRACT: By utilizing MSC/NASTRAN DMAP in an existing NASA Lewis Research Center coupled loads methodology, solving modal equations of motion with initial conditions is possible using either coupled (Newmark-Beta) or uncoupled (exact mode superposition) integration available within module TRD1. Both the coupled and newly developed exact mode superposition methods have been used to perform transient analyses of various space systerns. However, experience has shown that in most cases, significant time savings are realized when the equations of motion are integrated using the uncoupled solver instead of the coupled solver. Through the results of a real-world engineering analysis, advantages of using the exact mode superposition methodology are illustrated.

SPACE STATION SOLAR ARRAY POINTING SYSTEM CONTROL/STRUCTURE INTERACTION STUDY USING CO-ST-IN FOR MODAL REDUCTION (Acrobat 976KB) #6893

Tarun Ghosh -- Rockwell International Corporation

Benigno Muniz Jr. -- Rockwell International Corporation

Joseph Cheng -- Rockwell International Corporation

Marsha Nall -- NASA Lewis Research Center

ABSTRACT: The control/structure interaction problem of orienting the Space Station Freedom (SSF) PhotoVoltaic arrays is solved to achieve desired system pointing performance using the Beta Gimbal Drive Mechanism. The vibration modes of the on-orbit SSF Stage Configuration 17 are calculated using MSC/NASTRAN finite element models which presently comprise a total of 250,000 degrees of freedom. In-house Direct Matrix Abstraction Programs and post-processors are developed for more efficient and accurate Craig-Bampton modal reduction with geometric stiffening and either modal displacement or modal acceleration data recovery. Structural Dynamic Research Corporation's CO-ST-IN post-processor is used to rank the vibration modes for control system analysis. The problem of solving an actual case of ranking modes using CO ST-IN for large-scale SSF application is illustrated. Examples of calculated control system response including the effects of reduced flexible mode dynamics are shown.

SSF FLEXIBLE MULTI-BODY CONTROL/STRUCTURE INTERACTION SIMULATION (Acrobat 711KB) #1593

Timothy T. Cao -- NASA JSC/ES

Huang-Tsang Chang -- DYNACS Engineering Co.

Tuyen Hua -- NASA JSC/EG2

ABSTRACT: For a large structure with the size of Space Station Freedom, it is important to consider the structural flexibity in the dynamics and control simulation. Conventionally, structural flexibility is obtained from the MSC/NASTRAN structural models. The interface between MSC/NASTRAN structural models and multi-body simulation programs is achieved by a flex-preprocessor. This approach becomes formidable as the size of finite element models grows with the evolution of SSF structures. The use of superelement and Craig -Bampton reduction in MSC/NASTRAN introduces difficulty in interfacing MSC/NASTRAN structural models with the Station/Orbiter Multi-Flex- Body Berthing Analysis Tool (SOMBAT), a program developed at JSC to perform flex-multibody dynamics and control simulation. To handle the dynamic reduction approach in MSC/NASTRAN, the flex-preprocessor in SOMBAT has been modified, and appropriate DMAP sequences have been developed. This paper presents the enhanced capability of SOMBAT for the space station freedom dynamics and control simulation. The procedure is validated through control structure interaction simulation for space station stage two scenarios.

A SUPEREFFICIENT, MSC/NASTRAN-INTERFACED COMPUTER CODE SYSTEM FOR DYNAMIC RESPONSE ANALYSIS OF NONPROPORTIONALLY DAMPED ELASTIC SYSTEMS (Acrobat 748KB) #1493

Rong C. Shieh -- MU, Inc.

ABSTRACT: With the emphasis on frequency response analysis case, development of the title computer code capability and application of the latter in evaluation of the computational efficiency of the MSC/NASTRAN code itself in the dynamic structural response analysis of nonproportionally . damped elastic systems are made in this study. In this system, MSC/NASTRAN is used mainly for physical or modal structural (mass, damping, and stiffness) matrix assembling. The newly developed CMODEAN (Complex MOde/DEcoupling ANalysis) module uses the structural matrices as input for complex normal modes (state eigenmodes) calculation and equations of motion decoupling. Computational efficiency of CMODSTAN over MSC/NASTRAN for frequency response analysis of nonproportionally damped systems is demonstrated by an example problem with 225 dynamic degrees of freedom.

Electromagnetics

EDDY CURRENT SIMULATIONS FOR THE SSCL LOW ENERGY BOOSTER CAVITY (Acrobat 420KB) #1693

Y Goren -- Superconducting Super Collider Laboratory

L. Walling -- Superconducting Super Collider Laboratory

ABSTRACT: Eddy currents are developed in the tuner of the Superconducting Super Collider Low Energy Booster (LEB) cavity during the LEB frequency sweep. The two main difficulties created by the eddy currents are excessive tuner-surface heating, and more important, a reduction in the time response of the tuner. We present a detailed analysis of the eddy currents for various tuner designs. The analysis has been done using 2D and 3D time-domain finite element codes: PE2D by Vector- Field and MSC/EMAS by the MSC.Software Corporation. Non-linear analysis was performed utilizing B-H curves. The codes have been bench marked analytically and by using measured data for different slotted pillbox structures.

INDUCTION FURNACES: INTEGRATED MSC/EMAS, MSC/XL, AND MSC/NASTRAN ANALYSIS (Acrobat 873KB) #2093

Vern D. Overbye -- STM Consulting, Inc.

Glenn H. Stalker -- STM Consulting, Inc.

Scott M. Schaefer -- STM Consulting, Inc.

Roy H. VanderHeiden -- STM Consulting, Inc.

S. Rajendran -- Mobil Solar Energy Corporation

ABSTRACT: A generic induction-heated high temperature furnace is modeled using MSC/EMAS to determine power loss density distribution in a 3-D segment with a cylindrical coil surrounding the furnace. MSC/XL is used to display color plots of furnace induced power loss density as well as magnetic flux density within furnace and coil components. MSC/XL is next used to generate internal heat generation records for an MSC/NASTRAN nonlinear steady state thermal analysis using a full radiation matrix (assuming radiation shields closely spaced). An alternative to full radiation analysis is presented using a theory developed for radiation - equivalent convection. The two methods are compared, and furnace temperature contours are presented using MSC/XL. Results of a C-language program are presented to calculate heat flux and display the results graphically. Finally, the authors present several recommendations for forthcoming releases of MSC/NASTRAN to improve user friendliness.

MAGNETIC SENSOR DEVELOPMENT USING FINITE ELEMENT ANALYSIS (Acrobat 990KB) #1893

T.J. Wagener -- Honeywell Systems and Research Center

James R. Campbell, Jr. -- Honeywell Systems and Research Center

James E. Lenz -- Honeywell Systems and Research Center

ABSTRACT: Magnetic finite element analysis has proven to be valuable in the research and development of magnetic sensors and actuators. The emphasis of this research is not entirely on design detail, but also on the basic magnetics and material properties on which devices are based. Examples of this research include: 1) magnetic shielding for ring-laser gyros, 2) magnetic force modeling for accelerometers based on the type-two high-temperature superconductors, 3) proximity and position sensors for aircraft applications, and 4) the dynamics of and losses in electromagnetic actuators and bearings. These examples represent the wide diversity of magnetic problems faced by researchers, and demonstrate how solutions to these problems range from simple two-dimensional models to larger and more complicated three-dimensional models.

MSC/EMAS NONLINEAR TRANSIENT ANALYSES OF MULTITURN COILS WITH ATTACHED CIRCUITS (Acrobat 622KB) #2293

John R. Brauer -- The MSC.Software Corporation

Charles R. Figer, Jr. -- The MSC.Software Corporation

ABSTRACT: MSC/EMAS can be used to model multiturn nonlinear coils with attached circuits. The circuits are modeled by 0D finite elements, which can be attached to 1D line elements that represent multiturn windings. The 1D elements can then be MPCd to the 3D or 2D finite element model containing nonlinear magnetic material. The complete finite element model is analyzed by nonlinear transient Solution 305. Example computations of time-varying currents and electromagnetic fields are shown for an inductor, a transformer, and an induction motor. The induction motor solution is shown to be aided by the new ADAPT nonlinear transient method that is available in Version 2.5.

3D NONLINEAR MAGNETIZATION STUDIES OF THIN AMORPHOUS RIBBONS (Acrobat 558KB) #1993

Richard L. Copeland -- Sensormatic Electronics Corporation

Markus B. Kopp -- Sensormatic Electronics Corporation

ABSTRACT: Three-Dimensional analysis of a thin amorphous magnetic ribbon is studied using MSC/EMAS. A uniform static magnetic field is applied at various angles with respect to a nonlinear isotropic magnetic ribbon by using HSURF loads at certain boundary grid points. Different finite element meshes were examined in terms of generating a very uniform magnetic field to a region in the model center representing the thin amorphous ribbon. A simple uniform mesh produced a more uniform field than a mesh with much more refinement at the center. The importance of the proper boundary conditions along with the HSURF loads are discussed. The results clearly demonstrate the expected closure field around the ribbon and help in understanding the magnetization distributions in and around thin magnetic regions under an applied uniform field.

USING FINITE ELEMENT SOFTWARE TO PREDICT EMC PERFORMANCE FROM ELECTRICALLY SMALL SOURCES (Acrobat 555KB) #1793

ABSTRACT: A major limit of the finite element method in the past has been its relative difficulty to model three-dimensional open configurations found in EMC problems. MSC/EMAS with its open boundary elements now enables problems to be solved in three-dimensions that were previously considered solvable only with Method of Moment (MoM) codes and other techniques. MSC/EMAS is used here to accurately compute radiation fields from an electrically small source (less than a wavelength) that was previously solved with method of moments codes, and then to analyze printed circuit radiation that cannot easily be predicted by MoM.

USING MSC/EMAS IN STIMULATING A PARALLEL MICROSTRIP TRANSMISSION LINE (Acrobat 298KB) #2193

Markus B. Kopp -- Sensormatic Electronics Corp.

Vichate Ungvichian, Ph.D. -- Florida Atlantic University

Richard L. Copeland, Ph.D. -- Sensormatic Electronics Corp.

ABSTRACT: During a crosstalk study of two parallel and coplanar microstrip lines having identical width and thickness, the MSC/EMAS code was used to calculate the total current flows on the receptor line, and demonstrate the relationship between substrate height and crosstalk. The frequency used in the simulations was 200 MHz. One microstrip trace was driven by a 100 mA ideal current source and terminated with a 50 ohm resistive load. The other trace was terminated at both ends by two 50 ohm loads.

MSC/EMAS produced results that are accurate to within 4 percent of the closed form solution. It also clearly demonstrated that the magnitude of crosstalk, as a function of substrate height, asymptotically approaches a final value.

Nonlinear Applications

LARGE DEFORMATION HYPERELASTIC ANALYSIS IN MSC/NASTRAN VERSION 67.5 (Acrobat 586KB) #5693

Katerina-D. P. Papoulia -- The MacNeal Schwendler Corporation

Steve S. Hsieh -- The MacNeal Schwendler Corporation

ABSTRACT: Version 67.5 of MSC/NASTRAN includes finite deformation analysis for problems that involve large strain and large rotation. The material law is Green-elastic (hyperelastic) with a strain energy function of the generalized Rivlin type, extended to include the effect of compressibility at the nearly incompressible limit. The stress-strain relations are discussed in some detail as well as the approach taken to avoid the occurrence of volumetric locking. Examples are presented that illustrate the capabilities of the formulation to model problems with large strain and large rotation.

NONLINEAR SEISMIC ANALYSIS OF BRIDGES: PRACTICAL APPROACH AND COMPARATIVE STUDY (Acrobat 721KB) #5793

Yohchia Chen, Ph.D., P.E. -- Pennsylvania State University

ABSTRACT: A simplified numerical model with an efficient computational scheme is proposed for nonlinear seismic analysis of bridges. The results obtained from the simplified model are compared to those from the refined model and other methods. The proposed model is shown to be especially effective for obtaining maximum responses, and is practical and economical. Effects of bridge skews on responses are also carefully examined. The paper concludes with a number of bridge examples and design recommendations.

NONLINEAR SUPERELEMENT ANALYSIS TO MODEL ASSEMBLY PROCESSES (Acrobat 1.09MB) #5993

Mark J. Stone -- Eastman Kodak Company

Vic Genberg -- Eastman Kodak Company

ABSTRACT: It is often desirable to know the residual effect of assembly processes on the final product. Although only a linear static analysis may be desired, the changes in loads and constraints throughout the assembly process prevent the use of a simple linear solution. If the product is fairly complex, it is often necessary to use a large finite element model. This paper describes the use of a single nonlinear solution (MSC/NASTRAN V66A SOL 66) using superelements to analyze a large model subjected to varying loads and constraints. As the load and support change in each subcase, the structural deflection changes by adding to or subtracting from the previous deformation state. The model used represents the High Resolution Mirror Assembly (HRMA) for NASA's Advanced X-Ray Astrophysics Facility (AXAF). The process represented is the alignment and assembly of the AXAF mirrors to the support structure.

SNAP-THROUGH BUCKLING ANALYSIS OF A SHALLOW GEODESIC DOME USING MSC/NASTRAN (Acrobat 758KB) #5893

S. Loganathan, BSc Eng Civil (Hons), M. Eng, PhD -- BHP Engineering

R.C. Morgan, BE, M.Eng Sc MIE Aust -- BHP Engineering

ABSTRACT: This paper illustrates the nonlinear analytical and experimental study of a shallow geodesic dome comprising thin walled circular hollow sections. A 156-member shallow geodesic dome that has a rise to span ratio of 1:10 (i.e. a rise of 0.6m to span of 6.0m) was constructed and tested experimentally. The present investigation is focused to study the snap-through phenomena of the dome subjected to a static load at the centre of the dome. The dome is discretized as 156 beam elements and its perimeter supports are assumed as ideal pin supports. Displacement control of the load point was employed to trace snap-through and this prevents any possible dynamic jump in the vicinity of the snap-through region. Experimental observation shows that the members near the loading point deformed severely under the applied load without resulting in any damage to the welded joints, apart from material yielding in certain members.

MSC/NASTRAN Version 67 installed on the BHP Research CRAY-YMP (EL) SuperComputer was used as the numerical tool to validate the complex nonlinear behaviour of the dome. The results of geometrical and material nonlinear analysis of the dome from MSC/NASTRAN compare well with the experimental results for cases where the displacements can be reasonably measured with the linear transducers employed. A second nonlinear large displacement analysis was carried out with MSC/NASTRAN on a Schwedler dome in which the connection details are based on the Harley Spaceframe structures. The Schwedler dome considered has a rise-to-span ratio of 1:4 (i.e., a rise of 2.5 m to span of 10.0 m).

Nonlinear Methods

IMPROVEMENTS IN LINEAR BUCKLING AND GEOMETRIC NONLINEAR ANALYSIS FOR MSC/NASTRAN'S SHELL ELEMENTS (Acrobat 543KB) #4993

Claus C. Hoff -- The MSC.Software Corporation

ABSTRACT: This paper presents an improved approach for linear buckling and geometric nonlinear analysis in MSC/NASTRAN. The differential stiffness and the internal forces of the QUAD4 and TRIA3 shell elements have been corrected in MSC/NASTRAN Version 68. The linear stiffness of the shell elements has not been changed. With the corrections in Version 68, two major capabilities have been improved. The eigenproblem in linear buckling analysis of thin shells is free of spurious modes which have been observed in Version 67.5 and earlier Versions. Furthermore, the shell elements converge better in geometric nonlinear analysis. The theoretical concept of the corotational formulation in MSC/NASTRAN is summarized briefly. The corrections of the internal forces and tangent stiffness are explained. Examples are presented which illustrate the improved behavior of the shell elements in Version 68 as compared to Version 67.5.

NONLINEAR ANALYSIS USING A MODAL BASED REDUCTION TECHNIQUE (Acrobat 530KB) #5193

D. Shalev -- Israel Aircraft Industries

A. Unger -- Israel Aircraft Industries

ABSTRACT: This paper presents a solution to nonlinear formulated problems using eigenfunctions computed by a linear free vibration solution. The system of equations is extremely reduced. The solution is unique in its formulation as the governing equations represent the problem continuously and do not require an iterational or incremental solution. Energy consideration is used and the Ritz method is applied to render the governing equations. An integrated system was built in which the current analysis functioned as a MSC/NASTRAN dummy module integrated with MSC/NASTRAN SOL 3 and SOL 24 to render the mode shapes and geometrical and material properties respectively. Several numerical examples are presented and compared to solutions from the literature.

NONLINEAR GAP-TYPE SOLUTIONS USING A LINEAR F.E.A. CODE (Acrobat 733KB) #5093

R. D. Hilton -- Bell Helicopter Textron

ABSTRACT: The most common form of structural Finite Element Analysis (FEA) is the linear static solution, in which the behavior of each element can be characterized as a linear equation. Linear static FEA cannot be used for problems with nonlinear gap-type elements, as their load vs. deflection behavior cannot be expressed with a single linear equation. Examples of gap-type elements include a cable (an axial element which can transfer tension between its ends, but not compression) and a bearing contact (two interfering surfaces that can compress against each other, but do not adhere when separated). For a gap element, the load vs. deflection equation depends on the sense and magnitude of deflection each loading condition imposes on the element.

Many common FEA codes do not support gap elements; for those that do, adding a gap element complicates the solution by requiring extensive changes to the linear model, and by increasing the CPU time required (often several times over). As the gap behavior can vary from one loading condition to the next, a separate solution for each condition must be obtained.

The Enforced Strain Method uses an approach in which a compensating enforced strain is used to give linear elements gap-like load vs. deflection behavior. The technique can be used with linear FEA codes that do not support gap elements, or can be used as an alternate solution for gap-capable codes. Benefits of the method are reduced CPU requirements, the ability to run multiple loading cases, and no need for superelements.

The Enforced Strain Method is a more efficient gap solution, particularly when a given model has a relatively small proportion of gaps, and when multiple loading conditions are required. An example problem is presented in which the required CPU time was reduced by 43% as compared to the fastest MSC/NASTRAN gap solution. Though presented as a program external to MSC/NASTRAN, the method could be implemented through DMAP alters to the standard linear static solution. Run as a DMAP, CPU time savings for the example problem would have increased from 43 to 66% as compared to MSC/NASTRAN's nonlinear gap Solution 66.

THREE DIMENSIONAL SLIDELINE CONTACT (Acrobat 938KB) #4893

Rakesh Allahabadi -- The MSC.Software Corporation

ABSTRACT: Slideline contacts in MSC/NASTRAN model the separation and sliding of finite amplitude between three dimensional deforming bodies. The modeling of contact requires the user to specify slideline planes in which the interaction can occur. The sliding between bodies occur along lines, specified by lists of grid points, within the slideline planes. The bodies can have large relative motions within the slideline planes. However, relative motions outside the slideline planes are ignored; therefore, they must be small compared to a typical slideline element.

The finite element formulation for both contact and friction is based on the penalty method. But, the user need not specify the penalty values as the program automatically calculates them. The slide line element force vector and the stiffness matrix are derived from a variational principle using a consistent linearization procedure. The formulation is capable of modeling nonlinear contact geometry and inelastic material behavior including large deformation.

The three dimensional slideline contact is a standard feature in Version 68 for quasi-static analyses (SOL 106).

However, a special DMAP is required for Version 67.5.

Optimization

OPTIMIZATION OF DAMPED STRUCTURES IN THE FREQUENCY DOMAIN (Acrobat 625KB) #2893

Tsung-hsiun Li -- Iowa State University

James E Bernard -- Iowa State University

ABSTRACT: This paper presents a way to efficiently compute the sensitivities of steady state resonant response and discusses the utility of these sensitivities in redesign and optimization. The resonant response sensitivities are calculated by combining the new capabilities of MSC/NASTRAN v67 in SOL 108,111 and DMAP solution sequences. Two examples illustrate the approach.

OPTIMIZATION TRIAL ANALYSIS OF A JOURNAL/THRUST BEARING STRUCTURE (Acrobat 779KB) #2993

Takao Miki -- Mitsubishi Heavy Industries, Ltd.

Mitsuru Kondo -- Mitsubishi Heavy Industries, Ltd.

Fumio Mizuguchi -- Mitsubishi Heavy Industries, Ltd.

Yasuhisa Ogino -- Ryoyu System Engineering, Ltd.

ABSTRACT: Recently, due to the need to minimize structural weight and reduce material cost, several programs are offering optimization capabilities. An optimization capability has been added to MSC/NASTRAN in V66 and has been enhanced in V67. With V68, it will be also possible to optimize the shape of a structure.

This paper presents a trial analysis of optimization capability using the current version (V67) performed on a journal/thrust bearing structure. While supporting the static load and satisfying design constraints on stress and displacement, weight is minimized.

This trial analysis demonstrates the effectiveness of the optimization capability in MSC/NASTRAN in achieving satisfactory results while saving much of the designer time which is currently used in a manual iterative optimization procedure. Improvements such as easiness of use and shape optimization would help to put this capability to extensive use in design.

USING DESIGN SENSITIVITY FOR STATISTICAL RESPONSE ANALYSIS (Acrobat 690KB) #2793

Ken Blakely -- The MSC.Software Corporation

ABSTRACT: Statistical response analysis computes the statistical distribution of responses, given the distribution of design variable values. Variation in design variable values can result from manufacturing tolerances, and it can also be attributed to a level of uncertainty about the finite element input. Statistical distributions discussed herein are the standard deviation and variance.

Statistical response analysis can be accomplished in MSC/NASTRAN by using SOL 200 and DMAP alters in Version 67.5. This paper describes the theory, alters, and examples.

Pre- and Postprocessing

A COMPARISON OF THREE ADAPTIVE REMESHING TECHNIQUES (Acrobat 742KB) #4693

Joseph L. Walsh -- FEGS, Incorporated

ABSTRACT: Adaptive Mesh Refinement and Adaptive Remeshing of finite element meshes based on results to ensure result accuracy is a growing field of interest in the analysis modeling community. This paper will discuss and compare three methods of Adaptive Remeshing used in conjunction with the FAM Analysis Modeling System.

EXPOSING THE MYTHS OF DESIGN TO ANALYSIS DATA EXCHANGE (Acrobat 673KB) #6093

Joseph L. Walsh -- FEGS, Incorporated

ABSTRACT: Design/Analysis integration is a major issue for most companies performing finite element analysis. This paper will outline some of the underlying issues and difficulties along with exposing the deficiencies of several attempts at Design/Analysis integration. This paper will also outline the approach and capabilities of data exchange based on IGES using MSC/XLfromCAD and FAM along with data exchange based on STEP.

"GEOMETRIC" FE ANALYSIS WITH MSC/XLPLUS (Acrobat 937KB) #4493

Celso A. Barcelos -- Aries Technology

ABSTRACT: Historically, finite element modeling has been done independently from the underlying geometry used for design. Finite element models have largely been created using geometry which was manually entered, and to a large part, redundant with the geometry used for design purposes. Lately, advances in solids modeling and FEM software has enabled engineers to achieve significant productivity gains and unparalleled ease of use by bridging the gap between these two disciplines.

JOINING TETRAHEDRA TO HEXAHEDRA (Acrobat 416KB) #4593

Alok Mahajan -- Optimal CAE, Inc

ABSTRACT: In the context of creating finite element mesh from a solid modeler, for example, ARIES/ConceptStation, the user can choose between techniques of mapped meshing and free meshing. Mapped meshing provides greater choice of elements and more control of mesh density. Mapped meshing usually results in the most computationally efficient mesh. Free meshing provides for the most rapid production of mesh, sometimes with great sacrifice in computational efficiency. The best of both worlds would exist, if the user could choose to use mapped meshing in geometrically simple regions and free meshing in geometrically complex regions. Because currently available free meshing algorithms only provide tetrahedra, and because the most efficient (mapped) mesh consists of hexahedra, a methodology is required for joining these two non-conforming elements. This study examines a variety of methods for joining these elements, considers a multiplicity of load conditions, and demonstrates that joining with small error is possible.

SHAPE FUNCTION INTERPOLATION OF 2D AND 3D FINITE ELEMENT RESULTS (Acrobat 679KB) #4793

Victor Genberg -- Commercial & Government Systems, Eastman Kodak Company

ABSTRACT: A finite element program such as MSC/NASTRAN provides displacements, rotations or temperatures at grid points only. There are a variety of applications which require results at other locations in the model:

1) Obtaining displacement boundary conditions on finer resolution breakout models
2) Optical ray tracing on deformed mirror surfaces
3) Placing temperatures from a coarse thermal model on to a finer structural model.


All of these applications involve interpolation of results over the finite element model. This paper describes a general purpose post-processing program to accurately interpolate over a model, using element shape functions. The user may specify a choice of linear interpolation for thermal models and solid elasticity models, or cubic interpolation for plate and shell models.

TOPOLOGY BASED ANALYSIS MODELING AND PARAMETER DRIVEN ANALYSIS MODELS IN FEGS FAM 3.0 (Acrobat 455KB) #6193

Joseph L. Walsh -- FEGS, Incorporated

ABSTRACT: Finite element analysis is being used more and more today to evaluate design alternatives and drive design decisions. The ability to quickly modify an Analysis Model based on design geometry parameters requires a Topology Based Analysis Model and may also require Parameter Driven Analysis Models. This paper will discuss the implementation of Topology Based Analysis Modeling in FEGS FAM3.O Modeling System allowing for multiple geometric representations and assignment of analysis attributes to the topological entities. This paper will also discuss the application of parameter symbols and macros in FAM3.O to achieve Parameter Driven Analysis Models along with user definable extensions to FAM3.O.

VARYING PRESSURE DISTRIBUTION OVER A GROUP OF ELEMENTS USING MSC/XL (Acrobat 436KB) #6293

Pathuri Rama Murthy -- WIPRO Infotech Limited

ABSTRACT: A commonly encountered problem in FEA is applying pressures of linear or non-linear nature, over a group of elements. This paper presents a simple but effective way of specifying these kind of pressures using MSC/XL.

Shape Optimization

AN ALTERNATIVE APPROACH FOR IMPROVING THE ACCURACY OF SEMI-ANALYTIC DESIGN SENSITIVITIES IN SHAPE OPTIMIZATION (Acrobat 510KB) #3993

GaoWen Ye -- The MSC.Software Corporation

Shenghua Zhang -- The MSC.Software Corporation

ABSTRACT: The semi-analytic method has been adopted for shape sensitivity calculations because it does not require special code for analytic derivatives of element matrices. However, this method may have serious accuracy problems for which several approaches have been proposed. These approaches, however, are either lack of generality or difficult to be integrated with a general-purpose FEM package. Here, an effective approach to the problem has been developed in which an iterative algorithm used. It not only improves the accuracy but also provides error estimators so as to ensure the quality of calculated sensitivities. In addition, it can be easily integrated with MSC/NASTRAN In this paper, the basic idea of the approach is first described. Then, a general algorithm based on the approach is given. Finally, its effectiveness is shown through numerical results.

COMPUTER AIDED SHAPE OPTIMIZATION FOR ENGINEERING DESIGN WITH MSC/NASTRAN (Acrobat 566KB) #3793

Shenghua Zhang -- The MSC.Software Corporation

ABSTRACT: The goal of shape optimization is to find a best shape of a structural component so as to minimize an objective function subject to various design constraints including functional and manufacturing constraints. Version 68 in MSC/NASTRAN provides a tool to solve shape design problems systematically and automatically. This paper will first define the general shape optimization problem. Then, a new user interface to generate basis vectors is described. Through several example problems, the paper shows that the general shape optimization capability in MSC/NASTRAN can optimize complex shapes of two and three dimensional engineering components.

FORMULATION OF PARAMETER-BASED SHAPE SENSITIVITIES FROM FREE-FIELD SIZE SENSITIVITIES (Acrobat 911KB) #3893

Afshin Mikaili -- Iowa State University

James Bernard -- Iowa State University

ABSTRACT: Conventional finite-element-based structural shape optimization requires the calculation of grid sensitivities which are much more expensive to compute than size sensitivities. This paper presents an alternate method of shape optimization which is based on using size sensitivities (computed using MSC/NASTRAN, Version 66A) to guide shape redesign. In this method, a thin layer of plate elements is cast on selected free surfaces of structures modeled with solid selected free edges are covered with a thin lining of beam elements. While the performance of the structure remains virtually unaffected by introducing these nearly zero section elements, the sensitivities of the structural response with respect to the thickness of these elements provide qualitative insight on the behavior of the structure as well as a quantitative basis for shape optimization. This paper also addresses the challenge of calculating parameter-based shape sensitivities (e.g., sensitivity with respect to a shaft diameter, or a shaft fillet radius) from the computed free-field size sensitivities. The method is applied to a pin geometry under two different static loading conditions.

Superelements

DATA RECOVERY AND MODEL REDUCTION METHODS FOR LARGE STRUCTURES (Acrobat 761KB) #2393

Hyoung M. Kim -- McDonnell Douglas Aerospace - Space Systems

Theodore J. Bartkowicz -- McDonnell Douglas Aerospace - Space Systems

David A. VanHorn -- McDonnell Douglas Aerospace - Space Systems

ABSTRACT: This paper demonstrates important factors for the application of mode-superposition methods and component mode synthesis to transient response analyses of large structures. A theoretical review is presented and numerical results are evaluated for three case studies. Data recovery techniques based on the mode-superposition method are evaluated with respect to different types of force input, model reduction, model size, and computational resources. Cutoff frequency selection at the component- and system-level of component mode synthesis is discussed for accurate dynamic response calculations. This paper not only shows the theoretical differences between different data recovery methods, but also provides physical insights at each computational stage.

DMAP ALTERS TO APPLY MODAL DAMPING AND OBTAIN DYNAMIC LOADING OUTPUT FOR SUPERELEMENTS (Acrobat 561KB) #2493

Ted L. Rose -- The MSC.Software Corporation

ABSTRACT: A pair of DMAP alters are presented for use in the structured superelement dynamic solutions in MSC/NASTRAN Version 67.5. The first DMAP alter allows the entry of a "modal damping" matrix for superelements by using TABDMP1 entries and the frequencies of the calculated component modes. This allows the use of damping results obtained from dynamic testing of individual components in an assembly analysis.

The second alter implements the calculation and output (including plots and PSD) of the applied dynamic loads on superelements. This provides the user with a tool to verify that dynamic loads applied to superelements were entered properly.

DMAP ALTERS FOR NONLINEAR CRAIG-BAMPTON COMPONENT MODAL SYNTHESIS (Acrobat 1.43MB) #2593

Herand Bedrossian -- Rockwell International Corp.

Ted Rose -- The MSC.Software Corporation

ABSTRACT: A set of DMAP (Direct Matrix Abstraction Program) alters for MSC/NASTRAN have been developed for generating reduced matrices for nonlinear structures using the Craig-Bampton [1] method. These DMAP alters are currently being used in conjunction with the standard MSC/NASTRAN component modal synthesis approach to create a system model for coupled loads analysis of Space Station Freedom (SSF) on-orbit configurations. Nonlinear Craig-Bampton models of the Photo Voltaic (PV) arrays are created and combined with the linear Craig-Bampton models of the various other SSF components using the external superelement approach. In addition, a modal selection DMAP Alter is provided. This alter is used in combination with an external FORTRAN program to select a reduced set of component modes based on the modal strain energy criteria. The procedure is illustrated with an example problem.

IMPLEMENTATION OF THE BLOCK-KRYLOV BOUNDARY FLEXIBILITY METHOD OF COMPONENT SYNTHESIS (Acrobat 917KB) #2693

Kelly S. Carney -- National Aeronautics and Space Administration

Ayman A. Abdallah -- Analex Corporation

Arthur