Lars Olovsson – LSTC
Recent eﬀorts to model fully coupled CFD-airbag processes have motivated the implementation of a gas mixture model in LS-DYNA. The new model, *MAT GAS MIXTURE, allows the mixing of up to eight diﬀerent gases. It is designed to conserve the total energy of the system. Kinetic energy dissipated in the ALE advection process is automatically transformed into heat. The article serves as a description of the gas mixture model and of the accompanying keyword commands.
Janka Cafolla, Roger W. Hall, David P. Norman, Iain J. McGregor – Corus Automotive
Improving the accuracy of virtual prototypes helps to shorten product development times and reduces the number of physical prototypes required. One way in which the accuracy of crash analysis can be improved is to include the effects of forming in the material properties. Corus has developed a three-step “Forming to Crash” process to account for formed properties in crash analysis during both concept and detailed vehicle design stages. The first step uses a selection procedure to identify the parts most sensitive to the inclusion of formed properties. The second step uses an approximate “Forming to Crash” method to rapidly estimate the formed properties and reduce the time taken to conduct the analysis during concept design. The third step links a detailed forming analysis to a crash analysis to provide a full “Forming to Crash” technique for use during detailed design development. This three-step process is used by Corus to support customers in the application of advanced high strength steels. Introduction The process of forming a component changes the properties of the material being used. This is generally ignored in the design and validation process of automotive structures even though the changes in material strength and thickness may be substantial. Finite Element tools are now able to predict the as-formed material properties and use these in subsequent crash analysis. However, although the forming effects on the performance of individual components have been reported in the literature , there are few papers reporting the consequence of including formed properties in full vehicle models . This paper presents Corus developed procedures, which are used to include the results of forming simulations in the full vehicle crash model. Corus “Forming to Crash” Three-Step Process A vehicle body structure consists of hundreds of formed components. A detailed analysis of the stamping process can take between 5 to 10 days per part to complete. In order to minimise the time taken for vehicle analysis, it is therefore important to understand which components in the vehicle body structure are sensitive to forming and how the formed properties affect the vehicle crash performance. The identification of the key parts in which to include formed properties is the first stage of a “Forming to Crash” (F2C) three-step process, which has been developed by Corus.
Yih-Yih Lin – Hewlett-Packard Company
As MPP LS-DYNA uses the message-passing paradigm to obtain parallelism, the elapsed time of an MPP LS-DYNA simulation comprises of two parts: computation cost and communication cost. A quantitative approach for determining the communication cost and, hence, the computation cost and the speedup of an MPP LS-DYNA simulation is presented. Elapsed times, characteristic—latency and bandwidth—of interconnect networks, and message patterns are first measured, and then the method of least square errors is applied to estimate the two costs. This approach allows one to predict the performance, or the speedup, of MPP LS-DYNA simulations with any interconnect network whose characteristics is known. Also, while conducting this performance study of MPP LS-DYNA, loss of accuracy in single-precision (32-bit) MPP LS-DYNA simulations has been found. This finding and the advantage of double-precision (64-bit) arithmetic are presented. INTRODUCTION – Theory for Performance of MPP LS-DYNA To run an N-processor MPP LS-DYNA simulation, or job, an interconnect network, or called simply as interconnect, must first be established to connect the N processors; the collection of the N processors and the interconnect is called an N-processor cluster. In this paper, we will consider only the case that the N processors are of the same kind. For such a job, MPP LS-DYNA starts by decomposing the geometrical configuration of the model into N sub-domains. Each of the N processors is assigned to perform computation on one of the sub-domains; meanwhile, messages are passed among all those processors so that necessary physical conditions, such as 1 2 N force conditions, can be enforced. Let T comput, T comput, …, T comput be each 1 2 N processor’s computation cost, and let T comm, T comm, …, T comm be each processor’s 1 2 communication cost. Define the computation cost Tcomput as max (T comput, T comput, N 1 2 N …, T comput) and the communication cost Tcomm as max(T comm, T comm, …, T comm), respectively. Then the job’s elapsed time can be described as: T elapsed = Tcomput + Tcomm (1) For a given decomposition, the computation cost Tcomput is fixed. In contrast, the communication cost Tcomm varies with the characteristics of interconnects used. The term “speedup” is defined as the ratio T elapsed, 1-processor / T elapsed, N-processor. In general, speedups are smaller than N. Since for the 1-processor job the communication cost Tcomm is zero, the perfect speedup of N folds can be realized only under the unrealistic conditions of zero communication cost, i.e., Tcomm = 0, and perfectly 1 2 N balanced decomposition, which renders T comput = T comput = …= T comput. Assuming that the N processors are of the same kind, the variation of T comput, 2 N T comput, …, T comput arises out of the unbalanced decomposition of the N subdomains. It is extremely difficult to find a universal algorithm to decompose a model with a balanced decomposition. MPP LS-DYNA does provide features, as documented in pfile in parallel specific options, for users to provide hints to get a more balanced decomposition than the default.
Anders Jernberg – Engineering Research Nordic AB, Sweden
There is always a certain amount of springback deformation in a metal forming process. The high strength steel is becoming a more attractive choice for many applications when low weight of the structure is of importance. One major drawback with the high strength steel is that the springback deformation increased compared to low strength steel. All currently known techniques to reduce the springback deformation are used in tool design today, so it can be really hard to reduce the springback deformation further. The only possibility to obtain the desired geometry for the sheet after springback is to have a shape of the tool that is different than the desired final shape of the sheet. A method to modify the tool geometry in order to compensate for the springback effect is here presented. It is a heuristic method, based on the difference between the sheet after springback and the desired shape. No parameterisation of the tool geometry is needed. Since no parameters are describing the shape of the tool geometry, it can be modified in an arbitrary way without the restriction of the design space spanned by design parameter. The method is demonstrated on a doubly curved sheet and it is shown that the method gives a very fast convergence for the desired shape of the sheet. It is most likely that the method will work on a variety of different shapes. The method is applicable regardless of the yield strength of the sheet material, but most interesting are probably those materials which gives a large springback deformation, such as high strength steel and aluminium.
S. Fréchard, A. Lichtenberger, F. Rondot, N. Faderl – Institut Franco-Allemand de Recherches, France, Redjaïmia – Laboratoire de Science et Génie des surfaces, France, M. Adoum – CRIL, France
Quasi-static, quasi-dynamic and dynamic compression tests have been performed on a nitrogen alloyed austenitic stainless steel. This alloy achieves a high hardening modulus and a good ductility at all strain rates. In addition, this steel is very sensitive to strain rate. The temperature sensitivity has been determined for temperatures varying between 20°C and 400°C. Microstructural analysis has been performed on samples subjected to different loads in order to relate the microstructure to the material behaviour. Johnson-Cook and Zerilli-Armstrong models have been selected to fit the experimental data into constitutive equations. These models are unable to reproduce the behaviour of this type of steel over the complete range of tests. A new constitutive model that better fits all the experimental data at different strain, strain rate and temperature has been determined. This empirical model supposes that the influence of the main parameters is independent. Single Taylor impact tests have been realized to validate the models. Live observations of the specimen during impact have been achieved using a special CCD camera set-up. The overall profiles at different times were compared to numerical predictions performed with LS-DYNA.
R.S. Sadeghi, W.A. van der Veen – MSC.Software Corporation
The paper presents a new simulation method for airbags with multiple compartments. In recent years many accidents happened with passengers that were out of position (OOP) when their airbag deployed. Therefore new safety regulations have come into effect in the USA and an airbag now has to meet several requirements that concern out of positioning. In meeting these requirements simulations of deploying airbags are very useful and are widely used. In OOP the airbag hits the passenger when it is still deploying and the pressure field is far from uniform and so there is a complicated interaction between gas flow and the deploying membrane. This calls for an analysis in which the dynamics of the airbag membrane is coupled with a CFD (Computational Fluid Dynamics) solver. To weaken the frontal blow, an airbag with multiple compartments is sometimes used. The idea is to divert the gas stream in the airbag away from the frontal direction into sideways directions. Since these airbags especially need to operate well in OOP setting the CFD approach is required. The paper presents a new method to simulate a multi-compartment airbag in which each compartment is modeled by the CFD approach and each with a separate Euler domain. Before discussing multi-compartment simulations, we present background on simulations for one airbag using one Euler mesh. In these simulations the gas flow is described by the conservation laws of mass, momentum and energy. Also modeling of the interaction between CFD solver and airbag membrane solver is discussed. The conservation laws are applied to the 3D objects that are formed by taking an Euler element and cutting away everything that is outside the airbag surface. The boundary of this 3D object is in general a multi-faceted surface that consists of: (1) Polygons that connect an Euler element to another Euler element and that are called faces. (2) Polygons that connect Euler elements to the airbag surface, and are called polpacks. In the conservation laws several surface integrals occur. Many of these integrals have to do with transport and are computed by running over faces and polpacks. This gives rise to mass being transported across faces. In most cases there is no transport across polpacks since they are part of the airbag surface. For multi-compartment simulations we have to consider multiple Euler domains. Each Euler domain models one airbag. Communication between the airbags proceeds via the holes that connect one airbag with the other. Holes are modeled as a piece of airbag surface that is fully porous. Given that the largest possible timestep depends on the mesh size, each Euler domain could be advanced with a specific timestep. This would require subcycling. For the multi-compartment application mesh sizes are expected to be of comparable magnitude and all Euler meshes are advanced with the same timestep.
Paul A. Du Bois – consulting engineer
The simulation of rubber materials is becoming increasingly important in automotive crashworthiness simulations. Although highly sophisticated material laws are available in LS-DYNA to model rubber parts, the determination of material properties can be non-trivial and time consuming. In many applications, the rubber component is mainly loaded uniaxially at rather high strain rates. In this paper a simplified material model for rubber is presented allowing for a fast generation of input data based on uniaxial static and dynamic test data.
Willem Roux, Nielen Stander – Livermore Software Technology Corporation
System identification of ‘noisy’ structural design optimization problems: the sources of uncertainty, the competing roles of bias and variance, and the interaction of uncertainty and deterministic effects. Two test problems are used to clarify the effect of different approaches.
Happee R., Janssen, A.J., Fraterman E., Monster J.W. – TNO Automotive
The LS-DYNA – MADYMO coupling has been extended so that users gain maximal flexibility in their choice of software and models. Where the traditional coupling allowed only MADYMO ellipsoids and planes to contact entities in the coupled code, now contact can also be defined with MADYMO FE. This enables state of the art MADYMO dummy models, subsystem models, barrier models and human models to be applied in the coupling with LS-DYNA.
Dirk Ulrich – SOFY GmbH, Germany
Since the first applications of Finite Element Analysis (FEA) in automotive design this technology has gone a long way to become a standard tool fully integrated in the design process of automotive industry. Strongly promoted by progress in mathematics, information and computer technology simulation results have today reached a level of accuracy comparable to physical testing. In parallel developments in software technology – ranging from CAD to graphical interactive pre- and postprocessing – have helped FEA to develop from a work performed by specialists into an industrial tool used by design engineers. Within a product design engineers often have to judge the performance of dozens of design variants by doing very similar Finite Element Analysises. This leads to the requirement to automate the preand postprozessing of FEA to reduce errors and turnaround times and to improve quality and productivity. This paper demonstrates how RADE – the Rapid Application Development Environment – of the commercial pre- and postprocessing package SOFY is used in industry to automate analysis processes from input creation to report generation.
Dawn Rintala – Advanced Micro Devices, Inc.
Mordaka J., Gentle C. R. – Nottingham Trent University
Whiplash is the most common soft tissue injury sustained in car accidents. The term is commonly associated with hyperextension of the neck as the head rotates backwards in rear end collisions but the exact injury mechanism is not fully understood because the neck is an anatomically and mechanically complex structure. Experimental studies of the mechanism of injury are limited by several ethical and practical factors, so biomechanical computational simulation, based upon experimental research and mathematical modelling, appears to be the most appropriate method of investigation. During the last decade, significant progress has been made in improving car occupant safety through the use of safety devices, such as airbags and advanced seat belts, as well as the construction of the car body itself. Much still needs to be done, especially for female occupants, because statistically they incur twice the risk of whiplash injury as male car occupants. No simple explanation has so far been found for this difference. It is thought that the anatomic dissimilarity of the sexes is the principal reason, but there are undoubtedly a number of secondary, sociological reasons: women tend to drive smaller cars than men and are more likely to be passengers. The lack of a full explanation arises from the fact that, although there have been several FE-models of the male cervical spine reported, female models are rarely documented. This paper addresses the problem by developing a biomechanical FEM model of the 50th and the 5th percentile female cervical spines, based on the earlier published male model created at the Nottingham Trent University, which relies on grafting a detailed biomechanical model of the neck and head onto a standard HYBRID III dummy model. All numerical analyses have been undertaken using LS-DYNA. Special attention was paid to the behaviour of the scaled down male model in comparison with the model, which included female characteristic features. FEM models of males and females in a representative seat were therefore subjected to 9.5 km/h rear-end simulated collisions and were compared against reported experimental tests. The detailed behaviour varied significantly with gender. The female models revealed greater and earlier peak horizontal acceleration of the head and smaller peak relative extension than the male models. It was concluded that the presented FE models were reasonably in accordance with available crash data on instrumented volunteers in terms of head motion.
Matthias Eick, Dr. Lars Fredriksson, Dr. Jochen Seybold – Altair Engineering GmbH
The meshing, modeling and evaluation of LS-DYNA crash models for vehicle crash simulations is constantly subjected to an increase in complexity, since the model sizes grow, several vehicle variants are instanced from one single CAE data set and since the number of load cases and the requirements on the evaluation quality are constantly increasing. Current models sometimes contain more than 1000 different components and 1 million elements. In recent years, the model complexity has also been increased since interior parts and dummy models are included. The increased complexity makes it difficult to handle the models since changes in parts in many cases influence many different vehicle variants and load cases. V-CESS is a “Virtual Process Automation Framework” designed to simplify and standardize this type of CAE processes and is developed together with DaimlerChrysler AG, department PBE/DAM . The system was developed in a pilot project with the aim to manage the CAE data of the “Sprinter Successor” in order to create, manage and evaluate crash models in LS-DYNA. In engineering projects together with DaimlerChrysler AG, department PBE/DAM, we have shown that the return on investment of V-CESS is reached at about 1.5 virtual prototypes. The meshing cost can be reduced by 15% and the engineering cost saving is approx. 30%.
Sattler M., Finkel A.
Erich Schelkle, Herbert Klamser – Dr. Ing. h.c. F. Porsche AG
Simulation technologies are methods which have been traditionally applied in automotive engineering for a long time. Over the decades, enormous progress has been achieved in both, the simulation methods and the CAE-programs used. Thanks to the high efficiency levels of the current computer generation and the use of economically priced commercially available hardware such methods are being applied on a widespread basis today. Visionary concepts of the past are turning into reality. When considering the process of automotive engineering and the futuristic potential inherent in those methods, it becomes obvious that virtual automotive development still is in its infancy. This applies to the whole range of options from the coupled parallel/sequential simulation of manufacturing processes to the cross-functional simulation, including efficient management systems designed to handle the entire CAE process . At the same time, the CAE model transfer between the various expert departments on the one hand and between the OEMs and the system suppliers on the other has to be optimized while developing suitable documentation- and DataMining systems . What must also be mentioned is the need for continuous updating of the traditional methods in terms of numerical data and technical content. Currently, in the early phases of automotive engineering, the development activities are mostly handled in a sequential manner [3,4]. That is where the newly conceived CAD/CAE methods come in quite handy: They allow component geometries to be prepared on the basis of topologies and parameters and subsequent modifications to be implemented quite rapidly . For the synergy effects of these innovative design tools to be made full use of it is necessary, however, to combine the parametric concept geometry model with mathematical optimization methods. This approach allows the inherent design potential to be fully opened up and thus the defined targets to be reached in the most optimum way . Even though such numerical design strategies have already been used in certain areas, their wide-spread and consistent introduction into conceptual design is yet to come. It is with these innovative CAD/CAE strategies that the present paper is dealing.
Dr. Lars E. Jonsson, Dr. William R. Magro – Intel Americas, Inc.
The Intel® Itanium® 2 microarchitecture is based on a 64-bit processor architecture that is ideally suited to compute-intensive applications such as LS-DYNA. In fact, Hewlett Packard* has demonstrated outstanding performance of their Itanium 2-based systems on a range of technical computing applications, including LS-DYNA. Given the success of the platform in achieving performance on a per-processor basis, we turned our attention to the speedups achievable with tightly coupled clusters of Itanium architecture-based servers. In this paper, we study the scalability of an Itanium 2-based server cluster consisting of 4-CPU SMP nodes, connected by gigabit Ethernet* and by a high-performance InfiniBand* Architecture interconnect. We study the relative performance of these interconnects, relating the observed application-level performance to the underlying performance characteristics of the interconnect.
Ala Tabiei, Ivelin Ivanov – University of Cincinnat
Curd-Sigmund Böttcher, Steffen Frik – Adam Opel AG
The confidence level of crash simulations is mainly determined by a well-defined finite element representation of the vehicle structure, correct modelling of the kinematics, and the material properties being applied. In the past, materials were described by quasistatic – or, if available – dynamic stress-strain characteristics. Besides this, each sheet metal part was assumed to have a uniform gage and material characteristics. However, it is a well-known effect that the physical properties of steel can alter significantly during the manufacturing process. This comprises an increase of material stiffness due to plastic deformation as well as gage changes. The amount of these changes is of very local nature and cannot be covered by simply scaling material properties and gages. In the past, crash software tools didn’t support the introduction of these local effects, so that they couldn’t be taken into account. In the meanwhile LS-DYNA has the capability to import information provided by stamping tools such as PAMSTAMP or AutoForm. Thereby a very important part of the material properties can be introduced into the crash simulation models, leading to a significantly increased correlation to test results. The impact of this effect on crash performance was analysed for a recent vehicle project and will be discussed in detail.
Stan Posey, Nick Meng – SGI
Manufacturing industry and research organizations continue to increase their investments in structural analysis and impact simulations such that the growing number of LS-DYNA users continue to demand more from computer system environments. These demands typically include rapid single job turnaround and multi-job throughput capability in a multi-user high-performance computing (HPC) environment. What is more, many LS-DYNA simulation environments coexist with other mechanical computer aided engineering (MCAE) software for structural analysis and computational fluid dynamics that all compete for system resources such as CPU cycles, system bandwidth, memory, disk storage and I/O. This paper examines the computational efficiency of structural analyses and simulations for relevant applications in LS-DYNA. Parameters such as model size, element types, and simulation conditions can produce a wide range of computational behavior such that consideration should be given to how system resources are allocated and configured. The computational characteristics of the SGI® Origin® 3000 servers, based on IRIX® and MIPS®, and the SGI® Altix™ 3000 servers, based on Linux® and Itanium® 2 from Intel®, are examined for both turnaround and throughput requirements that include industrial-size examples. In addition, simple guidelines on proper system configuration and innovative use of available SGI system resource management tools are provided that are designed to maximize LS-DYNA productivity. Introduction Mechanical design and manufacturing organizations increasingly rely on highperformance computing (HPC) technology and mechanical computer-aided engineering (MCAE) applications to drive innovation in product development.
Susanne Dörr, Hartmut Chladek, Armin Huß – Ingenieurbüro Huß & Feickert
Nielen Stander, Willem Roux – Livermore Software Technology Corporation, Mathias Giger, Marcus Redhe – University of Linköping, Sweden, Nely Fedorova – Snezhinsk Institute of Physics and Technology, Russia, Johan Haarhoff – University of Pretoria, South Africa
This crashworthiness optimization study compares the use of three metamodeling techniques while using a sequential random search method as a control procedure. The three methods applied are (i) the original Successive Linear Response Surface Method, (ii) the Neural Network method and (iii) the Kriging method. It is shown that, although NN and Kriging seem to require a larger number of initial points, the three metamodeling methods have comparable efficiency. The random search method is surprisingly efficient in some instances, but by nature much less predictable.
Dr. Hallquist J. O. – Livermore Software Technology Corp.
LSTC’s Perspective on the future Version 970 status Recent developments for crash Arbitrary Lagrangian-Eulerian Developments Implicit Developments EFG (Mesh-free) Developments MPP Outlook
Dr. Hallquist J. O. – Livermore Software Technology Corp.
Thomas Frank – Daimler Chrysler AG, Artur Kurz – LASSO Ingenieurgesellschaft mbH, Martin Pitzer – PENG GmbH, Michael Söllner – Porsche AG
Cing-Dao (Steve) Kan, Dhafer Marzougui, Nabih E. Bedewi – The George Washington University
Due to significant improvements in computer technology and finite element (FE) code capabilities, it has become more feasible and effective to incorporate occupant models in the analysis and evaluation of vehicle crashworthiness and safety. Using Detailed FE models that incorporate the vehicle, restrains systems, and occupants, in automotive safety analyses have shown advantages over the traditional methods where the vehicle and occupants are uncoupled. This paper describes a finite element model of a 50 percentile Hybrid III dummy. The model was developed with several fundamentals in mind. First the resulting FE model must accurately represent the actual dummy. This is accomplished by incorporating correct geometry, material properties, and connectivity for all components. Second, the dummy model should be easily positioned and incorporated in vehicle and vehicle compartment models. This is achieved by using a systematic scheme for numbering and organizing the parts and joints in the model. The model should also be efficient, robust, and reliable. For this, special modeling techniques, such as using a uniform mesh for all important components and using one “automatic single surface interface” to treat the contact between all dummy components are incorporated in the model.
Melissa Adoum, Vincent Lapoujade – CRIL Technology Correspondence
The aim of this study is to understand how the *USER_LOADING option works and to give some examples of its use. This option uses the subroutine loadud, its usage will be described step by step and it is illustrated with four examples : • Plate loaded with time dependant pressure, • Plate loaded with displacement dependant pressure, • Randers-Pehrson & Bannister Tests, • Plate loaded by an explosion of 5kg of TNT. Some of those examples use loadings that can be also applied through LS-DYNA existing options, thus the results obtained with the user loading option could be compared and validated.
Ulrich Franz, Peter Schuster, Werner Schmid, Oliver Graf – DYNAmore GmbH
Detailed finite element side impact dummy models of the USSID, EUROSID, and ES-2 have been developed in cooperation with the German Association for Automotive Research (FAT) during the last 5 years. All models are validated using tests at material and component levels as well as fully assembled models. The models are used by nearly all car manufacturers worldwide who use LS-DYNA for occupant safety simulations. The paper describes modeling aspects of the dummies and gives an overview of their performance in sled tests. Furthermore emphasis is put on difficulties and potential pitfalls that might arise during the everyday work with the models in predicting occupant injury risks. In addition to the knowledge gained during the development process, experiences from the support for and the consulting with the FAT dummy models are presented.
Anantha Ram B.S., Joachim Danckert, Torben Faurholdt – Aalborg University, Denmark
Wear prediction necessitates the investigation of elastic stresses developed in the workpiece material due to impact and sliding of abrasive particles in tribological contact situations. LS-Dyna implicit ﬁnite element analysis is used to investigate these contact stresses in the workpiece material under imposed Hertzian pressure loading. A line contact condition (cylindrical body on a plane surface) is assumed and the predicted stress ﬁeld is compared with analytical solutions. The model is parametric, two-dimensional and built in LS-Ingrid in terms of semi contact width ‘ ’. Application of hertz pressure and implicit control cards are some of the issues of this study. The study is a preliminary step in the extension of the model where the workpiece will be modeled with a thin hard layer/coating under the same contact situations. It is found that the ﬁnite element model developed, predicts the elastic stresses in close agreement with the theoretical results and the model can be suitably extended for analyzing contact situations of layered systems.
Haryanti Samekto – Universitaet Stuttgart, Karl Roll – DaimlerChrysler AG
Superplastic forming process has been a standard manufacturing process in aircraft industry and its applications in other industries are increasing. Superplasticity is utilised in forming parts which can not be produced technically or economically using materials with ordinary ductility. As superplastic deformation should be carried out under certain strain rate in which m value is maximal, the finite element method is applied to model the forming process in order to optimise the process through generating a pressure-time curve. In this paper, the dynamic explicit solution procedure was taken as an alternative solution due to its efficiency, as most of the current simulations of SPF used static implicit. The material parameters of Aluminium alloy 5083 SPF were first determined and the creep constitutive model was chosen. The finite element analysis results from dynamic explicit were verified with experimental results and then compared with static implicit. The simulation of bulge forming process with the geometry of a cup was conducted. The effects of m value and friction coefficient value were investigated.
Trevor Dutton – Arup Solihull, UK
M’hamed Souli – Universite des Sciences et Technologie de Lille, Lars Olovson – LSTC
Numerical problems due to element distortions limit the applicability of a Lagrangian description of motion when modeling large deformation processes. An alternative technique is the multi-material Eulerian formulation. It is a method where the material flows through a mesh that is completely fixed in space and where each element is allowed to contain a mixture of different materials. The method completely avoids element distortions and it can, through a Eulerian-Lagrangian coupling algorithm, be combined with a Lagrangian description of motion for parts of the model, see  and  The Eulerian formulation is not free from numerical problems. There are dissipation and dispersion problems associated with the flux of mass between elements. In addition, many elements might be needed for the Eulerian mesh to enclose the whole space where the material will be located during the simulated event. This is where the multi-material Arbitrary Lagrangian-Eulerian (ALE) formulation has its advantages. By translating, rotating and deforming the multi-material mesh in a controlled way, the mass flux between elements can be minimized and the mesh size can be kept smaller than in an Eulerian model.
Fuchun Zhu, York Huang – First Technology Safety Systems, Inc.
SID-IIs, the small side impact dummy, is the smallest side impact crash test dummy currently available in the market, representing the anthropometry of a small female or a thirteen-year-old child. It is designed specifically to evaluate the performance of advanced occupant protection systems, such as side air bags, in automotive side impact situations. This paper presents the development and validation of new FTSS SID-IIs dummy model for LS-Dyna solver. FTSS started the SID-IIs model development in early 2001, and accelerated the developing process in September due to the increasing demand for this crash test dummy in both physical and FEA versions. Now a fully validated SID-IIs FEA model is available in version 1.0. The SID-IIs model faithfully represents the physical dummy hardware from head to toe. It includes all the hardware components, accelerometers, load cells and linear potentiometers to measure rib deflections. Pro-Engineer CAD model was used in the model geometry creation. Nine common materials were used in the model, and their performance has been proved stable in the other FTSS dummy models. Totally the SID-IIs model consists of about 47,000 nodes and over 65,000 elements. The minimum time step has been controlled as 1.03 µs. Validation of the model has been done on various levels, which includes material and component tests, such as standard SID-IIs head drop test, lateral neck flexion test, Rib drop test, arm drop test, shoulder plug drop test and pelvis plug drop test, and full body calibration tests on different areas, such as shoulder, thorax, abdomen and pelvis. The model was finally validated against sled test data. The criteria for model validation accuracy were controlled at less than 10% off for component level, and 15% for full body dummy level. The model showed stable and robust performance with reasonable accuracy in prediction. Details about model development and validation will be presented in this paper. It is believed that the SID-IIs FEA model can serve as a useful tool in the early stage of vehicle design and safety simulation.
Arthur B. Shapiro – Livermore Software Technology Corporation
LS-Dyna can solve steady state and transient heat transfer problems on 2dimensional parts, cylindrical symmetric parts (axisymmetric), and 3-dimensional parts. Heat transfer can be coupled with other features in LS-DYNA to provide modeling capabilities for thermal-stress and thermal-fluid coupling. This paper presents several examples using LS-DYNA for modeling manufacturing processes (e.g., upseting, extrusion, welding, casting).
R. Malcusson – Saab Automobile AB
Guangye Li, Jeff Zais – IBM Deep Computing Team
Khaled Sennah – Ryerson University, Magdy Samaan – University of Windsor, Ahmed Elmarakbi – University of Toronto
Cronin D. S., Kaufmann C. – University of Waterloo, McIntosch G., Bui K., Berstad T. – LSTC
Ceramic materials are commonly used in protective armour applications and may be subject to high-energy ballistic impacts in these situations. Under simple loading conditions, ceramics may be regarded as elastic-brittle materials. However, when considering ballistic impacts, the post-yield response of the ceramic becomes significant. One of the most widely used constitutive models for simulating the postyield response of ceramic materials is the JH-2 ceramic model. This constitutive equation was developed by Johnson and Holmquist and incorporates the effect of damage on residual material strength and the resulting bulking during the compressive failure of a ceramic material. The relevant equations describing the response of the material are described. In particular, the model parameters currently available for common ballistic ceramic materials are presented. The JH-2 constitutive model has been implemented in LS-Dyna as material 110 (*MAT_JOHNSON_HOLMQUIST_CERAMICS). Validation against the available test cases in the literature is discussed, and a sample calculation of a sphere impacting a ceramic material is presented. The JH-2 model in LS-Dyna has also been used by Kaufman et al. to successfully simulate the ballistic impact of 12.7 mm armourpiercing projectiles on supported alumina tiles.
Daniel Hilding – Engineering Research Nordic AB, Erik Schedin – AvestaPolarit AB
This paper describes the implementation of a new material model into LS-DYNA version 960, LSTC (1-4), a material model capable of predicting the TRIP-effect of HyTensX for different forming operations in different temperature scenarios. The implementation is verified by comparing measurements from three tension tests with simulations of the tension tests. The comparison shows good to excellent agreement, which is a strong indication that the implementation is correct and that the material model can be used to predict the hardening behavior of the material with good accuracy.
Dr. B. Maker, Dr. R. Grimes, Dr. C. Ashcraft – Livermore Software Technology Corp.
Bernhard Schott, Christof Westhues – Platform Computing GmbH
Gunther Blankenhorn – Universität Karlsruhe, Karl Schweizerhof – Dynamore GmbH, Hermann Finckh – Institut für Textil- und Verfahrenstechnik Denkendorf
Past perceptions to the processes of the penetration mechanisms of projectiles acting on textile structures  are often based on continuum models  or simplified models  and admit only limited conclusions concerning the real behavior of protective clothing made from several layers of fabric. Only a few investigations are known up to now with models based on single yarns as a major component for discretization . Thus, for the prediction of the protective effect of several layers of high-strength fibers in a textile, a structural approach is chosen by a separate modelling of each fiber by a shell or continuum based element. The single modelled fibers interact over a contact formulation with the adjacent fibers in the same way the fibers in the different layers do. This allows to model the in-plane motion and deformation of each fiber separately, as well as the failure of fibers thereby avoiding artificial localization effects to a great extent. With the so-called explicit finite element code LS-DYNA  different possibilities of the discretization of the fiber bundles are investigated. Also the description of the fiber material by available material models is varied modifying the load deformation relation and the damage evolution. The goals of the current project  are first to achieve a geometrically consistent model of the layered structure and second to better understand the phenomenological process of the impact of ballistic projectiles on such textiles. Finally, the particular effect of different layer setups can be studied. INTRODUCTION Up-to-date bullet proof vests are consisting of several layers of fabrics made of high performance fibers like Kevlar® (Du Pont), Twaron® (Tijin) and Zylon® (Toyobo). These vests have a specific weight up to 2000 g/m². Many so called “trial-and-error” tests have to be performed to improve the weight of this vests. So far, only few attempts are known to predict the behavior of new constructions, like new fiber materials or different fiber materials for different layers, by way of exploring the mechanical phenomena. Numerical simulations by a finite element program could be a useful tool to detect this phenomena and they would allow the developers to optimize their products. To perform a numerical simulation, some research into fiber and fabric geometry and their possibilities to approximate them through finite elements is necessary. Also, the material behavior of the fibers exposed by high velocity loading must be measured to choose a suitable material model. The objective of this paper is the description of the fabric geometry, the discretization by shell and solid elements and first investigations of the model behavior in the analysis. Weave geometry and fiber properties Fabric and fiber geometry Following the approach to discretize the weave through a collection of single yarns, the shape of the cross section and the curve through the center of gravity of the cross sections must be acquired. To get this information of an unloaded weave, two specimen of a Kevlar® weave were embedded in epoxy resin. One of them was roG – I – 08
Gernot Oberhofer, Harry Dell, Dmitri Dell, Helmut Gese – MATFEM Partnerschaft Dr. Gese & Oberhofer Horst Lanzerath, Jürgen Wesemann – Ford Forschungszentrum, Elke Hombergsmeier – EADS Corporate Research Center
The Crash Simulation of Magnesium Structures with Finite Element Methods demands the use of suitable material and failure models. An associated plasticity model describing the complex asymmetric yield behaviour in tension and compression of Mg extrusions has been developed during the InMaK-project (Innovative Magnesium Compound Structures for Automobile Frames) supported by the German Federal Ministry for Education and Research (BMBF). Differences to the material model 124 in LS-DYNA are exposed. In order to describe the failure behaviour of Mg extrusions under multiaxial loading in FEM crash simulation this constitutive model has been combined with a fracture model for ductile and shear fracture. The fracture model has been added to the user defined constitutive magnesium model in LS-DYNA. The experimental investigations carried out on model components are compared with numerical derived results. Experimental methods for fracture parameter evaluation are shown and general aspects of metal failure due to fracture as well as different modelling techniques are discussed.
Dr. A. Rößler – ICIDO
Matej Vesenjak, Zoran Ren – University of Maribor
The road restraint systems on public streets are used to prevent a vehicle to veer off the road or its breakthrough to the opposite side of the road. The road restraint systems designed according to the EN 1317 standard are intended to provide certifiable levels of vehicle containment, to redirect errant vehicles and to provide guidance for pedestrians and other road users. Its proper design is therefore crucially important for safety of all road users. Practical observations of installed systems indicate that the current design of road restraint system is far too stiff. This results in unacceptable decelerations during the vehicle impact. The global stiffness of the road restraint system is largely attributed to the design of the distance spacer in the initial phase of an impact. The purpose of this research is to evaluate several new designs of a distance spacer with increased strain energy absorption due to controlled deformation during the vehicle impact. The impact severity and stiffness of various designs have been evaluated with dynamic nonlinear elasto-plastic analysis of a three-dimensional road restraint system within the framework of the finite element method with LS-DYNA. The computational analyses prove that the currently used distance spacer is indeed too stiff and that new designs assure controllable elasto-plastic deformation and crash energy absorption which in turn decreases the decelerations of an impact vehicle and consequently increases the safety of vehicle passengers.
Pravin Kulkarni, Shashikiran Prabhakar – Cessna Aircraft Company
Forming of aluminum sheets in T-temper is a much sought after industrial process, especially in the aircraft industry. However, the success of this process largely hinges on the ability to predict springback accurately. Aluminum sheets in T-temper exhibit approximately twenty percent variability in material properties and also the amount of springback is very large. This makes tool design for aluminum in T-temper an iterative and difficult to control process. Traditionally aluminum has been formed in the O-temper and then heat-treated to T–temper, as recourse to reduce springback. This research is aimed at developing a predictive finite element technique for springback, using experimental validation. A parametric study was conducted to determine the influence of geometric parameters and tempers on springback. The study characterizes springback of aluminum in different tempers and investigates the effect of forming strain-rates on springback. The study focuses on springback in Aluminum 2024 using hydroforming process.
J. Brockmann – Faurecia Autositze GmbH & Co. KG
Adrian Hillcoat – Hewlett-Packard Ltd Bracknell, UK
In the area of crash simulation, LS-DYNA has traditionally been a good application for measuring performance of leading computer architectures. This paper considers the architectural design of the Intel Itanium® processor, as well as the system architectures into which Itanium fits, with the purpose of understanding the benefits of the novel 64-bit EPIC architecture over and above conventional 64-bit RISC architectures as well as 32-bit Intel XEON processor architectures. The paper will provide details of LS-DYNA performance results achieved using distributed memory parallel execution. It will conclude with a look into the future to predict what might be achievable with further generations of this processor architecture for crash codes.
Kelly Carney, Mike Pereira, Duane Revilock – NASA Glenn Research Center, Paul Matheny – Florida Turbine Technology
In the rare event of a fan blade separation, the fan blade must not penetrate the case of a commercial jet engine. Due to this requirement the fan case is the heaviest single component of a jet engine. With a goal of reducing that weight, a simulation of a fan blade containment system was tested at the NASA GRC Ballistic Impact Lab and analyzed using LS-DYNA. A fan blade simulating projectile was shot at two alternate geometric containment case configurations. The first configuration was a flat plate which represents a standard case configuration. The second configuration had a surface curved outward from the projectile. The curved surfaced forces the blade to deform plastically, dissipating energy before the full impact of the blade is received by the plate. The curved case was thus able to tolerate a higher velocity of impact before failure. The LS-DYNA analytical model was correlated to the tests and a weight savings assessment was performed.
Harry Schlagenhauf – science+computing
Rainer Emrich, Udo Jankowski – Tecosim GmbH
Hervé Le Sourne, Nicolas Couty, François Besnier – Principa Marine, France, Cyrille Kammerer, Hervé Legavre – DCN Ingénierie, France
Ship and submarine structures have long been studied thanks to finite element methods. Their large dimension and complexity, the coupling with heavy fluid and the presence of a free surface raise numerical problems on the field of dynamic analysis. This is particularly true for extreme or accidental situations such as collision and grounding, underwater explosion and severe fluid impact. This paper describes first numerical methods and tools developed in LS-DYNA and used in the simulation of ship-submarine collisions, with a focus on the outer collision dynamics, i.e. global motion of the impacted structure. Using a new version of the rigid body dynamic program MCOL included in LS-DYNA, the influence of hydrodynamic effects is highlighted in ship-submarine collisions and in military surface ship collisions. The second part of this paper is dedicated to dynamic response analysis of surface ship or submarines submitted to underwater explosions (UNDEX). Coupled with the Underwater Shock Analysis (USA) code, LS-DYNA is a powerful tool used to simulate the response of ships or submarines to both the shock wave induced by the detonation and the bubble gas effects. Through some examples of 3D numerical models of military vessels, the paper presents the capabilities of the LS-DYNA/USA tool and some difficulties encountered in such an analysis. Impulse loads with high pressure peaks may occur when a ship bottom hits the water with a high velocity. This is often called “slamming”. Sometimes ships suffer from local damage from the impact load or large buckling on the deck. In the last part of this paper, LS-DYNA is used to simulate such water-entries. Interaction between lagrangian bodies and multi-materials eulerian fluids (air and water) is taken into account thanks to a penalty coupling method.
Len Schwer – Schwer Engineering and Consulting Services, USA
LS-DYNA provides several beam element formulations, see the keyword description for *Section_Beam in the User’s Manual. Several of these beam element formulations support user supplied integration of the cross section, via the *Integration_Beam keyword. While most LS-DYNA users are familiar with the similar through-the-thickness integration algorithm for shell elements, which is made trivial by the rectangular cross section geometry assumed for shell elements, the numerical integration of even simple beam element cross sections requires more effort, and as will be demonstrated, more planning. In this article, a detailed explanation of the beam element cross section integration algorithm is presented. Simple suggestions for calculating, and checking, user provided integration rules are illustrated through several examples. The examples also provide suggestions for improving the LS-DYNA Standard Cross Section Types, available via the ICST parameter of the *Integration_Beam keyword.
John U.S. Hanlon – EDAG Engineering & Design AG, Bernward Platz – Teraport GmbH
At EDAG, the increasing demand for crash simulations with LS-DYNA required the expansion of the existing compute resources in the beginning of 2002. As at this point there was a stable and performant MPP-version of LS-DYNA available, EDAG decided – despite of the new technology – to purchase a Linux-cluster consisting of 8 nodes and 16 processors. Meanwhile EDAG has deployed over 14 clusters at three different locations. Whereas at the beginning, purchase costs and performance played a major role, due to the increasing complexity further demands became important which were realised by the EDAG partner, Teraport: fast, convenient and reproducible installation and administration of the cluster systems parallel usage of several DYNA versions with various MPI-libraries easy extensibility convenient usage of cluster resources via web technologies high-available access to the cluster resources handling of high data transfer integration of further FEM-applications The paper describes a use case which reports the concepts of the realisation of these requirements as well as the problems such as hardware quality during the installation and setup of the cluster environment. At the end, a perspective concerning further possibilities of development such as web-based workflow optimization and result management will be shown.
Chris Galbraith, Dylan Thomas – Centre for Automotive Materials and Manufacturing currently with Honda R&D Americas, Inc., Mark Finn – Alcan International Ltd.
This paper presents the results of applying the finite element method to calculating the spring back of an automotive hood assembly, and its application to the functional build method. The assembly was comprised of six individual panels: an inner panel, an outer panel, a major reinforcement, a latch reinforcement, and two hinge reinforcements. Finite element simulations were conducted for forming each of the six components. Each component was formed, trimmed, and positioned in car position. The outer panel required several secondary forming operations including a re-meshing, remapping, trim, and flanging operation. Once in car position, the components were moved so that they just contacted each other, and were “spot welded” together through the application of nodal constraints. Mastic between components was simulated with tied contact. Contact between components was simulated with contact interfaces. Finally, a spring back analysis was conducted. The models clearly illustrate that it is possible to predict spring back of large automotive assemblies, and that the assembly process yields different final shapes than those obtained from spring back of individual components. With this newly developed tool it is possible to predict whether or not the assembly process can correct out-of-spec components, a key factor in utilizing the functional build method.
Enver Serifi – Universität Stuttgart, Andreas Hirth, Stefan Matthaei – DaimlerChrysler AG, Heiner Müllerschön – DYNAmore GmbH
Foam materials are widely used in automotive industry such as energy absorbers and comfort enhancers. Because of high energy absorbing capability of foams, they became very important in vehicle crashworthiness. So in this manner, FE modelling of foam materials also becomes important. Although foam materials are very promising materials, not that much study has been done about foams comparing to other commonly used materials like steel, etc. Some different approaches are available to define the behaviour of foam materials. One micro-structural approach to define the mechanical behviour of foam materials, considers the foam material as a cubic model and uses the standard beam theories with solid-fluid interaction to describe the in- and out-flow of fluid inside the foam material(see Gibson and Ashby ). There are also macro-structural approaches those consider the foam material as a continuum with solid-fluid interaction in order to describe the in- and out-flow and pore-fluid in the foam material (e.g. Ehlers ). In contrast to such quite sophisticated models, in LS-DYNA for practical engineering pruposes the foam model *MAT_FU_CHANG_FOAM (MAT83) is available. The main assumption of MAT83 is, that Poisson’s ratio is equal to zero for foams and therefore no coupling between the material axes is present. This leads to a one-dimensional material law, where experimental curves of uni-axial test can be used directly.
Enver Serifi – Universität Stuttgart, Andreas Hirth, Stefan Matthaei – DaimlerChrysler AG, Heiner Müllerschön – DYNAmore GmbH
Foam materials are widely used in automotive industry such as energy absorbers and comfort enhancers. Because of high energy absorbing capability of foams, they became very important in vehicle crashworthiness. So in this manner, FE modelling of foam materials also becomes important. Although foam materials are very promising materials, not that much study has been done about foams comparing to other commonly used materials like steel, etc. Some different approaches are available to define the behaviour of foam materials. One micro-structural approach to define the mechanical behaviour of foam materials, considers the foam material as a cubic model and uses the standard beam theories with solid-fluid interaction to describe the in- and out-flow of fluid inside the foam material (see Gibson and Ashby ). There are also macro-structural approaches those consider the foam material as a continuum with solid-fluid interaction in order to describe the in- and out-flow of the pore-fluid in the foam material (e.g. Ehlers ). In contrast to such quite sophisticated models, in LS-DYNA for practical engineering purposes the foam model *MAT_FU_CHANG_FOAM (MAT83) is available. The main assumption of MAT83 is, that Poisson’s ratio is equal to zero for foams and therefore no coupling between the material axes is present. This leads to a one-dimensional material law, where experimental curves of uni-axial test can be used directly. The aim of this work is to provide a method and to develop a computer program in order to generate reliable input data for the simulation of EPP foam with MAT83 in LS-DYNA. Experimental raw data have to be prepared and extended respectively. In addition, suitable density laws have to be developed in order to provide LS-DYNA input data for intermediate densities, where no experimental data are available. To verify the reliability of the results, simulations with the generated curves are compared to an independent experimental database and to some real experimental applications.
W. Rehm, B.Wang – Forschungszentrum Jülich, B. Binninger – University Aachen, C. Nae – University Bucharest
In this this paper, we describe main results obtained within the scope of joint project activities concerning the numerical simulation of reacting flows in complex geometries. The aim is the refinement of numerical methods for applied computational fluid dynamics (CFD) using high-performance computations (HPC) to study explosion processes in more detail, especially for hydrogen safety and environment aspects of innovative technical systems. The R&D work is mainly focused on the modelling of the accident-related behaviour of hydrogen in safety enclosures ranging from slow to fast or even rapid flames with resluting explosion loads. Therefore, we have established a modern field code cluster with supercomputing and special modules for specific studies, including fluid-structure responce This recent methodology allows the assessment of adequate safety measures to control deflagration-to-detonation transition (DDT) processes and to suppress fires or explosions with relevance to industrial safety, to reduce combustion loads and structural deformations. For instance, visualization/animation of flow and crash tests are presented.
Peter Gantner, Herbert Bauer – Aalen University, David K. Harrison, Anjali K. M. De Silva – Glasgow Caledonian University
This paper is concerned with closing a gap in the process chain of metal forming. Tools for simulating the metal forming process like LS-DYNA® produce output geometries and stress information which cannot be easily re-imported in CADSystems or structure analysis programs for further processing. A concept has been developed and implemented with a corresponding program which allows the re-importation of parts with certain topologies (tube, plane) from LSDYNA® into any STEP-conformant CAD program. This method is mainly based on using the interconnection information which is contained in the LS-DYNA® output file. This information allows the construction of interpolating cubic B-Spline surfaces which can be represented in the STEP standard format. Thus, it is not necessary to reinvent general purpose surface reconstruction programs but rather to harvest the additional information available in the given situation. Furthermore, a method to make the strength hardening information available in structure analysis is represented. This hardening results from the forming process and should be considered to obtain a more realistic virtual prototype and is of assistance to save weight and material costs. Introduction The forming simulation by means of finite element analysis (FEA) is becoming more and more important in the field of process quality assurance and process design of mechanical and fluid media formed components. Using the finite element simulation in the development process of hydroforming components from the first draft through to the serial production of a component provides an enormous saving of development time and costs. However, due to the constantly increasing competition in terms of costs, development time and quality, a further reduction of processing time and costs is necessary. Moreover, there is the demand for even more exact predictions and results in the area of the virtual component, in order to reduce the weight and to ensure that the component produced will have enough stability and low material costs. Due to this demand the integration of the forming simulation into the process chain must be improved. In order to point out the optimisation potential, the sequence of the processes from the design phase to the finished component is represented in Figure 1.
Wenliang Chen, Dingyu Chen, H.Xie, Quanqing Yan, Arthur Tang, Chin Chun Chen – Engineering Technology Associates, Inc., USA
Over the past two decades, the finite element analysis (FEA) has emerged as one of the most important engineering tools in the many industries, due to its flexibility and accuracy in prediction. Nowadays, it is widely used in sheet metal forming industry to predict the forming feasibility of wide variety of complex components, ranging from aerospace and automotive components to household products. As the demand of FEA accelerates, the need for a robust and streamlined process based pre-processor has become crucial, especially in the tool and die manufacturing industry. As a continuous effort, the DYNAFORM Version 5.0 has been developed to meet the needs. This paper will discuss the capabilities of two newly developed key features embedded in the DYNAFORM Version 5.0. These “state of the art” features are Quick Setup (QS) and Die Face Engineering (DFE) module. The QS module aims to reduce the setup time for stamping simulation, while the DFE module allows stamping engineers to create, re-design and reengineer the tooling from part geometry.
L. Kangwook – Hyundai Mobiles
Melissa Adoum – CRIL Technology Correspondence
Safety nets are used to protect skiers during downhill competitions. However, although these nets are now able to retain skiers in almost all cases, the deceleration during such impacts can cause severe harm to skiers including hyperflexion injury or vertebra compaction. Experience showed that the behaviour of the nets is highly dependant on: • the material of which they are made of • the boundary conditions (installation and fastening). The aim of this study was : • to analyse the net constitutive material under dynamic loadings to determine its characteristics. • to analyse well defined impacts to be able to simulate them numerically • to improve the net behaviour during real impacts • to provide some recommendations concerning the geometry of the complete system. This study continues Fayçal Ben Yahia’s work which was presented in Paris in June 2001. The whole study was performed under the funding of the French company Dalloz Montagnes and with the technical collaboration of the International Skiing Federation (FIS). As a first step, we used the tests on the net thread to model it in static and dynamic conditions. Then we compared the results from impact tests on small safety systems and simulations with LS-DYNA which led to the validation of the whole model. Finally, simulations of impacts on real size nets were used to study the influence of the boundary configuration on the net behaviour. The model developed during this study makes it possible to optimise the geometry of the whole system in order to increase skier protection.
Alexander A. Ryabov, Vladimir I. Romanov, Sergey S. Kukanov, Dmitry V. Roschihmarov Sarov – Open Computing Center, Russia
Development of a durable flight recorder’s protective case is a very difficult engineering problem. To save information, the case must be strong enough to withstand intensive accidental thermal and mechanical loading in aircraft crash. So the case must meets strong requirements. One of the toughest of these demands is that any flight recorder’s case must be strong enough to withstand a penetration resistance test – an impact of pin at any angle at a speed of 7.7 m/s. This paper presents the results of a series of 3D numerical simulations of penetration resistance test of a flight recorder’s protective case. The simulations performed with LS-DYNA. The computer simulations showed the most severe direction of pin impact to the case.
Dmitri Fokin, Nitin Lokhande, Lars Fredriksson – Altair Engineering GmbH, Germany
In the present paper a basic finite element model of an ALE thorax side airbag with a simplified gas generator will be considered. In particular, it will be discussed how to define boundary conditions and properties of the inflating gas. A possible general approach to ALE airbag validation to fit results of a standard body block test will be described. Finally numerical results for a push-away test for an ALE and corresponding CV airbag will be compared.
William W. Feng, John O. Hallquist – Livermore Software Technology Corporation
The Hill-Ogden elastic constitutive equation for incompressible and compressible rubber-like materials is presented. The derivation and computer programs to determine the material constants for these equations from uniaxial and biaxial tests are included. These constitutive equations and the computer programs for determining the material constants have been implemented into LS-DYNA. A few examples are shown. Some special cases are given to demonstrate the versatility of these constitutive equations. The Mooney-Rivlin constitutive equation is a special case. The Feng-Christensen viscoelastic foam model in one-dimensional compression, developed in 1986, can be written in a mathematical form and implemented in finite element codes.
Michael Walter, Hartmut Chladek, Dr. Armin Huß – Ingenieurbüro Huß und Feickert, Germany
For a prototype of the new LANDROVER cockpit, developed by Siemens VDO Automotive, the design of the co-driver airbag area had to be optimized with the help of simulation in order to fulfill the guidelines of ECE-R 21. Furthermore a more limited inhouse-target had to be reached. The actual state of the prototype was illustrated by simulation and the influence of different measures on the crash behavior was examined. Among other things the following modifications were accomplished: • Changing the stiffness of the instrument panel by specific design of the rib structure • Reducing the stiffness of the airbag box and its connections • Partial absorption of the impact energy using foam depositors between instrument panel and airbag unit The influence of different strategic ways to optimize the crash behavior, in respect of head impact will be shown on a simplified cockpit model. Finally, a comparison of the simulation and the testing of the new RANGE ROVER cockpit will be given.
K. Altmeyer – Fujitsu Siemens Computers GmbH
Dr. A. Findling – NEC Europe GmbH
Dr. P.O. Marklund – Engineering Research Nordic AB, Dr. B. Pipkorn, C.F. Lindh – Autoliv Research AB
Michael Keigler, Herbert Bauer – University of Applied Sciences, Aalen, Germany, Richard Hall – University of Wolverhampton, UK, Prof. Musa Mihsein – De Montfort University, UK
Clemens-August Thole, Liquan Mei – Fraunhofer Institute for Algorithms and Scientific Computing
In crash simulation, small changes of the model or boundary conditions may result in substantial changes of the simulation results. For a BMW car model, the node positions of the crashed model show differences of up to 14 cm between several executions on a parallel machine for the same input deck. For the Dodge Neon testcase, small variations of the barrier position result in substantial scatter of the intrusion. Detailed investigations of several models have shown, that in some cases numerical effects might be responsible for the scatter in the results. In most cases, however, the instable behaviour of the simulation results is caused by bifurcations. These bifurcations result from numerical algorithms or are a feature of the car design. In the Neon model the scatter is a result of the interaction between the axle and the engine block. In the case of the BMW car model, the scattering of the simulation results is a direct consequence of buckling of the longitudinal rail. A slight redesign of this part causes stable results for parallel machines. Stable crash behaviour of a car model is a design target for the following reasons: • Simulation results might be misleading, when the impact of changes of the model or model parameters is investigated. • The numerical model is always only an idealized representation of the real car design. A stable crash behaviour simplifies the prediction of the crash behaviour of the real car from simulation results for the idealized model. • Smaller bounds for the scattering of the characteristic crash values will improve the possibilities of the engineer to find the best compromise for the car design with respect to the targets of the different load cases. Due to the nature of crash simulation many parts might show instable behaviour. Usually, only a small subset has a real impact on those values, which measure the crash behaviour (like intrusion). Measuring the scatter of simulation results for these characteristic values is a first step. In order to improve the design, it is necessary to trace this scatter back to its origin in space and time. DIFFCRASH is a tool, which allows one to measure scatter and to trace this scatter back to its origin. It allows the engineer, to understand the mechanisms of propagation and amplification of scatter during the crash itself as a basis for the improvement of the stability of the car design.
Masami Iwamoto, Kiyoshi Omori, Hideyuki Kimpara, Yuko Nakahira, Atsutaka Tamura, Isao Watanabe, Kazuo Miki – Toyota Central R&D Labs., Inc.,, Junji Hasegawa – Toyota Motor Corporation, Fuminori Oshita – The Japan Research Institute, Ltd.
A finite element model of total human model for safety, which is called THUMS, has been developed in order to study human body responses for impact loads. In previous report, a mid-size adult male occupant model of THUMS was developed and validated for several impact loads. This paper briefly describes recent advances in THUMS. Individual internal organ models and a detailed brain model have been developed to be integrated with the occupant model. In addition to the mid-size male occupant model, a small female occupant model and a mid-size male pedestrian model have been also developed and validated in order to simulate impact responses of female occupants and male pedestrians.
Youn-Seo Roh, Henry Fong – Sun Microsystems, Inc.
A recent effort of optimizing the performance of LS-DYNA running on SPARC(R)-SolarisTM servers is described. With new releases of compilers, generated executables benefit from the additional performance of latest UltraSPARC(R) CPU’s for SMP servers. Also, new release of Sun HPC ClusterToolsTM cluster environment includes tools that facilitates tuning of LS-DYNA-MPP executables and the MPI environment. A collection of development tools targeted for SPARC performance improvement results in faster simulations, fully benefiting continuously updated hardware performance. Those developments are exhibited with customer benchmark examples. With Sun ONE Grid Engine products, more efficient simulation environment is viable for LS-DYNA simulation.
P. Wriggers, K. Fischer – University of Hanover, A. Rieger – Continental AG
During the last years considerable eﬀort was devoted to better numerical treatment of contact problems. This fact is due to the growing computing power which lead to more and more sophistication and detailed technical models within engineering analysis. Due to the more precise modelling within the associated discretization process often unilateral constraints have to be considered. Hence better discretization techniques, especially for ﬁnite deformations, are needed to solve problems with contact constraints in an eﬃcient and robust way. In this paper we will discuss some recently developed discretization schemes and algorithms for the treatment of contact constraints. The presentation is split into two parts. The ﬁrst one is devoted to discretization techniques for contact problems which fulﬁll the BB-condition needed for a stable contact discretization scheme. This leads to a discussion of weak enforcement of the contact constraint conditions which results in so-called mortar methods for linear and nonlinear problems. Here also special remarks are made with regard to eﬃcient solution schemes which are based on a total gap vector at the contact interface. The second part of the presentation is related to adaptive ﬁnite element methods for large deformation thermo-mechanical contact problems. Here a special staggered scheme is developed in which diﬀerent ﬁnite element meshes are combined to solve the thermo-mechanical contact problem. Based of the methodology of the Zienkiewicz, Zhu error indicators based on superconvergent patch recovery special error indicators are developed for the the mechanical and thermal part of the problem including the contact constraints. Furthermore an error indication in time is derived for the thermal heat conductance equation based on a time-space discretization which uses a continuous Galerkin scheme for the time integration. Using such integration algorithm one can derive again a error indicator by assuming superconvergent time points. This method is applied to solve an example with known analytical solution which allows the computation of eﬃciency indices. Here it can be shown that the developed adaptive time stepping scheme results in very good eﬃciency of the method. For all parts, the basic theoretical basis is derived, algorithmic implications are discussed and explanatory examples are presented to show the properties of formulations when compared to existing ones.
Horst Raich – DaimlerChrysler AG
During the development of truck cabins the safety of the driver and the front seat passenger in an accident is considered. The cab must be designed in such a way that in an accident a sufficient survival space is guaranteed. The legal requirements of cabin safety are fixed in Europe in the regulation ECE-R29. In order to reduce the number of iteration loops during the development process, a computational simulation method for the load cases roof strength test, front impact test and rear wall strength test of the ECE-R29 was introduced. The explicit finite element program LS-DYNA was used for that purpose. The deformations of the driver‘s cab and the loads of the individual components within the elastic and plastic range of the material behaviour can be determined before the first tests are carried out. These tests can then be limited to a minimum by the numeric simulation. In this paper, the application of this numerical method by the example of the new of ACTROS Megaspace cab is presented and compared to the results from the acceptance test according to ECE-R29.
Ove Sommer, Thomas Ertl – science+computing, Norbert Frisch, Dirc Rose – University of Stuttgart
This paper presents a new preprocessor for the assembly of independently meshed car body parts. The assembly process gains more and more importance in the preprocessing of crash-worthiness simulations. It is desirable to take simulation results into consideration for construction decisions already in the early phase of the car development process. At this stage, CAD data unfortunately does not contain any information about constraints like spotwelds or adhesive bondings between substructures. This lack of data has to be resolved by the simulation engineer. We provide an adequate preprocessing tool for this purpose, called scFEMod. scFEMod supports the simulation engineer in effectively defining missing constraints such as point, surface, edge, or line links. All of these can be specified interactively with the mouse pointer. Hierarchical data structures guarantee quick and automatic flange detection so that the assembly process is significantly accelerated. Furthermore, scFEMod can be used to replace separately meshed car body parts by variants which then need to be adapted to the adjacent mesh structure. Subsequently, initial perforations and penetrations can be detected, visualized, and selectively removed. Sensor points can be positioned and oriented in order to compare simulation results with those of physical crash tests. scFEMod allows to distribute non-structural masses over all car body parts they are connected to. Finally, the proper assembly of the whole car body model can be validated.
Tobias Müller, Martin Lawerenz – University of Kassel
Smart materials and smart structural concepts in ﬂow control have the potential for signiﬁcant impact on the design and performance of modern turbocompressors. While the beneﬁts of an airfoil whose geometry is variable were investigated in detail in the area of adaptive wings for airplanes, this is a new ﬁeld for the application in turbomachines. The main focus is on simulations of novel ﬂow control concepts to allow a structural ’morphing’ and thus changing the aerodynamic characteristic of the airfoils. LS-DYNA 970 Implicit is used for the calculations. Additionally, shape optimiziations are performed using LS-OPT in conjunction with a parametric mesh generator.
Ashwini S. Gokhale – Wichita State University
Manufacturing process simulation using finite element techniques has immensely contributed to ensuring the success of concurrent design methodologies. However, Finite Element Methods (FEM) is computationally expensive and consequently unsuitable for design and manufacturing optimization in a production environment. In this research, a coupled Artificial Intelligence (AI) and FEM technique was developed to simulate and predict process response to changes in part design. Generic process models of part families are developed using Artificial Neural Networks (ANNs) and FEM. The generic models are used to predict the response of the manufacturing process to variations in geometric, material and process parameters, in real time. The predicted results are graphically displayed in a Virtual Reality environment. Standalone software VRForm was developed based on this methodology. VRForm can be used to optimize component, tool and process designs.
Jens Buchert, Herbert Bauer – University of Applied Sciences, Germany, David K. Harrison, Anjali De Silva – Glasgow Caledonian University, UK
Active hydro mechanical forming (AHMF) has been developed in order to meet the demand of the automotive industry for economical production of sheet metal parts with more individuality in small lot sizes. Conventional deep drawing of automobile parts which have large outer surface areas (such as roofs, doors and hoods) leaves them with a very small dent resistance. This is caused by the low deformation degree in the middle of the part. This low component stiffness has a negative effect on the crash resistance of vehicles. By using the AHMF technology a consistent plastic strain distribution can be brought into the part and therefore its stability will be improved. Within the design of conventional deep drawing tools the spring-back of a part is well understood because of the long experience with the process and its influence on the stress distribution inside the deformed part. In AHMF, as with any new process, the spring-back phenomenon is still under investigation. There are a considerable number of parameters to control in order to regulate the stress distribution and the shell thickness, which leads to more spring-back possibilities. This paper presents a simulation of spring-back in AHMF to find an efficient method of tool design and to generate the optimum process parameters, in the prototyping or later production. The FEM simulation is based firstly on the validation of the results of the AHFM process given by the mathematical calculation with the computer generated model. Secondly, spring-back simulation is introduced to see the influence of pressure curves and the blank holder forces or the parameters of the pre-bulging step. The quality of the FEM simulation is verified using practical applications in the automotive industry.
Tatsuya Fukushima,, Hitoshi Shimonishi – NISSAN MOTOR CO.,LTD, Kimihiro Hayashi – The Japan Research Institute, Limited, Masaki Shiraishi – SUMITOMO RUBBER INDUSTRIES, LTD
A simplified FE tire model has capability to solve a large deformation of a tire on a vehicle running simulation with acceptable computational time. We tried to simulate a vehicle running on to a curb, one of vehicle strength tests, by using simplified FE tire models and a vehicle FE model which has detailed suspension models. It is necessary to solve the force to a wheel with good agreement with a car test, in order to estimate the force to suspension systems or a body as a vehicle strength problem. We confirmed that the simulated force to a wheel agreed with the force by car test.
CAROLINE J.FIELD – Ove Arup & Partners
This paper focuses on the finite element simulation of two full-scale tests of high performance, seismic-resistant structural frames using the LS-DYNA 960 implicit solver. The frame was physically tested as part of the design validation for the new Stanley Hall building on the University of California Berkeley Campus. The pseudo static non-linear analyses, showed excellent correlation with the measured test data. Two sequential tests were performed on the same frame but with different brace configurations, hence residual stresses and strains, and the process of brace replacement were important. This work illustrates the convenience of implicit LS-DYNA for structural applications – transferring this technology to the built environment. It also provides confidence in and verification of the software. The construction industry tends to shun non-linear analyses, deeming them too complicated; however it is ideal and indeed, essential for seismic applications. This simulation provides an alternative approach to full-scale testing for the future evaluation of this type of structure. It also provides the opportunity for the development of new and improved structural details as well as the retrofit assessment for existing structures.
Vikas Patwardhan, Babushankar Sambamoorthy, Tuhin Halder – Lear Corporation
Federal Motor Vehicle Safety Standard 202 applies to automotive seat head restraints, and their attachment assemblies. The regulation is aimed at reducing the frequency and severity of neck injuries due to huge forces resulting from vehicle crashes. The main objective of this paper is to discuss the LS-DYNA IMPLICIT code vis-à-vis the simulation of the proposed FMVSS 202 regulation. The proposed changes to the FMVSS 202 standard incorporates a permanent set requirement for the head restraints. Since a quasi-static FE simulation cannot do permanent set calculations, LS-DYNA IMPLICIT code was used for this purpose. The paper explains in detail about the setting up of a seat model for the test, the various modeling techniques used and the correlation of results for the seat model.
Dr. R. Borg – Engineering Research Nordic AB
Metal-Plastic-Metal (MPM) sheets consists of two metal sheets bonded together by a thin polymer layer. Modelling this sheet structure with a single shell element gives a too stiff response. The large shear deformation in the polymer layer can not be represented correctly. A method is presented on how to simmulate the forming process of MPM-sheets by which the two metal sheets are modelled with two layers of shell elements. The bonding polymer is not modelled as elements. Instead its effect on the structure is implemented as a Tie-Break contact algorithm. The shell thickness is necessarily accounted for in this contact interface such that the midsurfaces of the metal plates are possitioned correctly with respect to each other. Results from simulations and experiments are examined and forming criteria such as wrinkling, necking and failure are discussed.
Manfred Bischoff, Frank Koschnick, Kai-Uwe Bletzinger – Technische Universität München
The Discrete Shear Gap Method, initially proposed for the elimination of transverse shear locking in plate and shell finite elements is extended to a more general concept, applicable to other locking problems, typically causing severe trouble in structural analysis, especially in the case of thin-walled structures. The outstanding feature of the proposed formulation is the fact that one unique method is used to avoid various different kinds of locking phenomena. It is applicable to beam plate and shell elements, but also to two-dimensional and three-dimensional solid elements. The fact that approximation quality is often subject to strong sensitivity to mesh distortion can be alleviated with the help of stabilization methods.
Klaus Hessenberger – DaimlerChrysler AG
To guarantee proper function of the seat belt system, belt anchorages have to resist defined static test loads that represent an vehicular impact. ECE R14 and FMVSS210 are tests to ensure sufficient strength of all anchorage points. In these tests high forces are applied to the seatbelts over loading devices. All components of the sytems, namely seats, seat and belt anchorages have to resist the defined loads without damage. The loads are applied slowly and are sustained over a long period of time, so one can assume a quasi static test. The correct modelling and simulation of the complex load application system is essential for significant and accurate computational results. The experimental test with an existing drivers cab according to FVMSS 210 was simulated with Abaqus Standard (implicit) and LS-Dyna (explicit). During the application of both tools, problems specific to each system were encountered. In Abaqus, problems were caused by large deformations of the sheet structure and possible local buckling phenomenons. In the LS-Dyna calculations the presence of dynamic effects have to be minimized to yield a good correlation with the quasi static tests. The problems encountered and the approach used are presented and a comparison between test and analysis will be given.
Won Yong-Hee, Kim Jeong-Ho, Lee Young-Jun – LG Cable Ltd
The conventional progressive stamping process used in producing connector pin makes the pin-shift from side to side due to elastic recovery reffered to springback phenomenon. As the solution of this problem, the reverse stamping process was introduced to the production process of connector pin in LG Cable. In this study, we performed research on the feasibility of reverse stamping process to support a design guide of which process. For this research, we tested specimens made of the connector pin materials, and built up the finite element model to simulate the failure mode of the section of the connector pin. The nonlinear analysis for these stamping process was performed using LS-DYNA. Through these nonlinear finite element analysis, we suggest the level of lifter force letting the section shape of the connector pin better and we used LS-DYNA for the practical approach than theoritical appoach.
Seong-Sik Choi, Jeoung-Gwen Lee – Samsung Electronics, Korea
Display performance of LCD is going to be continuously high definition, high brightness, wide viewing angle. On the other hand, outline dimensions are demanded to be slimer and lighter. As the result of above demand, the space of each part becomes gradually narrow and tight. Therefore, the importance of design to prevent weakness for impact is embossed greatly. LCD products have to be subjected to various impact test for consumer’s using environments. During the design, toolcorrection or modification caused by damage or large deformation of weak part give rise to time consumption and cost-up. For development of LCD-TV module, these problems become more important because of its large size and heavy weight. To improve these problems, we performed impact analysis of 40 inch LCD-TV module using LS-DYNA and applied the results to the development. In this analysis, we found out the weak regions and obtained improved reliability after design modification. Impact analysis using LS-DYNA is going to be applied in the all of LCD product developments hereafter, and development period reduction and reliability improvement are highly expected.
Tsuchida T., Yamamoto S. -Toyota Motor Corp., Isomura K. – Toyota Industries Corp.
In an evaluation of crashworthiness for the cars that are made of aluminum alloys, the evaluations that consider a fracture phenomenon come to be needed because conventional aluminum alloys have low fracture strain (10 – 20%). Since an original damage & fracture material model of LS-DYNA, namely MAT_PLASTICITY_WITH _DAMAGE: MAT81 has a damage & fracture characteristic in case of compressive strain state, real collision phenomena can not be simulated in some cases. Therefore, we reviewed the damage & fracture criterion of this material model. We newly introduced some sort of a damage & fracture criterion into the MAT81 of LS-DYNA V960 in later revision and performed crashworthiness evaluations for an aluminum car using this improved damage & fracture model. This criterion has nondamage & non-fracture characteristic in compressive strain state and it is known as “Orthogonal an-isotropic (Orthotropic) damage & fracture model”.
Zeguer, T. – Jaguar Cars
Hyunsup Kim, Sungoh Hong, Seokgil Hong – KIA Motors, Hoon Huh – Korean Institute of Science and Technology, Korea
This paper is concerned with a crash analysis for vehicle structures considering forming effects. The properties of formed vehicle structures have been effected and changed by such as work hardening, non-uniform thickness distribution resulted from the forming process. The crash analysis of vehicle structures with the forming effects leads to different results from those without such effects. In this study, the forming effects of the front side member assembly of vehicle structures are considered. The plastic strain and thickness distribution from the simulation results of the forming process have been used as an initial condition for crash analysis. The crash analysis of a full vehicle structure has been performed with those forming effects, and the results are compared to those without the forming effects. The deceleration pulse and deformation from the results are calculated and investigated in order to identify forming effects. Analysis results demonstrate that crash analysis of vehicle structures with forming effects can be more effective for accurate approximation of deceleration pulse and deformation mode.
Dr. Herbert Cornelius – Intel EMEA
M. Arad, D. Touati , I. Latovitz
The effect of yaw on the ability of an eroding long rod projectile to penetrate oblique thin targets was investigated. Numerical simulations of an eroding long tungsten rod projectile penetrating an oblique thin steel plate target were conducted using the LSDYNA code with a user-written subroutine. The numerical results were found to agree with the experimental data. From the simulation results it may be concluded that following penetration of a thin target, for non-zero initial yaw values, the projectile nose bends toward the velocity vector, while for zero yaw the bending is negligible. In addition, for a non-zero initial yaw angle, the side of the projectile pointing in the direction of its velocity is damaged (this side is in greater contact with the target because of the velocity vector direction).
Chris Carmody – MG Bennett & Associates Ltd, Gaetano Burriesci – Sorin Biomedica Cardio, Ian Howard, E.A. Patterson – Sheffield University
Cleve Ashcraft, Roger Grimes, Brad Maker – LSTC
Martin Wilson, Ben Ricketts
Bombardier Transportation is the largest manufacturer of rail vehicles in the world. The current product portfolio includes a wide variety of vehicles from low speed ‘people movers’ through to high speed inter-city trains. Bombardier offers products in every sector of the passenger rail equipment market and therefore is required to meet a number of national and international crash safety requirements. These requirements range from simple static collapse loadcases to full collision events with other rail vehicles and obstacles. As part of the validation procedure for new designs, finite element (FE) models are produced to simulate new vehicle crash performance against targets set by these requirements. The simulation of bolted and welded aluminium structures is particularly important for the Carbodies part of the business, since recent ‘real life’ crash cases have shown bolt failure and weld ‘unzipping’ as critical collapse modes for extruded aluminium carbody designs. The current technique used for modeling welded aluminium sections and particularly the heat affected zone (HAZ) is presented. Bolt failure modeling within large structures is also addressed and results are presented from calibration tests and simulations carried out to evaluate the failure behaviour of Huck Bolted connections. This paper presents an overview of the current state of the art in the rail industry and describes, through various case studies, the approach that Bombardier Transportation uses for validation of new vehicles. These case studies also show novel aspects of new vehicle design, which increase safety and highlight the commitment of Bombardier Transportation to a ‘design for crashworthiness’ approach to new passenger vehicles.
Ming-Chang Yang, Tien-Chi Tsai – Metal Industries R&D Center, Taiwan
The authors would like to present the whole procedure of using dynamic explicit finite element method to aid multi-stage miniature stamping die design. How to cost down and time saving are the key issues for tool maker in Taiwan. It has been applied to many automobile components and metalwork successfully by computer aided design and engineering analysis. In this paper, we dedicated our application of these technologies to small scale stamping parts. The trick of mass scaling for speed up the computation of LS-DYNA and control its effect in dynamic behavior for miniature blank sheet will be described in this paper. Die designers can operate die try out many times during one day on their desk top personal computer for their different processes changed. A mini scale ball bearing retainer stamping part was demonstrated. The history of metal flow and its thickness variation for pre-form stage and finished-form stage in real test were matched with the simulation results. It has been approved that the accuracy of numerical result is good for miniature sheet metal part. By this way, we have saved many try and error tests for die modify and try out in reality.
Friedhelm Birk – 3Dims GmbH Frankfurt, Germany
Since simulation of complex things like crashing a car becomes more precise and more affordable its only natural that one wants to analyze the results in a way a humans nature can intake complexe data best: visually.