The current state-of-the-art FEM crash models of vehicles are built using advanced CAD and FEM tools as a collaborative endeavor of engineers, modelers and computer specialists. The models have very detailed geometry discretization, contain large of number of sub-models, parts, components, and systems with complex properties, interaction, connectivity, spatial and functional relations. Describing and documenting such a model in detail becomes an equally complex undertaking. The emerging web-based technologies provide a framework that can assist in presenting the information contained in the model and thereby facilitate its distribution and wide utilization. Most of the current approaches to web-based documentation of the computational models are based on static information that primarily consists of document libraries, presentations, simulation results and model input files. Although they provide considerably more information compared to the conventional, printed documents, they do not fully take advantage of the potential offered by the new technology. One of the most powerful features of the interactive web is the ability to interlink the data in hypertext documents such that they mimic the relations that exist in the model. Navigating the interlinked data is then very similar to navigating the actual FEM model. The sequence of steps for model examination is not fixed, so that the user can have certain level of control how to examine the information and acquire the knowledge. The focus of the interactive documentation system is how to effectively utilize the technology and find the correct balance of simplicity necessary for a wide use and ability for in-depth inquiry into the model demanded from the expert users. Ideally, the system will provide an ability to configure access mode to match the level of user's expertise. It should also be able to support model development process such that the evolving changes to the model are documented and can be manipulated similarly to software development systems. Such capabilities are outside the scope of this project but are planned be pursued in the future.
The web-based manual developed during this project is an attempt to develop such a comprehensive documentation system, albeit of a limited functionality commensurate to the project's scope. Although it started with an objective to document one vehicle model, the project resulted in several models for tractor and trailer, which can be combined into combination vehicles for crash simulations. The initial approach for a web-based documentation was based on the interactive manual previously developed for the Single Unit Truck (SUT) FEM model . That version was primarily developed using direct generation of the content and presentation programs by the developers. Such an approach obviously necessitates a lot of human effort and it scales proportionately to the number of the models to be processed. As the models change during the course of their development, detecting, updating and tracking the differences becomes quite tedious and time-consuming. In order to deal with an increased number of models and their variations, we have decided to develop a system that can accommodate these requirements and can process the models in relatively short time. In doing so, we have started on a path towards a more general system for FEM model documentation that is relatively independent of the models under consideration.
A major difficulty in developing a more general interactive FEM model documentation system is the variety of existing file formats, model syntax, model building procedures and specializations within models for particular vendor's simulation software. Several efforts on standardization of FEM models have been initiated in the past and have resulted in some standardization primarily for geometric representation. However, connection and interaction between parts in FEM models is still specialized to the end simulation software. New technologies such as Extensible Markup Language (XML), the Extensible Style-sheet Language Family (XSL), and XSL Transformations (XSLT) bring forward new tools that can be used to simplify transformations between FEM models from different vendors and simulation types without a need to enforce common model syntax.
The web documentation system developed in this project consists of three major components. The first component is the underlying data model and relational database that describe the FEM model information and relations within the model. One of the possible approaches is to work directly with the ls-dyna crash files and dynamically extract its content as queried by a user. An obvious drawback is the size of the models, which if anything are going to grow as the model become more sophisticated. Therefore, some underlying condensation of information and simplification of the model is necessary if a system is to be reasonably fast. The second component is the multi-media content generation software to be used for presenting the characteristics of the model and relations within it. The last component is a web-based interface for querying and displaying the information. In the previous version of the web-based manual for the SUT model , all of these components were interwoven into one and could operate primarily on static information that was not easily extendable. In this new version, adding a new FEM model to the documentation system does not require any modification of the display system, rather it takes only generation of the model data and media, images, Virtual Reality Modeling Language (VRML), movies, etc., and their entry into the data model. In this way, the content is disassociated with the presentation engine and the two can be developed separately. In the following, the three components are described in more detail. The web documentation system can be tested at the provided web address.
Data Model and Relational Database
The most detailed examination of the tractor-trailer FEM ls-dyna model can be performed by using the commercial pre- and post- processors such as Livermore Software Technology Corporation software PrePost. However, that requires familiarity with the software and limits the model presentation to the expert group of users. Large-scale FEM models are usually developed as proprietary products for specific application and are not meant for wide distribution. On the other side, the FEM models developed by academia and government institutions are usually in public domain and their target audience is intentionally wide. World Wide Web technologies developed for variety of purposes can be adopted and applied to the documentation of FEM models. One of the most potent technologies is the Extensible Markup Language (XML), and related tools. Coincidentally with development of the markup languages, the syntax of FEM models was developing from strictly formatted numerical sequences (decks, with origins in card format), to the representation akin to markup languages. While some idiosyncrasies remain in the FEM model syntax (likely to be eliminated in new versions of the codes), the transformation from the current FEM model formats, such as KEYWORD format in ls-dyna, are relatively simple. Once a transformation to XML form is made, we have at our disposal an enormous toolbox for manipulating and transforming data. We have therefore selected the XML for the data model underlying the interactive documentation system. The relations existing in the model are encoded in the XML form and a relational database for the model is created. Data transformations are performed by a series of programs that extract from the FEM model file pertinent data, process that data, build cross-link references and links to media content (images and interactive 3D files), and finally store all that information in XML files. In essence these files represent a relational database stored in a format that can be viewed and edited in any text/word processor. This database is then used for various queries defined in the user interface. Although quite flexible and portable, such a system lacks management and security features provided by an integrated database management system such as mySQL. Hence, care must be taken on the server side to guard this content from unauthorized access and tampering. Initially, the system was driven by an mySQL engine, but due to the desired portability of the system, the mySQL database has been entirely replaced by the XML. Future plans for development include revisiting the concept of separate database server and/or combinations of transformations and XML-based databases. However, the current implementation considerably improves upon the SUT manual design in flexibility and generality of the system.
Multi-Media Content Generation
The wide spread of WWW have been driven in large part by the ability to effectively convey information using rich media content and describe relations within it by interactive, hyperlinked documents. Spatial relationships within the model are best described by images and 3D models, ideally to be manipulated by a prospective user. User's multi-media expectations have also grown in proportion to increased communication bandwidth and capability of computer hardware, so that the amount of media to be generated and presented is an order of magnitude larger than what was common just few years ago. We have developed computer programs to create image libraries and interactive 3D files for selected FEM model components. Additional structures on data may be imposed, such as grouping of parts into components and sub-systems as they are commonly grouped in the actual vehicles. The parts or part sets that may be grouped together by their physical association to a certain model component, or by utilizing specific material, section, or contact type.
A series of scripts have been written in various computer languages to generate images for the documentation. The programs used freeware programs, such as Medit and Persistence of Vision Raytracer to capture images of FEM model entities from different points of view and in different orientations. A series of images displaying the entity of interest both as standalone and as being a highlighted part of the model shows in detail what the entity looks like, and where it fits in the model. Another set of programs extracts from the FEM model data for creating the interactive 3D views. In order to reduce the size of the models, Level of Detail programs are used to minimize the number of polygons necessary for the display. We have also implemented graph presentation of data using public domain software and correlated visual and tabular representation of data to the exact location of their definition in the FEM model input file. Data retrieval and web display
The tractor-semitrailer web server was built on CGI interface that receive and process user requests, interact with the XML database type files and serve requested content. The available content is organized in six sections accessible through the horizontal navigation.
The Home section contains a summary of the project and excerpts from all chapters in the final report.
The display capability for the FE model offers model selection and includes briefs for the following LS-DYNA keywords: PART, SECTION, MATERIAL, DEFINE, CONSTRAINED, AIRBAG, CONTACT, and SET. Addition of other ls-dyna keywords is relatively straightforward. This subset was selected as it describes the essence of the model. A user first needs to select the model for analysis from the available database of models. The subsequent database queries by selecting options and links in the browser are conducted on the selected model. The presented data is stored in tables and cross-linked appropriately. For example, user has a link to view all the parts that are modeled with the selected material model, or view parts that share common section model, etc. A more detailed data view of the selected entity offers the user an opportunity to:
- launch a pop-up window to view the excerpt of the FE model file that served as an information source;
- view a series of static entity images (applicable for parts, sections, materials, contacts and airbags);
- use an interactive 3D application to get a more detailed model view (applicable for parts, sections, materials, contacts and airbags;
- launch a pop-up window to get curve graphs (applicable for tables and curves);
- launch a pop-up window to look at the LS-DYNA user manual for a particular keyword.
The Simulation section offers movies showing FE model simulations performed with individual tractors or tractor-semitrailer combinations. Cross links offer speedy connections to downloadable models featured in the simulations.
The Test section displays movies showing recordings of actual crush tests performed with tractors and tractor-semitrailer vehicles.
The Download section offers links to tractor and tractor-semitrailer FE models. Cross links enable speedy connections to simulations featuring these models.
The About section contains general information about the project and people that designed the web based interface as well as links to the web sites of the institutions that participated and financed this project.
This process results in a dynamic web server system that can continuously be updated as new versions of the FE model evolve, or to add new keywords and database query templates.