In FE modeling, the quality of the mesh is a critical issue if accurate predictions of the modelare of concern。 This is particularly true for explicit finite element modeling in which the one-point reduced integration scheme is the default integration method。 The explicit scheme decreases computational cost while at the same time improving solution accuracy byavoiding mesh locking。 Inthree-dimensional(3-

D)modeling, 8-node hexahedral elements can yield

King H。 Yang：Application of CAE Technology in the Development of Safer Vehicles 395

(Left) Block Structure of an Adult Brain, (Middle) Hexahedral Meshes of a Mid-resolution FE Model, and (Right) Hexahedral Meshes of a High-resolution FE Model。

better simulation results compared to degenerated 4-node tetrahedral or 6-node pentahedral elements。 Factors that can affect the quality of a mesh include: Jacobian, warpage, aspect ratio, skew angle, and internal (edge) angle。 Another key issue that ensures accurate simulation results is mesh convergence。 This problem has been overlooked, in part due to the fact that changing mesh density requires significant time and effort, especially in a 3-D space。 A new method of using multi-block approach has been used successfully in the development of hexahedral FE mesh in order to easily increase element density while maintaining high quality throughout the model [3]。 Figure 1 shows an example of using this method to develop high quality FE meshes with ease。

A hierarchical approach is needed when using CAD and CAE tools in the design of a crashworthy vehicle。 Based on this approach, FE models with varying degrees of complexity are developed and validated in three stages。Before developmentof a new car is considered, one must define design boundaries once market research is completed and target price ranges and safety ratings are determined。 In the first stage, CAE engineers are required to check the database to determine if mechanical properties of

any special materials to be used in the current design are not available in the CAE material library。 A simple example would be the use of light weight magnesium to improve the corporate average fuel economy (CAFE) rating。 As a novel material, its properties are not as well characterized as other materials like steel and aluminum。 CAE modeling and mechanical testing of the material sample will be needed in order to accurately calibrate the material properties, including strain rate dependency。 In the second stage, the energy absorbing capabilities of the main energy absorption components are calculated using a coarsely meshed CAE model。 The model-predicted crash pulse is then compared to an idealized crash pulse that is defined based on previous experiences or rigid body dynamics modeling。 Changes are made to the design in order to better match the idealized crash pulse。 Prototypes consisting of these main energy absorption components are made and crash tested to determine if the coarse-mesh model predictions match those obtained experimentally。 Iterative procedures are needed to finalize the design of these structures。 In stage three, a detailed model of the whole vehicle, including all interior components, fenders, power train, bumper, etc。, is developed based on the final iteration of the main energy absorption components。 Occupant and restraint models are also integrated to

396 汽车工程学报 第 1 卷

the detailed vehicle model and occupant responses calculated。 The theoretical safety rating will be evaluated from occupant responses in the model。 If the model-predicted rating does not meet expectations, changes to the restraint parameters need to be altered until the rating is satisfactory。 Several prototypes are then made and crash tested to determine if the safety rating obtained from physical tests concurs。 Only then, the design can be finalized for mass production。 Figure 2 shows the three stages of this set of processes before safety of a car design is finalized。论文网