Fig。2 Three Stages of CAE Modeling towards the Design of a Crashworthy Vehicle

for LS-DYNA (LSTC, Livermore, CA) simulations。 A least square based assessment index was then used as the objective function for minimization。 Finally, material constants can be determined from response surfaces generated[4]。 Figure 3 shows similarities between the actual tests and simulation results once the force-deformation curves are matched。 This material can then be used computationally for Step 2。 InStep 2, emphases are placed on finding the best coarse- meshed models to absorb the most energy with least deformation while using least amount of material to reduce overall weight。 Iterative procedures are used in this step and no examples will be provided in this article。

Fig。3 Comparison of Failure Mode and Deformation Pattern of a Magnesium Alloy Box Sujected to Four-point Bending

Magnesium and its alloy have been used in vehicular structures to provide strength while simultaneously improving fuel economy。 Mechanical properties of this material are strain-rate dependent and constitutive modeling of the failure of this material is currently not available。 One way to work around this lack of constitutive modeling is using other existing material laws in conjunction with optimization methods to determine the best fit material constants。 Flow stresses, obtained from tensile coupon tests of magnesium alloy, at a constant plastic strain of 5% were used to determine the parameters governing strain rate effects。 A design of computer experiments (DOCE) method was used to select sampling points

In general, injury to the lower extremity is not life threatening but can result in functional loss, impairment, and incur great societal cost。 The knee- thigh-hip (KTH) complex has been studied extensively to determine the injury mechanism and corresponding threshold。 However, interactions between the KTH and various vehicular interior design parameters, such as knee bolster stiffness and seat belt load limiter, have not been investigated thoroughly in frontal automotive crashes。 More critically, available studies are based on crash dummies which only provide the risk of injury

第4期 King H。 Yang：Application of CAE Technology in the Development of Safer Vehicles 397

to the region, such as the tibia or femur, instead of clinically relevant terms such as split condyle fracture, femoral neck fracture, etc。 A 3-D FE model of a 50th percentile male KTH complex, which includes explicit representations of the iliac wing, acetabulum, pubic rami, sacrum, articular cartilage, femoral head, femoral neck, femoral condyles, patella, and patellar tendon, has been developed to simulate injuries such as particular fractures that cannot be predicted by crash test dummies。 The model was then validated against a range of experimental data available in the literature。 The validated model was then integrated into an upper torso model and validated against data obtained from whole body sled tests。

Clinically, different types of fracture present different levels of difficulty when surgeons are trying to repair functionality。 For example, a mid- shaft femur fracture is much easier to repair than an acetabular fracture。Fromavehicle design point of view, the load limiter force, percentage of seat belt elongation, inlet amount of the pretensioner, knee-knee bolster distance, knee bolster angle, knee bolster stiffness, toe board angle, and crash velocity with associated pulse all affect the potential risk of lower extremity injury。 Using design of computer experiments (DOCE) methods based on Taguchi’s approach, each of these eight parameters were set to have either two or three values to choose from。 The integrated model was run through these cases and best design levels for vehicular interior design parameters, which minimize the potential risk of KTH injuries, were revealed。 This study found a combination of a load limiter force of 6 kN, 11% of seat belt elongation, 80 mm of pretensioner inlet amount, 130 mm of knee-knee bolster distance, 50° of knee bolster and toe board angle, and 60 psi of knee bolster stiffness,