Developments and Applications of Structural Optimization and Robustness Methods in Vehicle Impacts

Finite element based full vehicle structural crash simulation is an analysis tool commonly used in automotive industry to evaluate vehicle impact performance. As the simulations are computation intensive, special optimization methods and processes are often required. This paper presents recent developments and applications of structural safety optimization and robustness methods for vehicle crashworthiness. It addresses advanced methods in gauge, size, shape, topology optimization, and robust, reliability-based design optimization methods. Recent applications in vehicle safety design are presented and discussed.

Multivariate Knock Detection for Development and Production Applications

Combustion knock caused by end gas autoignition continues to be a limiting factor in the performance of automotive internal combustion engines. As such, the availability of efficient knock detection methods is a prime requirement for the optimization of engine mapping and control. Current production knock control systems are based on the measurement of mechanical vibration induced by the acoustic resonance excited in the combustion chamber during autoignition. These vibrations are measured using accelerometers on the engine block. Conversely, knock detection in the laboratory environment during engine development or calibration generally involves either acoustic methods or acquisition of in-cylinder pressure. The purpose of this study is to develop an improved multi-transducer vibration-based knock detection method with applications in engine development and production. The possibility of replacing the pressure-based detection methods in the laboratory environment presents many advantages relating to cost and efficiency. Moreover, the economy of a vibration-based system coupled with improved correlation to laboratory methods represents great potential for performance improvements if applied to production applications.

Shell Elements Performance in Crashworthiness Analysis

Crashworthiness analysis, a type of large deformation transient dynamics, has been an important and active area of researches and engineering applications. Several shell elements have been implemented in the finite element software for crashworthiness analysis. Among them, the 4-node quadrilateral Belytschko-Tsay element, using lower order integration technique is most commonly employed, due to its efficiency, robustness and overall accuracy. However, the lower order integration brings in some uncertainty. This paper is to conduct an engineering evaluation on performance of various shell elements, including Belytschko-Tsay, Belytschko-Leviathan (QPH), Bathe-Dvorkin, discrete Kirchhoff triangular elements, available in the commercial explicit finite element software. The study uses several linear and nonlinear benchmark examples and high-speed impact examples, to investigate the performance of these elements. Results of engineering interest and efficiency of computation are reported. Also, the behavior of finite element convergence, observed from the results by a sequence of refined meshes is investigated.

Structural Design Optimization of Side/Rear Impact Guards Using a Coupled Genetic Algorithm Finite Element Method

Underride protective structure can reduce serious injures when passenger cars collide with the rear end or side of the heavy vehicle. This paper describes the use of Genetic Algorithm (GA) coupled with a dynamic, inelastic and large deformation finite element (FE) code LS-DYNA to search optimal design of the Side/Rear impact guards. In order to verify the accuracy of the FE model, the simulation results were compared with real experiments follow with the regulation ECE R73. The validated FE model then used to study the optimal design base on under running distance and total amount of energy absorbing capacity. The results from this study shown that this newly developed method not only can found multi-objective design parameters but also can reduce computational time significantly.

Energy-Based Crashworthiness Optimization for Multiple Vehicle Impacts

In this paper, we use a full-scale finite element vehicle model of a 1996 Dodge Neon in simulating two types of vehicle crashes, offset-frontal and side impacts. Based on an analysis of the vehicle’s histories of internal energy absorption under both impacts, we select twenty components as design variables in the optimization of the vehicle’s weight without decreasing the vehicle’s energy absorption capacity and energy absorption rate. We use the second-order polynomials in creating the metamodels for the response functions of energy absorption under both impacts. The optimization result shows a significant reduction on the total weight of the selected components. The LS-DYNA MPP v970 and a full-scale finite element vehicle model of 320,872 nodes and 577,524 elements are used in the simulations. A simulation of 100 ms offset-frontal impact takes approximately 17 hours with 36 processors on the IBM Linux SuperCluster, which has a total of 1038 Intel Pentium III 1.266 GHz processors and 607.5 GB RAM. A simulation of 100 ms side impact takes approximately 29 hours with the same condition as the offset-frontal simulation.

Numerical Analysis of the Biomechanical Characteristics and Impact Response of the Human Chest

The mass density, Young’s modulus (E), tangent modulus (Et) and yield stress (σy) of the human ribs, sternum, internal organs and muscles play important roles when determining impact responses of the chest associated with pendulum impact. A series of parametric studies was conducted using a commercially available three-dimensional finite element (FE) model, Total HUman Model for Safety (THUMS) of the whole human body, to determine the effect of changing these material properties on the impact force, chest deflection, and the number of rib fractures and fractured ribs. Results from this parametric study indicate that the initial chest stiffness was mainly influenced by the mass density of the muscles covering the torso. The number of rib fractures and fractured ribs were primarily determined by E, Et and σy of the ribcage and sternum. Similarly, the E, Et and σy of the ribcage, which is defined as the bony skeleton of the chest, and sternum and E of the internal organs contributed to the maximum chest deflection in frontal impact, while the maximum chest deflection for lateral impact was mainly affected by the E, Et and σy of the ribcage.

A Comprehensive Evaluation of NHTSA Rollover Test Data for Use in Computational Model Validation

Increased computational power and new software have brought occupant motion simulation into the mainstream for vehicle accident reconstructionists. Using programs available today, investigators are able to achieve numerical results that match actual physical results with a high degree of accuracy. It should therefore be possible to validate the performance of a software simulation using instrument data collected from a real vehicle test. For valid results, however, one must have valid data upon which to base the simulation. We attempted to validate MADYMO occupant motion simulation software by using data from the National Highway Traffic Safety Administration (NHTSA) vehicle crash test database for vehicle rollovers. In the course of our work, we discovered flaws in the NHTSA database that rendered it useless for both validating and disputing a computational simulation. These flaws included data that did not match the descriptions of vehicle travel in the written reports, entire channels of missing data, and others. NHTSA crash tests are often cited as reliable sources of data in vehicle crash situations. While not disputing the limited scientific value of these tests, this paper documents the problems with NHTSA test reports and concludes that the data contained therein can be unintentionally misleading and of little value for computational model validation of rollover simulations. This paper also presents testing improvement procedures that should allow a greater correlation of computational and testing data.

Dynamic Crush Behavior of Aluminum 5052-H38 Honeycomb Specimens Under Out-of-Plane Inclined Loads

The dynamic crush behavior of aluminum 5052-H38 honeycomb specimens under out-of-plane inclined loads is investigated. Honeycomb specimens were designed to minimize the secondary stresses under out-of-plane inclined loads. A test fixture was designed such that inclined loads can be applied in dynamic crush tests. A static linear elastic finite element analysis was performed to understand the stress distributions in honeycomb specimens under inclined loads. The computational results show that the secondary stresses of the specimens are limited to the region near the stress-free boundary. The results of dynamic crush tests indicate that the effects of the impact velocity on the crush strengths are significant. Under dynamic loads, as the impact velocity increases, the crush strengths increase. The trends of the inclined crush strengths for specimens with different in-plane orientation angles as functions of the impact velocity are very similar to that of the pure compressive crush strength. Honeycomb specimens under pure compressive and inclined loads show similar progressive folding mechanisms. The similar trends of the crush strengths as functions of the impact velocity are possibly due to the similar progressive folding mechanisms.

A Partially Validated Finite Element Whole-Body Human Model for Organ Level Injury Prediction

A finite element whole-body human model, which represents a 50th percentile male, was developed by integrating three detailed human component models previously developed at Wayne State University (WSU): a thorax model with detailed representation of the great vessels [1], an abdomen model [2], and a shoulder model [3]. This new model includes bony structures such as scapulae, clavicles, the vertebral column, rib cage, sternum, sacrum, and illium and soft tissue organs such as the heart, lungs, trachea, esophagus, diaphragm, kidneys, liver, spleen, and all major blood vessels including the aorta. In addition to model validations already reported at the component level, the new whole-body model was further validated against two sets of experimental data reported by Hardy [4]. In these experiments, human cadavers were loaded either by a seatbelt or by a surrogate airbag about the mid-abdomen, approximately at the level of umbilicus. It is believed that exercising a validated human model is an inexpensive and efficient way to examine potential injury mechanisms. In some cases, this can provide insight into the design of subsequent laboratory experiments.

An Integrated Robust Design Method for Occupant Restraint System

Computational analysis of occupant safety has become an efficient tool to reduce the development time for a new product. Multi-body computer models (e.g. Madymo models) that simulate vehicle interior, restraint system and occupants in various crash modes have been widely used. To ensure public safety, many important injury numbers, such as head injury criteria, chest G, chest deflection, femur loads, neck load, and neck moment, are monitored. In the past, deterministic optimization methods have been employed to meet various safety regulations. Further emphasis on product quality and the consistency of product performance, uncertainties in modeling, simulation, and manufacturing, need to be considered. There are many difficulties involved in the optimization under uncertainty for occupant restraint systems, such as (1) highly nonlinear and noisy nature of occupant injury numbers; (2) large number of constraints; and (3) computational intensity to obtain the statistic information of injury numbers by the traditional Monte Carlo method. This paper investigates an integrated robust design approach for occupant restraint system by taking advantages of design of experiments, variable screening, stochastic meta-modeling, and genetic algorithm. An occupant restraint system is used as an example to demonstrate the methodology, however, the proposed method is applicable for all occupant restraint system design problems.