Our website uses cookies. By using the website you agree ot its use. More information can be found in our privacy policy.


Modeling of component failure due to notch effects in press-hardened steel caused by mechanical and thermo-mechanical joints under crash load

The increasing application of press-hardened steel in combination with aluminum sheets in the construction of car bodies results in the use of mechanical joining techniques such as self-piercing riveting and thermo-mechanical joining techniques such as resistance element welding. These joints generally represent a notch within the component. The cause of the notch effect is different for the investigated joining techniques and can be distinguished in a geometrical notch and a metallurgical notch. Riveted joints result in a pierced hole with high plastic strains at the edge and thus represent a geometrical notch. Thermo-mechanical joints in press-hardened steel result in a softened heat affected zone (SHAZ) around the weld due to the applied heat during the joining process.

Using history variables in materials to reduce modelling effort and increase model accuracy

In crashworthiness simulation the definition of material properties is one of the key aspects to obtain reasonable results. However, a lot of materials come with properties that either change locally or are generally of stochastic nature. Additionally, production processes (e.g., welding) might change the behavior of certain materials. To overcome the necessity of defining an individual part for each region where material properties differ a new approach was developed. With the new keyword *DEFINE_TABLE_COMPACT it is now possible to define material properties by means of a multi-dimensional table with arbitrary variables controlling for example the plastic flow curve or the damage behavior. Secondly, the keyword *INITIAL_HISTORY_NODE enables the user to set these variables individually on each node in the model. This presentation shows possible applications of this approach and the benefits on parametric modelling and simulation.

Damage modeling of aluminum casting components considering defect distribution for crashworthiness prediction

Aluminum die casting components are widely used in vehicle constructions because of their good compromise between weight reduction and improvement of mechanical properties. The complex geometries of these components with inhomogeneous defect distribution are a relevant issue, as material with higher defect content shows lower fracture strain. It makes the analysis of the damage behavior for crash simulation more challenging. An extensive experimental investigation is required to quantify the scatter as well as the development of a suitable material model to describe it.

Creating a complete crash model with GNS. High accurate Barriers, handy preprocessing with dummy positioning and fast tailored result analysis

The current transformations in the automotive industry are forcing all car manufacturers to check their processes constantly and repeatedly for economic efficiency. Regarding the areas of body development, vehicle safety and occupant as well as pedestrian protection, economic efficiency can only be achieved through a high degree of automation in virtual product development. Such a degree of automation is only feasible, if the software used and the integrated models are well coordinated. In this presentation it will be shown how a high precision barrier model and several dummy models can be combined and positioned for any vehicle model within the pre-processor Generator4 with ease. Both direct positioning tools and simulation driven ones can be fully automated in a simple way. Information on positioning and kinematic relationships can be taken directly from the input files. This simplifies the following processes and reduces the error rate due to less manual data input. Generator4 is then able to directly start the simulation with LS-Dyna or other solvers.

Gradient enhanced damage: modelling, implementation and applications

Finite element (FE) simulations with constitutive models for softening materials, such as in the case of standard continuum damage mechanics based approaches, suffer from pathological mesh sensitivity as a result of strain localisation into a single element row. To overcome this major drawback, the local damage has to be enhanced towards nonlocal damage evolution. A suitable method for this purpose is the integral nonlocal formulation, available in LS-DYNA® by the keyword *MAT_NONLOCAL. However, its costly underlying search algorithm can result in a strong increase of the simulation time, leading to an impractical application for engineering problems.

Parameter Identification of Coating Parameters to Improve Webbing Bending Response in Passive Safety Crash Simulations

Seat belt is one of the main load bearing parts for restraining an occupant in a vehicle crash. Thus, accurate modelling of seat belt is important to achieve realistic interaction between belt to Anthropometric Test Devices dummy model in passive safety crash simulation. The belt modelling in the lap area is even more challenging because it also bears out-of-plane load during interaction with the pelvis, causing bending in webbing. Inadequate modelling of the bending response often results in rope-like effect in the lap belt during passive safety crash simulations, causing loss of contact area and eventually incorrect pelvis coupling. Such a behavior of belt is often observed with the application of THOR dummy in crash simulations, leading to an argument that simulations are not able to correctly predict submarining (slippage of belt over the pelvis to load the abdomen) and eventually incorrect estimation of the pelvis iliac forces and moments on the dummy [2]. Therefore, concerns are growing for improving the belt modelling. Besides other modelling aspects (e.g., mesh size, contact modelling, directional dependency of friction etc.), it is believed that including appropriate bending stiffness could improve dummy-to-belt interaction. There are several options available in LS-DYNA to model bending in webbing but the inevitable use of 2D slipring for its robustness and efficiency in the system models poses limitations.

Experience a complete virtual crash and safety laboratory with the aid of the ANSA pre-processor

With safety protocols and regulations becoming increasingly enhanced, safety analysts try to keep up replicating all possible crash event scenarios in laboratories using specifications that frequently change. In this pursuit, it is crucial for analysts to have at their disposal accurate and robust digital models that enable the tune and study of any real crash event parameter. Through the ANSA pre-processor, BETA CAE Systems offers an extensive crash and safety portfolio of automated tools for simulation modelling, to create a complete “virtual crash and safety laboratory”. Such tools include this of the seat and the dummy guide, from the identification of HPOINT of the seat to the positioning of the coupled restrained seat-dummy system according to a regulation or a test data position, available not only for standard crash dummies but also for human body models. Pedestrian and Interior tools ensure the proper marking of the exterior and interior of the vehicle but also the accurate positioning of the headform to the desired targets. The Impactor tool enables the positioning of barriers/impactors according to all available regulations. Moreover, the Knee Mapping plugin helps the analyst avoid knee modifier, while Airbag stitching and folding tools set up the pre-crash simulations required for the proper treatment of the airbags. The current paper presents all the afore mentioned tools and more handy features that crash and safety analysts need to set up detailed and accurate models for different regulations fast, and with the minimum human interaction.

CASE STUDY – Material models for depiction of unloading in low speed crash applications

Computer simulation is used in many fields of automobile development to shorten the development term and reduce tests. Large plastic parts such as instrument panels and bumper fascia take a lot of time to make a die, so it is necessary to determine the part shape quickly to shorten the term. Therefore, performance verification by simulation is important. However, due to issues such as material property and complicated part shapes, reproduction of phenomena such as deformation and failure are sometimes not sufficiently accurate in simulations of plastic parts. Creating a die, testing part, and evaluating its performance in such a situation may raise the possibility of inadequate performance. In that case, it is necessary to modify the die, which requires extra cost and time. Therefore, high simulation accuracy is necessary to avoid such risks.

Assembly of full-vehicle digital crash models using ANSA techniques

The level of complexity of the full-vehicle digital crash models varies between car manufacturers. Undoubtedly, however, it is very high and the trend is to become higher and higher, due to the fact that the development teams want their models to include even more details. During all these years a standard technique, that has become of common use, is the fragmentation of the full-vehicle model into *INCLUDE files. Each one contains a distinct physical subassembly of the vehicle, which we shall subsequently name module. This technique offers various advantages, but definitely sets a challenging task. That of connecting the modules to each other and assembling them into a full-vehicle model. Using the embedded capabilities of ANSA we have created processes and tools that enable : i) fast and robust intermodular connections represented by various element types, ii) efficient and reliable integrity check of these connections, iii) the assembly of multi-variant modules, iv) inspection of the connection areas in a lightweight view, v) an error-free update of modules’ versions, vi) the overview of the modules participating in the full-vehicle models with aid of graphs, vii) the reusability of already existing models, as the basis for the creation of new ones.

Modelling of Fracture Initiation and Post-Fracture Behaviour of Head Impact on Car Windshields

More than half of all road fatalities involve vulnerable road users, such as pedestrians and cyclists [1]. When involved in crashes with cars, the head is particularly susceptible to injuries, and especially if the road user hits the windshield of the car [2]. Impact tests are often performed to estimate the risk of head injury during such an event. A pedestrian head strike test normally involves a spherical headform, in which the likelihood of head injury is described by the head injury criterion (HIC) [3].

Analyzing Bicycle Accidents with Human Body Models

For the past 20 years, single-bicycle accidents have been the most common cycling accidents in Sweden (more than 70% of all injuries) [1-4] and in other countries where many people use bikes as means of transportation [5-6]. Vehicles were involved in a majority of the lethal bicycle accidents. Neck injuries were a small portion of all cycling injuries, but they were associate with a large risk of permanent medical impairment. Therefore, it is interesting to explore if head protective safety devices can provide safety benefits for the neck as well. Hövding is a head protective device that is worn as a scarf around the neck, with sensors that trigger inflation of an airbag in case of an accident. Theoretically, the portion of the airbag that surrounds the neck could protect from neck injuries (see Fig. 1).

Characterisation of an Energy Absorbing Foam for Motorcycle Rider Protection in LS-DYNA

A study on traffic fatalities [1] shows that the most frequent causes of motorcyclists’ deaths in accidents are multiple skeletal and visceral injuries. Brain and skull injuries are the second leading cause. An injury frequency and pattern data analysis of discharged riders of powered two-wheelers from hospitals [2] show that the most common injuries of non-fatal accidents are injuries of the lower extremities followed by upper extremity and traumatic brain injuries. With this data of fatal and non-fatal accidents it can be said that besides head protection, the protection of a two-wheeler rider’s whole body with lower and upper extremities is of major importance.

Predicting the results of the finite element simulation of a snowboarding backward fall with ODYSSEE

Skiing and snowboarding are very popular sports, associated with a high risk of injury (2.35/1000 skiing days in 2019 in France [1]). Among those injuries, the leading cause of death is the head injury which accounts for 5-10% of all injuries. Several studies have shown that helmet was effective in reducing the risk of head injury [2], [3] but the effect of the helmet in reducing certain types of brain injuries such as concussion is still unclear [4], [5]. To better evaluate and design effective helmets, it is critical to understand the head impact condition during the crash as well as the injury mechanism.

Comparing the Frontal Impact Responses of the VIVA+ Average Female and SAFER Average Male Human Body Models in a Generic Seat

The VIVA+ 50F average female Human Body Model (HBM), currently in early beta status, was compared to the SAFER average male HBM Version 9 with the aim of investigating differences between females and males in terms of kinematics and injury assessment in frontal impacts. The VIVA+ HBM is under development within the research project VIRTUAL and will be released as open source during the summer 2022.

Improvement in predictive capability of smalloverlap crash simulation with emphasis on GISSMO material model, weld rupture and detailed modeling

CAE tools are one of the best techniques in the auto industry to drive design and help product development with minimal physical tests. Physical tests are very time consuming and expensive which is driving the Auto industry towards virtual simulations to replace physical tests. CAE has become an integral part of product development to accurately predict physical testing and drive design direction. For CAE to accurately predict the physical test, it depends on details captured in the full vehicle model. In the small overlap load case it’s necessary to capture as much detail as possible for components engaged during the impact event. However, capturing too much detail leads to prohibitively large models with excessive computational time. So it is important to understand the load path to decide the critical vehicle components which play a vital role in the crash event. This includes the sheet steel/aluminum stamped parts, aluminum extrusion and also the fasteners and welds. In this paper an attempt is made to revisit the modeling of these critical vehicle components and later confirm the performance with respect to the physical test. The sheet steel/aluminum stamped parts and also the aluminum extrusions are finely meshed and GISSMO material models are implemented to define their rupture. The fasteners (bolts) are modeled using solid elements. Spot welds are modeled as solid nuggets with damage material model MAT_SPOTWELD_DIAMLERCHRYSLER and a simple elegant technique is used to define the aluminum MIG welds. The MIG welds are joining thick Aluminum parts in the cradle. MIG welds are represented by discrete beams with MAT119 material model. The stiffness, loads and rupture displacement parameters are adjusted to component tests and an envelope of rupture is created. This is carried on to the full vehicle as a predictive model and the designs are iterated. All of the above modeling methods and techniques helped to accurately predict velocities, intrusion, wheel kinematics and a good correlation to the physical test was achieved.

Reconstruction of Trimmed and Faceted Vehicle Models for Isogeometric Analysis in LS-DYNA

The traditional engineering design-through-analysis process is inadequate for modern needs. In it, an engineering designer will create a model in a computer-aided design (CAD) software, after which an analyst takes this smooth CAD model, defeatures it, cleans up the model, and ultimately approximates the intended shape as a faceted, semi-structured mesh for subsequent engineering analysis. Analysis, thus, no longer operates on the intended object, but instead evaluates physics on a faceted approximation, which may lead to compromised results [1]. Furthermore, while design and analysis are the primary objects of interest in the design-through-analysis process, the intermediary steps of geometry cleanup and meshing consume over 70% of the time spent in the design-through-analysis process [2, 3]. Naturally, this leads to increased associated costs [4].

Using JFOLD and LS-DYNA to Study the Effects of Passenger Airbag Folding on Occupant Injury

JFOLD is a software tool for simulation-based airbag folding in LS-DYNA®. This paper presents how JFOLD and LS-DYNA can be used effectively to research how slight changes in automotive passenger airbag folding can lead to significant changes in occupant injury prediction. The demands placed on today’s occupant safety teams continue to increase, driving up the need for airbag complexity and simulation accuracy whist driving down the time to deliver. Accurate airbag simulation is critical to improve occupant safety in an increasing number of crash scenarios and out-of-position cases, including passengers of autonomous vehicles. In addition, airbag simulation is now being used to assess the performance of interior trim components during early break-out and deployment.

Numerical and Experimental Correlation of a Survival Cell Designed for a Bus Body Structure

The behavior of mechanical structures, when subjected to impact load, is a matter of great relevance and its applications in terms of vehicle collision. When we analyze the superstructure of a bus, those vehicles must be tested according to prerequisites established in standards such as UNECE ECE 29 (European standard) or CONTRAN 629/2016 (Brazilian standard). The standards prescribe to use a pendular system to evaluate the frontal structure of the vehicle. In this regulation is defined the height and the mass that will collide with the structural modulus. However, the procedures described in these standards do not represent the real collisions involving these types of vehicles. This can be seen when comparing the energy imposed on the test module, detailed in CONTRAN 629/2016, where the energy imposed on the vehicle is approximately 20 kJ on each side of the test module, this corresponds to a collision of a 5 tons vehicle at 10 km/h or a 20 tons bus at 5 km/h.

Truck Frame Optimization Considering Crashworthiness, NVH and Static Responses

This paper demonstrates how to efficiently perform optimization for vehicle structures, taking into account nonlinear responses from LS-DYNA crash simulation, as well as responses from linear loading conditions such as NVH and Static. The optimization process is based on the Equivalent Static Load (ESL) method and uses an iterative process which utilizes the non-linear structural analysis results from LS-DYNA and the linear structural analysis and optimization capabilities of GENESIS. With this integration, the combined multidiscipline problem can be solved with only a few LS-DYNA simulations (5 to 10). In addition, large-scale optimization techniques, such as topology, topometry, topography and freeform, can easily be employed. The optimization process and results will be demonstrated using two examples: topology optimization of a beam cross-section under impact and static loading and topometry design of a truck frame under crash, normal modes, and static loading conditions simultaneously.

An enhanced Design exploration using Modal decomposition of Key events in Frontal crash simulation

In recent years, CAE has been used extensively in vehicle development, and parameter study of sheet metal thickness for design exploration and optimization is one of the major applications. Response surface method is commonly used for this application among various analysis tools. The concept is to connect input variables such as sheet metal thickness and output variables such as firewall intrusion with non-linear functions such as radial base, kriging, and neural network.