This paper presents an artificial neural network (NN) modeling approach for representing mechanical fasteners in large-scale finite element crash simulations for explicit analysis using LS-DYNA version R9.3.1. The NN-model is established to describe the local force-deformation response of point-connectors in automotive applications like self-piercing-rivets and flow-drill-screws. The behaviour from initial loading until failure or unloading is covered. Various architectures and complexities of feedforward NNs were evaluated and trained based on synthetic experiments generated from the constraint model proposed by Hanssen et al. [1]. The constraint model is available as *CONSTRAINED_SPR2 but was used in form of a cohesive element (8-noded, 4-point cohesive element with offsets for use with shells).
LS-DYNA models for industrial applications are often composed from several smaller sub-models. For example, in the automotive industry the different components of a car are usually modelled separately. These “include files” may be built-up in different departments of the same company or even developed by external suppliers. Due to the huge variety of features and functionalities in LS-DYNA, it is a good idea to set up general rules for the sub-models to achieve robust and efficient simulation models for production. Also, the successful assembly of models can be assured by defining generalized modeling guidelines, especially for the interaction of the different include files.
The commissioning of forming tools includes the mechanical spotting of the active surfaces and the identification of suitable actuator set values. It represents a time-consuming and expensive step in the tool development process. A major cause for the necessity of this manual die spotting is the elastic compliance of press and die that results in geometric deviations between predicted and produced part geometry as well as the control characteristics and the resulting accuracy of the drives. The interactions between the machine and the forming process can be computed numerically. An integration of simulation models that includes the machine behavior into the tool development process offers potential for time and cost savings for die manufacturers and machine operators.
From 2017 until mid’ 2020, DYNAmore collaborated with various partners from the manufacturing industry, universities, independent research institutions, and software vendors to develop a software neutral storage format for finite element data. The goal was to define a new standard which allows for the flawless exchange of all the required information, enabling the industrial partners to easily establish closed simulation process chains for production processes, where various software tools with non-consistent data formats have been a barrier in the past. During the project, an interface between the mapping tool Envyo® and the established VMAP standard has been realized and validated with test cases.
With the progress in CAE simulation leading to more complicated and integrated workflows, data control and transfer becomes essential. This is extremely important in the manufacturing industry where complicated simulation workflows are necessary in tracking material changes throughout the manufacturing proclete software interoperability.
Recently, many enhancements have been made to LS-DYNA’s implicit solver with regard to improved robustness and general functionality. New material models, element formulations, and features for convenient load history management all contribute to making the implicit solver more versatile and user friendly. The new features are of course relevant to general analyses, but here a case-study of a combustion engine analysis is presented to illustrate and promote these new features. Focus will be on thermal and mechanical analyses, to evaluate deformation, stress, plastic strain, gasket pressure etc. The built-in fatigue analysis capabilities of LS-DYNA will also be demonstrated since fatigue life is often the design target in structural analyses of combustion engines.
The ecological transition necessity makes the use of cryogenic fluids more and more relevant. However, experimental tests and associated modelling of those liquids dynamic vibratory behaviour remain extremely challenging. Indeed, security, control and conditioning are critical issues due to the intrinsic fluid instabilities. Among those critical fluids, liquid hydrogen and supercritical xenon are both highly used in the spatial propulsion domain. Because of their hazardous behaviour, only few experimental dynamic tests have been performed to improve the knowledge of their behaviour inside a vibrating tank.
A thermomechanical modelling technique was developed using LS-DYNA for simulating the heating and subsequent erosion of metallic elements by a continuous wave laser beam. Accurate representation of the laser-material interaction requires inclusion of several physical phenomena.
It is known that the simulation of fastening elements can be carried out by using different approaches. One common way is the use a force-displacement based approach for single-point-connections. While enabling LS-DYNA to calculate crash simulations in our crash simulation tool chain it was necessary to make several adjustments to the standard *CONSTRAINED_INTERPOLATION_SPOTWELD keyword to ensure that the currently available data can be used almost completely.
Sudden cardiac death commonly occurs due to heart rhythm disorders called arrhythmia. Although recognized as the most efficient treatment options, Cardioverter Defibrillator implantation and tissue ablation are still not used to their full potential. Recently, advances in computational modeling and the increasing use of imaging tools have proven that patients’ digital twins can play a role in addressing these limitations. This paper presents such an approach using the industrial software ADAS-3D and LS-DYNA. The workflow starts from Late Gadolinium Enhanced-Magnetic Resonance Imaging (LGE-MRI) data from a patient with structural heart disease. The left ventricle and fibrotic substrate were analyzed using ADAS-3D software, which enables to distinguish between tissue that is healthy, scarred, and intermediate, and to extract topological information. This segmentation and tissue classification are used to build, using LS-DYNA, a detailed electrophysiology model containing the relevant features for simulating arrythmia. Using LS-DYNA, this model is then used to simulate a normal heartbeat and a clinical pacing protocol for inducing arrhythmia.
The virtual process chain is an essential step for the sustainable digital transformation in the manufacturing industry. For the Body-In-White (BIW) sheet metal parts, the manufacturing joining simulation based on finite element method is used to simulate the joining processes in the body shop. The target is to predict the dimensional accuracy of assemblies after using different types of joining technologies. However, the assembly deviation is not only affected by the joining operations, but also by the initial part deviations. Therefore, an integration of geometrical variations in the joining simulation model is necessary to improve the prediction accuracy.
Neural network representation of mechanical fasteners in large-scale analyses