Eric Markiewicz

Riveted joint modeling for numerical analysis of airframe crashworthiness

Abstract Numerous rivets have to be taken into account to simulate the behavior of aeronautical frames under crash loading conditions. Until now this type of bonding modeling has not been judged satisfactory. The influence of structural embrittlement due to the riveting process, the strength of a riveted joint under dynamic loading and the characterization of a simplified rivet element, in particular, were sources of questions. For each topic, we did solid finite element modeling and carried out experiments to measure the influence of strain rates under quasi-static and dynamic loading conditions. We also determined the elastic–plastic and damage mechanical properties for sheet metal plates and rivet materials. Our results showed that it was essential to develop a new kind of rivet element taking material non-linearities into account. Experiments and solid finite element modeling of an adapted Arcan test procedure were then conducted and pure shear and tensile non-linear responses as well as parameters of a macroscopic criterion were identified. The use of this new rivet element was found to improve prediction of the dynamic behavior for a frame assembled with 700 rivets.

An alternative numerical approach for full scale characterisation for riveted joint design

This paper describes a pure numerical methodology (FE database) to improve the representativeness of joint equivalent models for airframe crashworthiness. This method is based on material constitutive models and failure criterion in accurate 3D FE simulations. The interest of FE simulations is to define the dynamic strength of many types of riveted joints with a reduced cost compared to a pure experimental way. The FE database method is carried out on elementary riveted joints to predict and to analyse: first the post-riveting initial strain and stress state, and second several kinds of joint failure (e.g., crack propagation, rivet shearing or pull-out). The aim of the first step is to start the mechanical strength simulations with a correct deformed shape and post-riveting state. The responses of 3D riveted joint simulations can then be considered as reference and be used to optimise the mechanical properties of equivalent joint elements. A new equivalent joint element is developed to improve the representativeness of an airframe crash simulation.

Experimental study of the bone behaviour of the human skull bone for the development of a physical head model

Abstract The first part of this work aims to develop an experimental protocol in order to identify the monolayer behaviour of the skull (cortical and spongy bones in the model are represented by a single homogeneous layer). Experimental tests on 20 skulls from unembalmed cadavers have been performed by the LAMIH and the CEESAR. Nineteen specimens were taken from each of the human skulls following a very precise cartography (five frontal, eight parietal, two temporal, one occipital, two on the coronal suture and one on the sagittal suture). Three hundred and eighty specimens were tested in three-point bending test. In the second part, a physical head model was developed. A reference model was realized by using stereolithography method. The main point was to find a resin whose characteristics are similar to human skull bone. Then, this prototype was validated by means of experimental tests similar to Nahum et al. (test in 6 m/s by means of a cylindrical impactor of 6-kg mass).

Identification technique of constitutive model parameters for crashworthiness modelling

Abstract This paper deals with a parametric identification technique, based on the conjugate gradient methods. As part of this research, compression tests are performed on a XC48 steel (AFNOR norm 1048AISI), which is strain rate sensitive. Two kinds of specimens, tubular and cylindrical “dumbbell”, are tested at different strain rates, from 10 −3 to 130 s −1 , and for impact velocities ranging from 5 mm/min to 6 m/s. The parameters of the Johnson-Cook constitutive model are evaluated directly from the XC48 steel compression stress-strain diagrams expressed in terms of true stress and strain measurements. These parameters are next used as input data for a FE crash simulation of the tubular specimen. Comparison with experimental results prove the validity of the identified parameters.

Material behaviour law identification for the various zones of the spot-weld under quasi-static loadings

The aim of this paper is to predict by FE simulation the non-linear behaviour and rupture of spot-welded assemblies subjected to quasi-static loadings and to show the feasibility of numerical works. The three material zones of the spot-weld (metal base, melted zone and heat affected zone) are analysed and studied on a metal sheet assembly structured 0.7/0.7mm Experimental works based on tensile-shear, cross-tensile and peeling tests are done, as well as a numerical sensitivity analysis on the three material zones. Different fracture modes are put into evidence. A methodology to identify the material behaviour law of the MB, HAZ and MZ is proposed. The rupture of the spot-welded structure is simulated with the help of the Gurson damage model. The damage parameters of the MB and HAZ are identified through an inverse method on the basis of tensile-shear and cross-tensile tests. The numerical/experimental comparison on the various applied quasi-static loadings validates this methodology.

Full scale experimental characterisation for riveted joint design

This paper deals with experimental works the objective of which aims at improving the design of riveted joints for airframe crashworthiness purposes. Complex assemblies are considered at this stage as the sum of simpler ones constituted of 1 rivet and 2 plates, the behaviour of which is investigated under different points of view. Research is divided in two major parts. The first one investigates the consequence of rivet processing in term of possible material and structural embrittlement (residual stress and strain in rivets and plates). In the second part, the overall behaviour and strength of the considered basic assembly (1 rivet and 2 plates) is studied. The aim is here to characterise basic failure modes of assemblies linked to rivet failure or crack propagation in punched metal plates. The measurements of local variables enable to assess the influence or not of dynamics on the different failure mechanisms. Eventually, an original test procedure based on the ARCAN test rig is presented, the objective of which is to give access to multi-axial failure criterions for rivets.

Numerical approach for assessment of dynamic strength for riveted joints

Abstract Numerous rivets have to be modelled for aeronautical framework crashes. A numerical procedure based on FE modelling and characterisation of material failure constitutive models is proposed in order to limit the experimental procedure. Quasi-static and dynamic experiments are carried out on elementary tension (punched) and shear (riveted) specimens. No strain rate sensitivity has been measured on the riveted joint assemblies failure. The experiments are used to identify, by an inverse method, the Gurson damage parameters of each material (2024-T351 and 7050 aluminium alloys for the sheet metal plate and the rivet). The characterisation gives rise to a satisfactory correlation between FE models and experiments. Optimised parameters are validated for each material by means of a uniaxial tension test for the sheet metal plate and an ARCAN type specimen in pure tension for the rivet. Results can then be used to identify macroscopic failure criterion to model the rivet behaviour in aeronautical framework crashes. FE tools can also resolve problems linked to limit-design or the design of new riveted joint assemblies more rapidly and cost effectively than experiments.

Identification of the spongy bone mechanical behavior under compression loads: numerical simulation versus experimental results

Abstract In the fields of crashworthiness, ballistic protections, and other medical applications, the accurate material constitutive law of spongy bone is needed to carry out valid finite element analyses. The direct identification of bone mechanical behavior law is not easy since it is a complex network of intersecting osseous spans (trabeculae), where the space in and around the trabeculae contains bone marrow and fluids. We propose in this work to overtake the bone geometrical dispersion by applying an inverse scheme identification method, based on the global mechanical response, correlated with the exact geometry. First step study was made on spongy bone cylindrical samples cut in beef ribs. Compression tests on these samples showed a large dispersion and suggested that the fluid effect can be neglected during the quasi-linear part of the mechanical response. The micro-architecture of each sample was acquired thanks to microcomputed tomography technique (μCT). After applying a threshold, we used the μCT data to build a micro-FE model of the spongy bone. This model is introduced in FE code in order to simulate quasi-static compression of the sample. An elastic plastic constitutive law is assigned to the spongy bone. An optimization procedure is then applied in order to identify the spongy bones behavior. The optimization function is based on the global response (force versus displacement) of the sample. This procedure was repeated for different samples in order to obtain average spongy bone behavior.

On the effect of marrow in the mechanical behavior and crush response of trabecular bone

The present paper focuses on the mechanical behavior analysis of bones at mesoscopic scale, paying a special attention to the trabecular bone and the bone marrow filling the porosities. Uni-axial quasi-static compression tests under unconfined conditions have been performed to identify the mechanical behavior of 46 trabecular bone samples. The bone marrow for 22 samples has been preserved to analyze the fluid flow effects on the crushing response. Although deformation patterns do not differ significantly, the average crush behavior of the trabecular bone shows an unexpected decrease of the mechanical properties when the marrow is kept in the sample (26% for the elastic modulus (E(a)), 38% for the maximum compressive stress (σ(max)) and 33% for the average stress (σ(mean))). An explanation is given by analyzing the contribution of the bone marrow viscosity which smooths the mechanical response. A numerical analysis on an idealized trabecula confirms that the marrow induces transverse pressure and extra local stress on trabeculae during its flow, causing the premature collapse of the trabecular network.

Towards a finite element head model used as a head injury predictive tool

The aim of this paper concerns the development of a new finite element head model. A particular attention is focused on the representation of the skull/brain interface and the use of hyperelastic material is recommended to represent the cerebro-spinal fluid. The validation of this model is achieved by the comparison of its response with results obtained from cadaver impact tests. A second topic of this paper deals with an original scheme which is to find a correlation between numerical model response and head injuries. An accident reconstruction methodology has been developed. Based on numerical and experimental approaches, dynamic head solicitations generated during the accident are identified. By using these solicitations as input in the head model, interrelationships can be established between finite element head model response and injuries observed on in vivo human victims. Applied here to one accident case, a link can be built between inner brain contusions and “von Mises impulsive stresses”.