S. Mukherjee

Prediction of crushing behaviour of honeycomb structures

A finite element methodology has been developed for predicting the behaviour of honeycomb structures. Dynamic analysis of hexagonal aluminium honeycomb structures is carried out using PAM-CRASH TM , an explicit FE analysis code, and the result are verified against experimental data. Relationship between the crushing behaviour of honeycomb and simulation parameters has been established. The simulation results are also compared with theoretically predicted values.

Six-degree-of-freedom three-wheeled-vehicle model validation

Abstract A spatial six-degree-of-freedom mathematical model of a three-wheeled vehicle used in Asian countries is developed. The governing equations are derived and integrated explicitly without linearization to yield a large displacement model that can be used for parametric studies. The model includes suspension model, tyre compliance, lateral force due to cornering stiffness, and rolling resistance at the tyre. The model is validated against vertical accelerations measured when passing over a road bump. The lateral steady state response is validated through a steady state circular test procedure.

FE simulations of motorcycle—car frontal crashes, validations and observations

Abstract ISO 13232 [1] requires 7 configurations for full scale tests (FSTs) between the motorcycle (MC) and the car (also called the opposing vehicle, OV). Of these, 3 involve the impact of the MC with the front of the OV. This paper presents Finite Element (FE) based simulations of these impact configurations. It does not evaluate the efficacy of any safety devices but refers to parts of ISO 13232 for developing these simulations. This paper analyses FE based simulations of the above-mentioned impacts. The simulations have been developed using the PAM-CRASHTM solver. In this paper an attempt has been made to compare the kinematics of the simulations with those obtained from the FSTs. The simulations indicate that the MC-OV impacts are sensitive to many phenomena. The objective of this paper is to highlight some of the important aspects of MC-OV simulations. This paper has been revised from an earlier version of the paper [2].

Mesh Generation for Folded Airbags

AbstractThis work presents the results of a novel approach to obtaining a geometric mesh of a given airbag after a series of folds are defined on it, for use in crash simulations. The process of airbag folding is simulated as a series of geometric transformations applied on to the airbag mesh, which is modeled as a stack of connected planar layers. Along with these transformations, optimization techniques are used to ensure that there is minimum change in the geometry and the area of the airbag. The results from this approach are compared with several other commercial airbag folding software, and it is observed that the algorithm proves to be very effective in ensuring minimum change in the area and geometry of the airbag during the folding process.

Predicting throw distance variations in bicycle crashes

This paper investigates the correlation between throwing distance and impact speed, point of impact, and angle of approach for varying bicycle–car crash configurations. Crashes between a bicycle and a car were simulated using multibody models developed in MADYMOTM. The Hybrid III 50th percentile male dummy model, available from the MADYMO library, was used. Crash configurations reported in the literature were used and parametric variations were done in speed, angle of approach and point of impact. It is observed that only some configurations show a monotonic dependence of the throwing distance on the car speed.

An Algorithm for Optimised Generation of a Finite Element Mesh for Folded Airbags

Abstract This work presents a novel algorithm for generating a geometric mesh of an airbag after a series of folds are defined on it, for use in crash simulations. The process of airbag folding is simulated as a series of geometric transformations to the airbag mesh, which is modeled as a stack of connected planar layers. Along with these transformations, optimisation techniques are used to minimise change in the geometry and the area of the airbag post-inflation. The results from this approach are compared with commercially available airbag folding software. It is observed that the algorithm is effective in limiting change in the area and geometry of the airbag in the folding process. The change in surface area is only 0.1% of the surface area of the airbag against the change of 11% in the commercial package. The current paper is a sequel to authors ‘Development of FE meshes for folded airbags’, (International Journal of Crashworthiness, 2005 10(3) 259–266), which reports initial results of the folding process.

Issues in ALE Simulation of Airbags

Abstract Airbag simulations have emerged as an important means to study their efficiency in occupant safety. The situation in which the occupant deviates from the nominal position is commonly referred to as an out-of-position (OOP) situation. The focus in this paper is on developing simulation models that can be used for OOP studies in the future. To capture the initial stages of inflation of the airbag in OOP simulations, the focus has now shifted towards discretizing the fluid flow inside the airbag using the Arbitrary Lagrangian Eulerian (ALE) approach. The interaction between the fluid (Eulerian mesh) and the structural elements (Lagrangian mesh) is defined through a coupling. In this work, a description of the process for generating the mesh for the folded airbag is presented. Some strategies adopted to resolve penetration and time step issues in folding have been highlighted. The folded airbag is inflated in simulation using both the CV (Control Volume) and ALE approach, and the results are compared. For validation, the inflation of the airbag was recorded using high-speed photography and a tank test conducted to characterise the inflator. The performance of the ALE method and CV method for the purpose of OOP simulations is compared with the experimental observations.

Development of FE meshes for folded airbags

Abstract A novel algorithm for generating Finite Element meshes of folded airbags for use in crash simulations is proposed and evaluated. The algorithm models the airbag folding process through geometric transformations on a given airbag mesh, followed by a series of numerical optimizations in order to achieve a folded state. An airbag model as a stack of connected planar layers is the input. It is shown that the proposed approach compares favorably with existing airbag folding techniques when evaluated quantitatively with the area and volume of the airbag post folding.

Rollover crashworthiness of a rural transport vehicle using MADYMO

Abstract In this work a full vehicle model of a rural transport vehicle (RTV) is developed in MADYMO. The steering, tyre and suspension are modeled using standard modules available in the package. A speed controller was also incorporated to maintain a constant speed of the vehicle in the simulations. Validation of the model was done against experimental accelerations measured over a bump. The validated model was then used to predict rollover characteristics using Slowly Increasing Steer, J-Turn, and Road Edge Recovery maneuvers. These maneuvers were conducted for three different loading conditions – RTV without passengers, RTV with unrestrained passengers, and RTV with restrained passengers – and the safe speeds are compared with those of other vehicles.

Inverse finite element characterization of soft tissues using genetic algorithm

Karthikeyan B, Chawla A, Mukherjee S Department of Mechanical Engineering, Indian Institute of Technology, New Delhi, India Human body finite element (FE) models for use in impact simulations require soft tissue characterization at high strain rates. The objective of the current work is to extract viscoelastic properties of passive muscle tissues at high strain rates and study their rate dependency. A procedure to identify the dynamic properties of passive muscle tissue under impact has been proposed using isolated-tissue experiments, FE simulations and Genetic Algorithm (GA) based optimization. Data from nineteen impact tests on unconfined isolated human muscles for strain rate ranging from 132/s to 262/s were used [1]. Tissues were compressed up to approximately 50 % strain and the force-time response was recorded. FE simulations of these impact tests have been performed in the present study by modeling the muscle as linear viscoelastic. RMS of the deviation between the experimental and FE force data, sampled at 10 kHz, was then minimized to predict the material parameters, bulk modulus, short-term shear modulus and long-term shear modulus. This parameter identification process was automated using PAM-CRASH