NUMERICAL MODELLING OF BLAST EVENTS TO DESIGN AN INNOVATIVE BLAST RESISTANT TEXTILE LUGGAGE CONTAINER

ROSARIO DOTOLI Consorzio Cetma, Materials and Structures Engineering Department, S.S. 7 km 706+030 72100 Brindisi Italy, rosario.dotoli@cetma.it http://www.cetma.it

DANILO BARDARO Consorzio Cetma, Materials and Structures Engineering Department, S.S. 7 km 706+030 72100 Brindisi Italy, danilo.bardar@cetma.it http://www.cetma.it Abstract

Keywords blast resistant; simulation; textile; luggage; container; blastworthy

1. Introduction This paper deals with the activities of numerical simulations and validation for the development of a textile container for the blast protection of cargo holds. The activities are carried out within the FLY-BAG Research Project Blastworthy textile-based luggage for aviation safety, co-funded by the European Commission, 7 FP, under the Transport Aerostructures research area. About 75% of the aircrafts in service are narrow-body aircrafts, and more than 70% of bombing attempts have been against these aircrafts. The risk that a small quantity of an explosive, below the threshold of the detection instruments, could get undetected cannot be discarded, and the introduction of countermeasures to reduce the effects of on-board explosions should be considered, especially for narrow-body aircrafts. Existing Hardened Unit Load Devices (HULD) have been developed to reduce the effects of on-board explosions, but they have some disadvantages which prevent their wider utilization: they are heavier and much more expensive than standard luggage containers and, notably, applicable only to wide-body aircrafts. The research and development of hardened containers for narrow-body aircraft are lagging behind the work on containers for wide-body aircraft. Limited research has been done on container role as part of a total architecture for aviation security. Explosion-containment strategies and development of a new concept of ULD (Unit Load Device) for narrowbody supported by numerical simulations is the aim of the study presented. In particular, the work points out the importance of the numerical modelling approach in order to virtually reproduce the blast tests carried out for the characterization of materials used for the innovative blastworthy luggage container. All experimental tests have been simulated and validated in order to predict materials behaviour and to perform sensitivity analysis. Numerical simulations are carried out to perform sensitivity studies and to investigate the effect of the test parameters variability, such as bomb placement, luggage filling ratio. Full scale blast tests are then performed for the validation of the numerical simulations and for the demonstration of the Fly-bag concept behavouir. All activities have been focussed on the containment of blast waves and on the reduction of the aircraft structure damage.

Rosario Dotoli, Danilo Bardaro

2. Blast modelling approach It is possible to study the effect of a bomb explosion into a textile container in terms of blast wave propagation and structural resistance by means of different numerical methods: traditional Fluid Structure Interaction (FSI), Arbitrary LagrangianEulerian approaches (ALE), Smoothed Particle Hydrodynamics (SPH). The first one needs two different codes: a CFD and a FE solver. At every iteration, the pressure distribution on the solid, achieved using Computational Fluid Dynamics software, is input in the FE numerical code; whereas the structure deformations, computed applying a Finite Element code, are transferred back to the CFD code. This data exchange occurs by means of a communicator, a third external software, which interpolates the results from two calculation domains. This approach is very complex and requires high performance computing resources. The Arbitrary LagrangianEulerian approach is an alternative way to solve problems involving fluid-structure interaction with high stressstrain rates as in the case of the Fly-bag project for the presence of blast loading condition. This method has been developed on the idea of separating the material and mesh displacements to eliminate mesh distortion, that is the main drawbacks of the Lagrangian approach. On the contrary, extra equations provide relations between the material points and the mesh points. The ALE method does not require any mesh generation algorithm or communicator between different solvers and does not change the topology of the problem, hence can be easily implemented in the existing Finite Element codes, such as LS-Dyna. This code allows the simulation of a blast event by means of a classical ALE approach and by others two alternative methods: Smoothed Particle Hydrodynamics (SPH) and Conventional Weapon (ConWep). In Ls-Dyna code the ALE approach allows a mesh movement independent from the material flow. Moreover each mesh element can contain a mixture of two or more different materials (Slavik, 2007). For explosion problems, an element of the Eulerian domain may contain one or two different entities, air and gas produced from the detonation. The mesh size of the finite element has a significant effect on the accuracy of simulation results, which is based on both the dimensions of the model and the computer performance. ALE approach needs to model the fluid part with an accurate mesh to precisely evaluate the shock propagation around the structure. Nevertheless, this generates big models because in order to better define the blast wave propagation from the explosive to the structures a fine 3D air mesh is needed. Consequently the ALE elements must be small enough to capture the nearly discontinuous shock front. In the case of a long distance between the explosive charge to the target structure (standoff distance), an enormous volume of air must be modelled. The result is a complex and time consuming model with small integration time steps due to the fine air mesh. The ALE method is computationally expensive and is appropriate for small standoff distances. On the other hand, Smoothed Particle Hydrodynamics is a meshless technique unaffected by the normal problems of grid tangling in large deformation problems. This is a Lagrangian approach thus allowing efficient tracking of material deformations and history dependent behaviour. Compared with Euler, it only needs to model regions where materials exist and it doesnt need to create all regions where material might exist as a void (Toussaint and Durocher, 2008). Being meshless, phenomenon such as fracture and fragmentation may be modelled properly. Fracture can occur arbitrarily without a priori artificial mesh modification. With all of its promises, however, SPH technology is relatively not fully developed if compared with standard grid based Lagrangian and Eulerian techniques. Several problems need to be solved before the technique becomes a fully developed computational continuum dynamics technique. There are remaining known problems in the areas of stability, consistency and energy conservation (Le Blanc et al., 2005). The Conventional Weapon (ConWep) method is a pure Lagrangian approach where an analytical loading of the structure replaces the computation of the shock wave propagation. This Lagrangian approach allows the use of a smaller model since only the structure needs to be modelled, without modelling the explosive and the air leading up to the structure. This kind of approach, based on an empirical model, defines an air blast function for the application of the pressure load due to conventional weapons explosives. The inputs needed for this function are restricted to the initial position of the explosive and the TNT equivalent explosive weight. However, this method is limited to the analysis of the explosions characterized by hemispherical charges on the ground or spherical charges in the air without ground interaction (Murat et al., 2008). This method is very fast and appropriate to perform sensitivity analysis and preliminary analyses.

Numerical modelling of blast events to design an innovative blast resistant textile luggage container

3. Simulation activities These activities aim at reproducing all the experimental test carried out starting from current aluminium container to the final design (Gatto and Krznaric, 1996). Preliminary experimental tests have pointed out the limits of the current container in term of blastworthy. Moreover the comparison between the numerical simulations and the experimental tests allows the set-up a validated numerical model that could be used for the sensitivity analysis. The explosive is modelled by means of a large number of solid particles (SPH approach) to avoid local effect on the structure and to correctly reproduce the blast loading. LsDyna Keyword *MAT_HIGH_EXPLOSIVE_BURN allows the modelling of the detonation of a high explosive. The inputs necessary for high explosive material are: mass density, detonation velocity and ChapmanJouget pressure (Williams et al., 2007). An equation of state must be associated with *MAT_HIGH_EXPLOSIVE_BURN, for the evolution of the explosive after ignition. In this case the JonesWilkins-Lee (JWL) equation of state defines the pressure in the elements as a function of the relative volume and internal energy. The comparison between blast test inside the current aluminium unit load device (ULD) and the numerical simulation is shown in Fig. 1 and Fig. 2 at two different times, 0.7ms and 2.2ms. The blast test carried out on a standard aluminium unit load device was modelled by using the LS-DYNA code, with the aim to set the parameters of the numerical model. Furthermore the experimental and numerical results have pointed the structural limits of the current ULD. Starting from these consideration all the activities have been focussed on the development of container able to better resist to the blast waves in order to cause a limited damage to the aircraft structure.

Fig. 1 Comparison of experimental test with numerical simulation, Sph approach simulation time t=0,7[m/s]

Rosario Dotoli, Danilo Bardaro

Fig. 2 Comparison of experimental test with numerical simulation, Sph approach simulation time t=2,2[m/s]

3.1. Belts sensitivity analysis Simulation activities were then directed to support the design of the textile-based container, through the evaluation of forces, displacements and stress during the blast event. In order to design the FLY-BAG device the first step was to numerical evaluate the contribution to the blast mitigation of the belts enveloping the bag. The comparison between different configurations has been made in term of variation of force on beams and bag displacements.

Fig. 3: Fly Bag reference model: external textile-based container and frame represented by 12 cylinders, in order to evaluate the force .

Numerical modelling of blast events to design an innovative blast resistant textile luggage container

The same boundary conditions in all the analyses were considered: the only difference was related to the position and the number of belts. The reference model is reported in Fig. 3, where the frame is schematized by 12 cylinders. The container is constrained in the fixing points placed in the upper side. In this configuration, the trial material used for simulating the textile was an aramid fabric. The position of the charge is on the left side as in experimental test on the aluminium ULD. The distribution of displacements of the bag without belts is reported in Fig. 4. It was calculated a maximum displacement of 175 mm.

Fig. 4 Textile-based container (no belts) Resultant displacement [mm]

In order to reduce the displacements other configurations with 2 belts and 4 belts were considered. All belts are fully constrained in the corners while fixing points are constrained in the upper side (represented by the black points in Fig. 5).

Fig. 5: Textile-based container with 6 fixing points and 4 belts

The results of the numerical simulations carried out on the configuration with 2 belts, is shown in Fig. 6: the explosion inside bag now produces a reduction of displacement, in comparison with the previous configuration,

Rosario Dotoli, Danilo Bardaro

of about 20%. The corresponding peak of force was found on the fixing point n 2, with a reduction of about 55% compared to previous one.

Fig. 6: Textile-based container (2 belts) Resultant displacements

Fig. 7 Textile-based container (4 belts) Resultant displacements

The simulations results of the 4 belts configuration provide a displacement reduction of about 30%. The corresponding peak of force on the fixing point n 2 shows a reduction of about 65% compared to the configuration without belts. In order to summarize the simulations results, a comparison in term of force and number of belts is illustrated in Fig. 8 and in Table I where displacement and forces are shown. All data were normalized respect to the configuration without belts. Fig. 9 summarize the bag displacement obtained by changing the number of enveloping belts. Data refer to the same detonation time t=0.7ms.

Numerical modelling of blast events to design an innovative blast resistant textile luggage container

Fig. 8 Comparison of resultant force increasing number of belts

Table I. Bag displacements, force on fixing point and cylinder n.9

Fig. 9 Bag displacement changing the number of enveloping belts at the same detonation time t=0,7ms

All these simulation activities have been directed to support the fly-bag design choices. In particular the selection of the belts enveloping the bag was based on the max stress calculated along the X direction at 1 ms (Fig. 10).

Fig. 10 Distribution of X stress on belts

Rosario Dotoli, Danilo Bardaro

3.2. Analysis of the multilayer concept The activities were then focussed on the simulation of the multilayered textile bag concept. The aim was to support the design choices dealing with the multilayer configuration and the material selection. A bang box experimental procedure has been set-up at Blastech Laboratory (a Fly-bag project partner); the tests were carried out with different charge levels, material types and layers configuration. The reference model and the testing facility is reported in Fig. 11.

Fig. 11 Set up of the bang box experimental test on multilayer

In order to correctly evaluate the behaviour and the strength of the fabric, the numerical model has been calibrated on the basis of this experimental test performed. The comparison between the experimental test and the numerical simulation is reported in Fig. 12. The first analysis deals with the detonation inside the bang box composed with only one layer of fabric.

Fig. 12 Comparison of Bang box experimental test and numerical simulation

On the basis of the calibrated numerical model a sensitivity analysis has been carried out with different layer thicknesses and number. The evolution of the detonation and its effect on a two layer configuration is summarized in Fig. 13.

Numerical modelling of blast events to design an innovative blast resistant textile luggage container

Fig. 13 Bang box experimental test simulation results: multilayer textile configuration

Different tests and simulations were carried out to identify the relative performance of the fabric materials to quasi static pressure and to identify the relative performance of multilayer configuration to the shock loading. Finally it has been identified a final multilayer configuration made of 4 layers of fabric material, (InnegraTM and TwaronTM) to be adopted for the final prototype. Analyzing the preliminary results of a blast simulation of the final configuration it can be assessed that, even if the four layers work independently, the whole fly-bag concept is suitable to reduce the total displacements of the container structure. Further analyses will be performed on the multilayer textile container in order to support the final prototyping phase. The aim is to verify and find the most advantageous blast resistant design of: sandwich panels configuration, multilayer textile material and manufacturing scheme.

3.3. Full scale final model The final simulation activities were carried out for the evaluation of the performances of the novel textile-based container during a blast event on board, inside the cargo area of an Airbus A320. For this activity a detailed model of the fuselage and the corresponding cargo area compartment was developed. This allowed to simulate the damage caused to the airframe by an explosive charge of the same magnitude as the one used in the validation test. The fuselage and the cargo area compartment have a length of 3,34 m and a diameter of 3,95 m. The internal structure of the aircraft shows ribs, stringers and the floor that separates the cargo area from passengers one Fig. 14. The 3d full scale model of the fuselage allows the evaluation of the effect on the structure during a detonation and offers the possibility to analyze different loading scenarios. Preliminary simulations have been carried out considering the effect of a blast load inside an empty cargo area, with the same approach and loading conditions used for the unit load device. The foreseen damage caused by the explosive is shown in Fig. 14 where it is possible to notice the rip in the fuselage near the detonation area. The stringers and the skin panels are completely removed while the floor is subjected to local bulges between the beams.

10

Rosario Dotoli, Danilo Bardaro

Fig. 14: Blast wave effect inside the empty cargo area

In order to evaluate the blastworthy contribute of the fly-bag concept, the final simulations are focussed on its interaction with the airframe structure. Fig. 15 shows preliminary results of a simulation of the textile based container inside the cargo area. The numerical model takes into account the initial propagation of the blast wave from the container sloping edge, which is the most critical part. The first simulation results show an encouraging reduction of the cargo compartment damage. The current operative step is still in progress and deals with the simulation of the final multilayer bag configuration inside the cargo area.

Fig. 15: Blast wave propagation inside the textile-based container with the SPH approach

Numerical modelling of blast events to design an innovative blast resistant textile luggage container

11

3.4. Conclusions A combined experimental and numerical approach has been considered for supporting the development of an innovative blast resistant textile luggage container, designed to protect the aircrafts from explosions caused by bombs concealed inside the checked luggage. The most important finding from the experimental work was to identify the need for the container to be able to both withstand intense localised shock loading from the initial detonation of a device, and contain the subsequent quasi-static pressure caused by the release of gas from the chemical decomposition of the explosive. Although experimental testing is always necessary, there are considerable motivations for an extended use and support of numerical simulations, as a tool to address the variability in test parameters, such as bomb placement, effect of the luggage inside the container, structure and explosive charge performance. The Smoothed Particle Hydrodynamics (SPH) method has been used to simulate the blast detonation phenomenon inside the textile container and to predict its structural response. The blast test carried out on a standard aluminium ULD was modelled by using the LSDYNA code, with the aim to set the parameters of the numerical model. Further analyses were performed on a textile container: several configurations were numerically tested for evaluating the influence of the belts around the container and to design the multilayer textile configuration. The last numerical analysis was carried out for predicting the effects of a blast wave propagation inside the cargo area of the forward fuselage of a narrow body aircraft, without luggage container. The analysis was performed considering the same explosive charge considered in the previous simulations. In this case of an explosion inside the empty cargo the damage of the fuselage is relevant: the stringers and the skin panels are completely removed while the floor is subjected to local bulges between the beams. On the other hand the first simulations that take into account the multilayer textile container are very promising and show an encouraging reduction of the cargo compartment damage. Further analyses will be performed on the textile container and different configurations will be numerically tested. The aim is to verify and find the most advantageous blast resistant design of: sandwich panels configuration, multilayer textile material and manufacturing scheme.

Acknowledgments The authors wish to acknowledge the financial support of the European Commission under the Contract No. 213577. The authors would also acknowledge the Fly-bag project partners for the support in developing project activities and in particular the Blastech team (United Kingdom) where the blast tests have been carried out.

References
Gatto, J. A. and Krznaric S.(1996). "Pressure loading on a luggage container due to an internal explosion," Proceedings of the 1996 4th International Conference on Structures Under Shock and Impact, SUSI Udine, Italy 1996, pp. 61-67. G. Le Blanc, M. Adoum, V. Lapoujade (2005). - External blast load on structures Empirical approach 5th European LSDYNA Users Conference. Murat Buyuk, Cing-Dao Steve Kan, Nabih E. Bedewi (2008). Moving Beyond the Finite Elements, a Comparison Between the Finite Element Methods and Meshless Methods for a Ballistic Impact Simulation 8th European LS-DYNA Users Conference. Todd P. Slavik (2007) - A Coupling of Empirical Explosive Blast Loads to ALE Air Domains in LS-DYNA 7th European LS-DYNA Users Conference. G. Toussaint, R. Durocher (2008). - Finite Element Simulation using SPH Particles as Loading on Typical Light Armoured Vehicles 10th International LS-DYNA Users Conference . K.Williams, S. McClennam, R. Duricher (2007). - Validation of a Loading Model for Simulating Blast Mine - Effects on Armoured Vehicles 7th European LS-DYNA Users Conference.

12

Rosario Dotoli, Danilo Bardaro

About the authors (optional)

ROSARIO DOTOLI, Email: rosario.dotoli@cetma.it Consorzio Cetma, Materials and Structures Engineering Department, S.S. 7 km 706+030 72100 Brindisi Italy, Master degree in Mechanical Engineering, simulation engineer His main research interests are the numerical modelling of explicit phenomena: impact analysis, crash analysis, ballistic analysis, blast simulation and fluid-structure interaction. He has 10 years experience in CAE technologies in particular with Ls-Dyna, Ansys, Hyperwoks numerical codes and high performance computing resources. DANILO BARDARO, Email: danilo.bardaro@cetma.it Consorzio Cetma, Materials and Structures Engineering Department, S.S. 7 km 706+030 72100 Brindisi Italy, He is the CETMA Modelling and Simulation Area manager. His main competences deals with the numerical modelling of complex phenomena (computational fluid-dynamics analysis, thermal analysis, implicit and explicit structural analysis) by means of advanced CAE (Computer Aided Engineering) technologies: finite element and CFD codes (Ls-Dyna, ANSYS, Fluent ). Other research interests are the numerical modelling and characterization of the behaviour of traditional and advanced ceramic materials.