‘To develop the technologies and knowledge for building capabilities needed to ensure the security of citizens from threats such as terrorism, natural disasters and crime, while respecting fundamental human rights and privacy’ is one of the objectives of the seventh framework programme of European Union policy on research and technological development. In this context, the research conducted in the Vulnérabilité des Constructions soumises A des Impacts et des explosioNs project (VULCAIN) funded by the Agence Nationale de la Recherche (ANR – Programme Génie Civil et Urbain – 2007) is developed to improve the evaluation of the structural safety of sensitive industrial structures (storage and protection) that might be submitted to terrorist or accidental threats. The project is focused on severe dynamic loadings such as impacts or explosions.
The quantification of the vulnerability of a protection or storage is still a large field of scientific investigation despite numerous works. Indeed, the design methods are often simplified: dynamic actions are replaced by equivalent static forces and the effect (of the impact or the explosion) is assumed to be localised. The design of the steel tank or a concrete wall under impact is then based on empirical formulae.
This special issue of the European Journal of Environmental and Civil Engineering gathers the main contributions of scientists involved in the VULCAIN project. Two main kinds of structures were studied experimentally and numerically. Firstly, the vulnerability analysis of steel tanks to accidental impacts or explosions is presented. Secondly, the analysis of concrete protection structures submitted to hard and soft impacts is carried out.
Much of the research effort in this project is: (i) to conduct relevant small-scale experiments to characterise both dynamic loads and structural responses (storage tank under blast wave and reinforced concrete wall under soft impact) and (ii) to provide simplified analytical and more refined numerical tools able to predict and quantify the non-linear structural or material response of real structures submitted to extreme loads.
Based on experimental data available in the literature, some representative steel tanks have been selected and the different modes of failure are identified; various methods are considered for the modelling of damage. Simplified models about buckling and perforation/penetration of projectiles impacting cylindrical vessels are proposed according to the experimental results. The sophisticated numerical models will help to understand and reproduce the behaviour of large concrete structures under impacts. They will also allow evaluating the validity of simplified formulas estimating penetration or perforation of concrete structures.
In this context, Noret, Prod’homme, Yalamas, Reimeringer, Hanus and Duong present a representative selection of accidental situations observed in oil and chemical storage sites. First, static and dynamic mechanical models are proposed to represent tank failure modes identified for impact and blast loadings. Then, some stochastic models of loading and structural modelling parameters are developed based on accident observations and expertise. Uncertainty propagation methods are used in order to evaluate the sensibilities and probabilities of failure associated to the selected models and geometries. This work reveals the prevalence of buckling failure mode and the importance of the dynamic feature to perform accurate accidental risk assessment.
Duong, Hanus, Bouazaoui, Pennetier, Moriceau, Prod’homme and Reimeringer (Parts I and II) present test results, performed on an explosion table, of a tank physical model submitted to a blast wave. The selected type of structure corresponds to the low pressure cylindrical steel storage tanks. The blast wave is generated by the detonation of a hemispherical charge of gas (propane-oxygen in stoichiometric proportions) confined in a bubble. The radius of confinement varies between 4 and 10 cm and allows incident positive pressures ranging between 70 mbar and 3 bar (associated positive phase duration between 4 and 22 ms) approximately, which is usual in hazard studies. In a first part, experiments performed on instrumented rigid models are used to quantify the resulting pressure loads in terms of time and space distribution. A computational fluid dynamics simulation (LS-DYNA) is used to show that fluid–structure interaction contribution remains negligible in the experiments. In a second part, with the use of experiments on flexible models, buckling failures modes are investigated. These test results are then used to assess the validity of simplified analytical models as reliable design tools and to verify the conformity of numerical simulations against experimental results. Finally, an oil storage tank of 60,000 m3 is analyzed and a critical curve for buckling in a pressure impulse diagram is constructed; deterministic and reliability analysis results are compared with the classical French practices for domino effect assessments.
The first three articles deal with industrial metallic tanks. The three following articles deal with RC targets. Nowadays, comparisons of empirical formula for predicting perforation or penetration of RC structures under impact rather concern the ballistic domain with hard impacts and impact velocities greater than 100 m/s. Therefore, VULCAIN proposes an experimental campaign of soft impacts with velocity ranges less than 100 m/s. Pontiroli, Rouquand, Daudeville and Baroth present this campaign of soft impacts on RC walls generated by steel tubes. Numerical simulations of these tests show the ability of the proposed finite element model to reproduce the different local mechanical effects (cracking, spalling, scabbing, and perforation) occurring during the impact. These tests have been carried out in order to get more data on aircraft-type impacts for the validation of numerical models.
The article proposed by Baroth, Daudeville and Malecot deals with empirical formulae for the perforation prediction of RC barriers. These formulae are usually validated for hard impacts only. Consequently, on the one hand, a simple method is proposed in the case of soft impacts. On the basis of experimental tests, including those presented by Pontiroli et al., the comparison between experimental results and numerical predictions of perforation shows a good agreement. On the other hand, recent triaxial compression tests performed on ordinary concrete under high confinement have shown that the unconfined compressive strength of concrete after 28 days, fc28, is not a relevant indicator of the mechanical response of concrete. Both the free water present in the cement paste and in the entrapped air voids and the granular skeleton exert a major influence. This new result is very important since most of design analytical formulae consider fc28 as the main material parameter. Therefore analytical models based on fc28 need to be considered with caution.
The efficiency of the discrete element method (DEM) for studying the fracture of heterogeneous media has been demonstrated, but it is limited by the size of the computational model. Because both the local analysis of damage in the vicinity of the impact and the global response of the RC building are necessary, a multidomain approach is used by Potapov, Faucher, and Daudeville for the simulation of soft impacts on RC walls. This multidomain approach is based on a DEM/FEM coupling. Potapov et al. propose a first attempt to reduce the calculation costs based on a multidomain MPI parallel version of the finite element code EUROPLEXUS. In the last article, Durand, Marin, Faure and Raffin propose to take advantage of graphics processor units (GPUs) to reduce the computation time. GPUs are massively parallel coprocessors increasingly used to accelerate numerical simulations. The algorithm and implementation of the DEM on GPU are detailed. Then the article presents performance results for simulations of a rock impact on a concrete slab.
All the papers presented in this special issue were peer-reviewed prior to publication. The editors want to express their gratitude and thanks to the three reviewersFootnote 1 for their conscientious reviews and useful recommendations.
Notes
1. Maurice Lemaire, professor at Institut Français de Mécanique Avancée, Clermont-Ferrand, and consultant at Phimeca Engineering SA; Alain Millard, senior researcher (CEA) and professor at the École Polytechnique, Palaiseau; and André Lannoy, formely senior researcher at EDF, vice-president of the Institut de Maîtrise des Risques.