J.Y. Cognard

Mechanical behavior of syntactic foams for deep sea thermally insulated pipeline

Ultra Deep offshore oil exploitation (down to 3000 meters depth) presents new challenges to offshore engineering and operating companies. Flow assurance and particularly the selection of insulation materials to be applied to pipe lines are of primary importance, and are the focus of much industry interest for deepwater applications. Polymeric and composite materials, particularly syntactic foams, are now widely used for this application, so the understanding of their behavior under extreme conditions is essential. These materials, applied as a thick coating (up to 10-15 cm), are subjected in service to: - high hydrostatic compression (up to 30 MPa) - severe thermal gradients (from 4°C at the outer surface to 150°C at the inner wall), and to high bending and shear stresses during installation. Damageable behavior of syntactic foam under service conditions has been observed previously [1] and may strongly affect the long term reliability of the system (loss of thermal properties).This study is a part of a larger project aiming to model the in-service behavior of these structures. For this purpose it is important to identify the constituent mechanical properties correctly [2, 3]. A series of tests has been developed to address this point, which includes: - hydrostatic compression - shear loading using a modified Arcan fixture This paper will describe the different test methods and present results obtained for different types of syntactic foams.

A Prediction Method of the Behavior of Adhesively Bonded Structures under Cyclic Shear Loading Based on a Characterization of the Viscous Aspects of the Adhesive in an Assembly

The capabilities of structural bonding are more and more used. Estimating the abilities of an adhesive to endure repetitive loadings and to keep stable its mechanical properties along service life is an essential point to analyze in order to conduct fatigue assessments. The aim of this study is to develop a predictive tool for describing the fatigue behavior of an adhesive in an assembly under cyclic loadings. The approach developed analyzes the influence of viscosity on the mechanical behavior of an adhesive in an assembly based on monotonic and creep test results. Thanks to the evaluation of viscous phenomena, it is possible to predict the cyclic response of the adhesive. The experimental approach uses a unique bonded joint designed to limit the stress concentrations and with a maximum stress state in the center of the adhesive. In this paper, following the strategy developed under monotonic loading, experimental results under cyclic loading are presented for different types of loading using several load ratios and amplitudes. These results underline that the evolution of viscous deformations depends on the loading type. Under shear loading and for a ductile structural adhesive, the experimental results are well described using a viscoelastic–viscoplastic constitutive model with nonlinear viscous parameters. This model makes it possible to analyze the influence of different parameters on the mechanical response of bonded joints under cyclic shear loadings.

Strategies for the analysis of the behavior of an adhesive in bonded assemblies

Experimental and numerical analyses of the mechanical behavior of bonded joints can be made particularly difficult by the influence of edge effects. Therefore, understanding the stress distribution in an adhesive joint can lead to improvements in adhesively-bonded assemblies. Such an analysis is proposed in the case of usual single lap shear specimens. Stress singularities can contribute to the initiation and propagation of cracks in the adhesive. Thus, in order to obtain reliable experimental data to analyze the nonlinear behavior of an adhesive in an assembly, tests which strongly limit the influence of stress singularities must be proposed. The design and the abilities of such a device for shear tests are presented. Moreover, some experimental results obtained using a modified Arcan fixture, which has been designed to strongly limit edge effects, are presented in the case of monotonic and complex history loadings. Furthermore, a 2D non associated elasto-visco-plastic model is proposed to accurately describe the experimental behavior under tensile-shear monotonic loadings. An extension of this model is also proposed to represent relaxation type effects under shear loadings.

Influence of edge effects on the load transmitted by adhesively-bonded assemblies

This paper presents a contribution to the influence of edge effects on the load transmitted by adhesively-bonded assemblies. Bonded assemblies are mostly characterized by the presence of stress concentrations at the bottom of the adhesive and in its thickness near the interface. This phenomenon, called edge effects, plays a very important role in the macroscopic observed behaviour of an adhesive in an assembly. Thus, to describe its behaviour it is necessary to propose solutions to reduce the edge effects for the bonded test specimens used to characterize the adhesive and for industrial applications. Mastering edge effects improves the efficiency of bonded parts and reduces the dispersions of the loads at failure. In the first part a modified Arcan test fixture is described to analyze the behaviour of an adhesive in an assembly under traction-shear loadings. The test specimens use a specific geometry to reduce the edge effects (beaks). In the second part, edge effects are analyzed by the finite element method under elastic hypothesis. In the third part, experimental results for traction-shear Arcan tests are presented. And finally, this method to reduce edge effects is applied to the bonding of a rail on a composite mast for high competitive sailing boats.

Numerical approach for the design of adhesively-bonded assemblies

This paper presents contributions of numerical modelling for the optimization of adhesivelybonded assemblies for marine and underwater applications. Difficulty in modelling the failure of even simple joints (lab shear specimens) highlighted the need for more reliable constituent input data [1 ; 2]. Therefore, in order to be able to study the behaviour of thin adhesive films up to failure, as a function of the normal stress, a modified Arcan fixture has been developed. Numerical simulations in linear elasticity, for bi-material structures show that the use of special geometry for the substrate (a beak close to the adhesive joint) makes it possible to eliminate the contribution of edge effects [3]. Non linear simulations taking into account the fixing system of the substrates on the supporting fixture were used to optimize the design of the complete fixture in order to prevent pre-loading of the adhesive joint. Moreover, in order to have more precise data for design, a study of the evolution of the stresses through the thickness of the adhesive joint under elastic assumption is proposed with respect to the shape of the edge of the adhesive joint. Non-contact extensometry and optimisation techniques have been used to determine the kinematics of the bonded joints deformation. For the epoxy resin Vantico™ Redux 420 the viscoplastic behaviour has been analyse for different radial loadings with the following variables: the applied load and the displacement of both extremities of the adhesive joint. This paper mainly presents the use of these experimental data, to develop a numerical tool to predict the behaviour of adhesively-bonded assemblies. To limit the numerical cost, we propose an approach using joint type elements which allows us to work directly with the relative displacement and the stress. As under elastic assumption, a non uniform stress field is observed in the adhesive joint, an inverse technique has been developed to identify the parameters of the model. For monotonic loadings, a plastic behaviour with isotropic hardening gives good results and allows us to analyse the non-isotropic behaviour of the adhesive with respect to tension-compression. In order to analyse the possibilities of this numerical approach, comparisons with experimental data are presented. A simple lab shear type specimen with thick substrates and beaks is proposed to analyse the influence of different parameters.