Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Non-Linear Behaviour of Concrete Structures Under Severe Impact

Télécharger au format pdf ou txt
Télécharger au format pdf ou txt
Vous êtes sur la page 1sur 8

CONSEC07 Tours, France

NON-LINEAR BEHAVIOUR OF CONCRETE STRUCTURES UNDER SEVERE IMPACT


COMPORTEMENT NON LINEAIRE DE STRUCTURES EN BETON SOUMISES A UN CHOC SEVERE Rainer ZINN1), Friedhelm STANGENBERG1), Michael BORGERHOFF1) Danile CHAUVEL2), Jean-Pierre TOURET2)
1 2

Stangenberg und Partner, Bochum, Germany EDF/SEPTEN, Villeurbanne, France

ABSTRACT - The protection against external hazards for nuclear power plant structures has been increased by taking into account high levels of earthquake and also high energy airplane crash. Methodologies have to be developed to evaluate the ability of protective concrete shells to cope with these evolutions of design requirements. In the present paper, non-linear analyses of reinforced concrete structures under severe impact loads with different finite element (FE) programs are compared with each other and with large-scale impact tests (Meppen slab tests). The post-computations of these Meppen tests with ADINA and SOFiSTiK show a good agreement between measured and computed results, and it can be concluded that those two programs are well-suited for impact analyses of reinforced concrete structures in the non-linear range. RSUM La protection des structures de centrales nuclaires contre les agressions d'origine externe a t amliore en tenant en compte des niveaux de sisme levs et galement dimpact davion de taille importante. Des mthodologies doivent tre dveloppes pour valuer la capacit des enveloppes protectrices rsister ces nouvelles exigences de conception. Ce papier prsente les rsultats des analyses non linaires de structures en bton arm soumises des chocs avec diffrents programmes aux lments finis et compars des essais grande chelle (essais Meppen sur dalles). Les simulations numriques de ces essais avec ADINA et SOFISTIK montrent une bonne concordance entre les rsultats de calculs et les mesures, et on peut conclure que ces programmes sont bien adapts pour des analyses dimpact des structures en bton arm dans le domaine non linaire. 1. Introduction The protection against more and more severe external hazards obliges us to investigate a domain beyond usual regulatory codes criteria, that is to say perform non linear calculations.The programs utilized for the non-linear dynamic FE analysis are ADINA (cf. Bathe 1989) and SOFiSTiK (cf. SOFiSTiK AG 2005). Both programs are suited for the calculation of static and dynamic effects of general loading on any type of structure, which has to be divided into an assembly of individual elements interconnected at nodes (FE method). Both programs include arbitrary non-linear assumptions for the materials concrete and reinforcing steel. Whereas SOFiSTiK is based on a layered concrete model regarding the reinforcement at both sides in the used shell elements, in ADINA the concrete is modelled by volume elements, and the reinforcing steel is modelled by truss elements. The

CONSEC07 Tours, France

material model of concrete includes cracking and crushing effects as well as tension stiffening and strain softening. The influence of high strain-rates within the structure, which causes property changes in all used materials, is regarded according to published experimental results. In order to validate the computational approach, post-computations of experimental extreme impacts on reinforced concrete targets are presented. The computations refer to the so-called Meppen slab tests in Germany, where large-scale impact tests have been performed in order to investigate the ultimate load-bearing behaviour of reinforced concrete structures under high-speed impact loads. 2. Computational Approach 2.1 SOFiSTiK The computer program SOFiSTiK (2005) is suited for the calculation of static and dynamic effects of general loading on any type of structure, which has to be divided into an assembly of individual elements interconnected at nodes (Finite Element Method). In the current investigation, this program is used for dynamic analyses of models, which consist of shell elements. The numerical methods of the utilized computer program include the consideration of the effects caused by non-linear material behaviour of the shell structures and the investigation of 2nd and 3rd order theory effects due to geometrically non-linear behaviour. The SOFiSTiK shell element is implemented as triangular as well as quadrilateral element. Since the element remains plane, the bending and membrane actions of the individual element are decoupled. For the plate action, the shell element is based on Mindlins plate theory with an extension by a non-conforming formulation. The element formulation of the membrane stress state is carried out by a classical isoparametric formulation or by a similarly classical non-conforming formulation by Wilson and Taylor. The analysis of non-linear effects in SOFiSTiK is done by iterations using a modified Newton method with constant stiffness matrix. The non-linear material behaviour of reinforced concrete in the shell elements is regarded by use of a layer model with arrangement of the crosswise reinforcement layers in their correct positions near the surfaces. The non-linear behaviour of the components of reinforced concrete is defined by non-linear uniaxial stress-strain laws of concrete (increase of strength due to biaxial compressive behaviour regarded), consideration of tension softening of concrete after cracking dependent on fracture energy, trilinear stress-strain laws of reinforcing steel. For geometrically non-linear analysis the initial stress matrix is added to the stresses of the primary stress state. In this way, a stability failure is recognized even in cases without unplanned initial deformation. A limitation of SOFiSTiK is that shear deformations of shell/plate elements are only approximately included by comparison of the elements shear stresses with an ultimate shear strength to be specified by the user.

CONSEC07 Tours, France

2.2 ADINA The computer code ADINA (Automatic Dynamic Incremental Nonlinear Analysis), (2005) and Bathe (1989), comprises solution capabilities for two- and three-dimensional non-linear finite element analysis of concrete structures. In the ADINA computations performed in this study, the concrete is modelled by volume elements, and the reinforcing steel (bending reinforcement bars and stirrups) is modelled by truss elements. The implemented concrete material law includes the consideration of triaxial non-linear stress-strain behaviour, tensile cracking, tensile stiffening, compression crushing and strain softening. To model the failure of the concrete material in tension and compression and to account for multiaxial conditions in the uniaxial stress-strain behaviour, failure envelopes are employed. As failure criterion, the ultimate strength surface according to Ottosen (1977) proposed in the CEB-Bulletin No. 156 (1983) is used. The stress-strain law for the truss elements representing the reinforcing steel, used in the computations, is elastic-plastic with kinematic hardening conditions. 3. Numerical Computations In order to validate the computational approach, post-computations of extreme impacts on reinforced concrete targets (so-called Meppen slab tests in Germany) are presented, where large-scale impact tests of reinforced concrete structures under high-speed impact loads have been performed, cf. Riech et al. (1984). The mass of the (deformable) impacting body was about 1000 kg, the impact velocities were in the range 200 up to 250 m/s. The projectile was a steel projectile with length 6 m, diameter 60 cm and wall thickness 7 resp. 10 mm. The target was a reinforced concrete slab of total dimension 6.5 x 6.0 m. Figures 1 3 show photos of slabs after tests, in the foreground of Fig. 1 there is the deformed projectile (residual length of projectile after test 1 - 2 m). The finite element model for the computations with the program SOFiSTiK is shown in Fig. 4 (left hand) together with the substitute load area of the plate with thickness 0.9 m and the fixed supports for out-of-plane motions. The number of plate elements of the FE model is 1,296. The ADINA model shown in Fig. 4 (right hand) consists of 2,645 nodes, 1,936 volume concrete elements and 2,553 truss steel elements.

Figure 1. Meppen test II/21, slab and projectile (left), rear side of slab (right) after test

CONSEC07 Tours, France

Figure 2. Meppen test II/12, front side (left) and back side (right) of slab after test

Figure 3. Meppen test II/20, front side (left) and back side (right) of slab after test

Figure 4. SOFiSTiK-FE model with supports and load area (left), ADINA model (right)

CONSEC07 Tours, France

In Table I, the data of projectiles and targets for selected tests of the last test series II/11 II/21 are listed. In addition to the values listed in Table I a 10 % increase of concrete and steel strength due to strain-rate effects as observed by Amman et al. (1982), Brandes et al. (1986), Eibl et al. (1999), Mller et al. (1983), is regarded in the performed calculations. Table I. Test data of Meppen slab tests II/11 II/21 Target Slab Test Mass Velocity Kg 1000 980 954 1060 972 1000 m/s 222.5 241.5 178.4 237.4 197.7 237.0 Thickn. cm 70 70 50 70 50 90 Concrete fc MN/m 32.9 39.6 33.9 31.2 33.6 33.2 Steel fyk 467 485 473 473 504 469 Reinforcement Upper f. lower f. stirrups cm/m 20.4 20.4 23.1 20.4 23.1 10.5 cm/m 40.1 40.1 56.0 40.1 58.2 25.5 cm/m 37.7 52.3 64.9 52.3 87.0 26.5

II/11 II/12 II/17 II/18 II/20 II/21

In this paper, comparative calculations for the tests II/12, II/20 and II/21 are documented (ADINA only II/12 and II/21). The slab dimensions and the arrangement of measurement equipment are shown in Fig. 5. In Table II additional information on the data of the plates of the 3 tests selected for post-computations are given. In Figures 6 - 8 measured displacements of the tests II/12, II/20 and II/21 are shown; they refer to displacement gauges W1 - W10 which are indicated in Fig. 5.

Fig. 5. Meppen slab with measurement equipment

CONSEC07 Tours, France

Table II. Additional data of plates tests II/12, II/20 and II/21 Test II/12 II/20 II/21 h (cm) d (cm) 70 50 90 65 44 85 as 25 e = 12.3 cm 20 e = 12.3 cm as 18 e = 12.5 cm 18 e = 24.2 cm as, 16.6 20 /m 27.7 20 /m 10.4 18 /m

25/28 e = 7/12 cm 18 e = 11.0 cm

Figure 6. Meppen test II/12, measured displacements

Figure 7. Meppen tests II/20 (left), II/21 (right), measured displacements

Figure 8. Meppen test II/21, measured displacements

Fig. 9 (left hand side) shows the load-time functions which have been derived from measured support reaction forces. The resulting displacement time-histories of the

CONSEC07 Tours, France

SOFiSTiK computations are shown in Figures 9 (right hand side, test II/12) and 10 (tests II/20 and test II/21, the corresponding time-histories of the ADINA computations are given in Figure 11. The computed values are in good agreement with the measured displacements concerning the displacement amplitudes, the residual plastic deformations as well as the vibration behaviour.
20
80
II/12 II/20 II/21

Displacement [mm]

16

60

Load [MN]

MW9 W10 W6 W8

12 8 4 0 0 10 20 30 40

40

20

0 0 0.02 0.04 0.06 0.08 0.1

Time [ms]

Time [s]

Figure 9. Load-time functions (left), comp. displacements II/12 (SOFiSTiK, right)


100

60

Displacement [mm]

80 60 40 20 0 0 0.02 0.04 0.06

Displacement [mm]

MW9 W10 W6 W8

40

MW9 W10 W6 W8

20

0
0.08 0.1

0.02

0.04

0.06

0.08

0.1

Time [s]

Time [s]

Figure 10. Meppen tests II/20 (left) and II/21 (right), comp. displacements (SOFiSTiK)
60

50

Displacement [m/s]

Displacement [m/s]

40

MW9 W9 W6 W8

40 30 20 10 0

MW9 W9 W6 W8

20

0 0 20 40 60 80 100

20

40

60

80

100

Time [ms]

Time [ms]

Figure 11. Meppen tests II/12 (left) and II/21 (right), computed displacements (ADINA)

CONSEC07 Tours, France

4. Conclusions The post-computations of the extreme impact tests at the Meppen test site (Germany) with the two different FE programs SOFiSTiK and ADINA showed a good agreement between measured and computed displacements, which had been to a far extent in the non-linear range with considerable residual displacements. The ADINA version 8 (ADINA 2005) shows a better representation of the behaviour after the displacement maximum in comparison with former program versions. It can be concluded that those two programs are well-suited for impact analyses of reinforced concrete structures in the non-linear range. A limitation of SOFiSTiK is that shear deformations of shell/plate elements are only approximately included, cf. chapter 2.1. The tests II/12, II/20 and II/21 for which postcomputations are documented in this paper predominantly showed bending deformations without significant shear or punching effects. The overall deformation behaviour of such tests can also be computed with simple 2-dof-systems, cf. Schwarzkopp et al. (1987), but with loss of information on stress and strain distributions over the plate.

5. References ADINA R&P, Inc, (2005) ADINA A Finite Element Program for Automatic Dynamic Incremental Nonlinear Analysis. Report 05-2 (ADINA 8.3). Ammann, W., Mhlematter, M., Bachmann, H. (1982) Zugversuche an Bewehrungs- und Spannstahl mit erhhter Dehngeschwindigkeit. Institut fr Baustatik und Konstruktion ETH Zrich, Bericht Nr. 7709-1, Birkhuser Verlag, Basel. Bathe, K.-J. (1989) Nonlinear Analysis of Concrete Structures. Computers & Structures Vol. 32, No. 3 / 4, pp. 563-590. Brandes, K., Limberger, E., Herter, J., Berner, K. (1986) Kinetic Load Bearing Capacity of Reinforced Concrete Members under Impact Load, Reinforcing Steel Tension Tests with High Strain Rates. Bundesanstalt fr Materialprfung, Forschungsbericht 129. CEB, Comit Euro-International du Beton, (1983), Bulletin dInformation No. 156. Eibl, J., Schmidt-Hurtienne, B. (1999) Betonstoffgesetze fr hochdynamische Beanspruchungen. Beton- und Stahlbetonbau 94, pp. 278-288. fib (fderation internationale du bton) (2001) Punching of structural concrete slabs. Technical Report, Bulletin 12. Mller, F. P., Keintzel, E., Charlier, H.(1983) Dynamische Probleme im Stahlbetonbau, Teil I: Der Baustoff Stahlbeton unter dynamischer Beanspruchung. Deutscher Ausschuss fr Stahlbeton, Heft 342, Berlin. Ottosen, N. S., (1977), Journ. Eng. Mec. Div. ASCE, Vol. 103, EM4. Riech, H., Rdiger, E. (1984) Meppener Versuche II/11 bis II/22, 1500408 (RS467), II. Technischer Bericht Schwarzkopp, D., Zinn, R. (1987) Simplified Nonlinear Analysis of Reinforced Concrete Structures under Projectile Impact Including Penetration and Projectile Deformation Effects. Proc. 1st Int. Conf. On Concrete for Hazard Protection, Edinburgh. SOFiSTiK AG (2005) SOFiSTiK, Analysis Programs, Version 21.0, Oberschleissheim.

Vous aimerez peut-être aussi