FRACTURE AND RESIDUAL STRESS BEHAVIOR OF STAINLESS STEEL WELDED PLATES: EXPERIMENTAL AND FINITE ELEMENT ANALYSES

E. Gonçalves M.A.C. Gonzales Escola Politécnica da Universidade de São Paulo – CEP: 05508-900 - SP - Brazil

Keywords: fracture behavior, residual stress, stainless steel welded, hole-drilling method, finite element. ABSTRACT Experimental and finite element analyses are combined to study fracture and residual stress behavior of stainless steel AISI304 welded plates. As the first step, 4 mm thick plates are TIG butt welded and residual stresses are experimentally evaluated. For the FEM computation, a rectangular plate is modeled with a mesh following the shape characteristics of the intended through thickness crack. In order to get a residual stress field, similar to one experimentally evaluated, plastic strains are induced in the modeled plate, by applying external displacements. By deactivating some of the finite elements, a crack is simulated and a modified residual stress field is obtained. By deactivating more elements, the crack grows and the residual stress field changes. As the crack length increases, the residual stresses and the stress intensity factors are evaluated. The finite element results were compared to the experimental ones, showing good agreement. INTRODUCTION Mechanical structures and components may fail during the service, even when submitted to loads smaller to the one established by the standard design approaches. The existence of residual stresses and cracks in welded joints, no doubt, are the main reasons for this behavior, [1]. The residual stresses are those stresses that remain in the body when this is not submitted to any type of external load. Usually, the residual stresses are caused by manufacturing processes, such as, conformation, foundry, shot blasting, welding, etc. Also, some of these processes can contribute to the emerging of cracks; highlighting among these, the welding processes. For this reason, in this work, emphasis has been given to the study of the fracture and residual stresses behavior of a welded steel structure plate. The main objective of this investigation is to evaluate the fracture behavior of a crack in the residual stress field. Usually, the residual stress numerical finite element evaluation of welded structures requires the simulation of entire process of welding, with enormous number of variables, as the ones related to structure and welded joint geometries, base and weld materials properties, different microstructures, gradients of temperature, heat transfer phenomena, etc. However, in order to obtain the residual stresses, a more simple method is considered in this work. Through the application of high intensities loads or displacements in the finite element plate model, it is pursued to obtain a residual stress field similar to the ones existing in the welded joints. Different sizes of cracks are considered to evaluate the residual stress field modifications and the stress intensity factors (K I).

1

EXPERIMENTAL PROCEDURE The test specimen, a TIG butt welded stainless AISI304 steel plate, as the one used in tanks of corrosive liquid transportation vehicle, are showed in Figure 1. The test specimen has the following dimensions: 250 mm long, 155 mm wide and 4 mm thick. The AISI304 steel mechanical properties are presented in the Chart 1.

FIGURE 1 –Tank transportation vehicle and the welded steel specimen. Modulus of Elasticity

193 GPa

Poison Coefficient

0.28 7900 kg/m3

Density Yield Limit

241 MPa

Ultimate Limit

586 MPa 55% Elongation Fracture Toughness 200 MPa . m1/2 Hardness 80 HRB CHART 1 - Mechanical properties of the AISI304 steel, [2]. The residual stresses are evaluated in five points of the test specimen. Two points on the weld line (A and B) and the other three points located at 8, 16 and 30 mm away from the weld line (respectively D, C and E). Figure 2 shows these points on the test specimen and the experimental apparatus used to evaluate the residual stresses by the Hole Drilling Method (HDM), [3].

y x

Point

Position x

A

0 mm

B

0 mm

C

16 mm

D

8 mm

E

30 mm

FIGURE 2 – Test Specimen with the strain gages positioned for the residual stress evaluation and experimental apparatus used to apply the hole drilling method.

2

In the Chart 2 are presented the experimental residual stresses Sxx, Syy and Sxy, where are also presented the calculated principal and equivalent residual stresses in the five points considered. The Sxx e Syy values are plotted in Figure 3.

A B D C E

σ1 σ2 σeq Distance Sxx Syy Sxy (mm) (MPa) (MPa) (MPa) (MPa) (MPa) (MPa) 0 185.7 243.7 -2.8 243.8 185.6 220.6 0 178.1 241.8 -22.4 248.9 171.0 220.5 8 70.0 45.3 -35.2 94.9 20.3 86.6 16 28.9 36.6 -2.3 37.2 28.3 33.7 30 -5.9 -56.5 11.8 -59.1 -3.3 57.5 CHART 2 – Residual stresses obtained by the hole drilling method 300

Residual stress (MPa)

250

Syy HDM

200

Sxx HDM

150 100 50 0 -50 -100 0

5

10

15

20

25

30

Distance from the weld line (mm) FIGURE 3 – Residual stress fields induced by the welding process. RESIDUAL STRESS FINITE ELEMENT ANALYSIS The plate is modeled using a finite elements mesh, composed by 7785 nodes, 6352 elements, with larger number of elements in the region around of the intended crack and following the geometry of the welded plate specimen. The stress numerical results obtained by using this mesh configuration and simple loading condition were in good agreement to the analytical ones. The Figure 4 shows the modeled samples and the finite element mesh. With the purpose to introduce the residual stress fields, a distributed and irregular node displacements field is applied to cause plastic deformations. The displacements were applied in the x and y directions. The displacement values are Ux1= -0.12 mm, Ux2= -0.13 mm, Ux3= -0.21 mm, Ux4= -0.37 mm, Uy1= 0.09 mm, Uy2= 0.32 mm, Uy3= 0.26 mm, Uy4= 0.06 mm and Uy5= -0.20 mm, which are applied in the positions indicated in Figure 5. Also, the boundary conditions are shown in Figure 5. The x and y d.o.f., respectively, are restrained for the nodes in the weld line plane (Rx) and for the nodes in the plane perpendicular to the welded joint (Ry). The model was developed and carried out by the commercial finite elements LS-DYNA- ANSYS code [4], for explicit modeling.

3

x

y z

FIGURE 4 – Modeling of a section of the welded steel sheet, isometric view.

FIGURE 5 – Application of the displacements and constraints on the plate model, top view. The time load evolution is presented in the Figure 6, in percentage of total load. The resulting residual stresses, evaluated by the finite element method, are shown in Figures 7 and 8, for x and y directions.

120% Application of displacements to induce the residual stresses

100%

Load

80% 60%

Crack introduction

40%

Application of the service load

20% 0% 0

0.2

0.4

0.6

0.8

1

1.2

time (ms) FIGURE 6 – Load evolution for the entire process.

4

1.4

1.6

a)

b)

FIGURE 7 – Residual stresses in the x (a) and y (b) directions.

Residual stress (MPa)

280 230

Syy HDM

180

Sxx HDM Syy FEM

130

Sxx FEM

80 30 -20 -70 -120 0

10

20

30

40

50

60

70

Distance from the weld line (mm) FIGURE 8 – Residual stress fields, in the x and y directions. FRACTURE BEHAVIOR Some finite elements were disabled to create a through thickness crack in the middle and transversely to the welded joint. It was considered three sizes of through thickness cracks, with 0.5, 1.0 and 2.0 mm half crack length (a). The mesh configuration near the crack is shown in Figure 9. It was observed that the introduction of the crack affected the residual stress field close to the crack. There are a stress relaxation due to the crack faces and a stress intensification due the crack tip.

5

FIGURE 9 – Deactivation of the finite elements to create the cracks.

Then, it was considered a static service load resulting in stresses parallel to the welded joint, uniformly distributed and equal to 120,5 MPa. This value of applied stress is equivalent having a design safety factor equal to two. The resulting stress fields are shown in the Figures 10 and 11.

a)

b)

FIGURE 10 – Stresses in the y direction, in the region close to the crack of 2.0 mm, with (a) and without (b) residual stresses, after applying a service load.

6

FIGURE 11 – Residual stresses in the y direction before and after the introduction of the cracks and the service load. The stress intensity factors (K I) are estimated by extrapolating to zero the K curves, as proposed by Chan et al [5] and schematically shown in Figure 12. The K values were calculated by using the Westergaard formulation below. K=

yy

. (2 r)1/2 / f( )

(KI)0.5 = 38.3 MPa.m1/2 (KI)1.0 = 33.0 MPa.m1/2 (KI)2.0 = 24.5 MPa.m1/2

FIGURE 12 – Stress Intensity Factor evaluation for different crack sizes and considering residual stress plus service loads.

7

The fracture safety factors (SFK) are estimated by relating the stainless steel AISI304 Fracture Toughness (KIC) to the Stress Intensity Factors (K I), for different load conditions and three crack sizes. Below, Chart 3 shows the fracture safety factor obtained for different conditions.

Load Condition Service Load Residual Stress Service Load + Residual Stress

a = 0.5 mm 41.9 4.9 5.2

SFK = KIC / KI a = 1 mm 29.6 5.6 6.1

a = 2 mm 20.9 8.0 8.2

CHART 3 – Fracture Safety Factors for different crack size (2a) and load conditions CONCLUSION The application and releasing of high-intensity displacements produces a residual stress field, in the x and y directions, similar to the one generated by the welding experiments. The introduction of the crack produces a redistribution of the residual stresses. This procedure showed to be adequate in evaluating the fracture behavior of welded structures, with no need to apply a very complex numerical welding simulation. The commercial finite elements code ANSYS [4] was appropriated for the non-linear analysis developed in this study. The existence of the residual stress may reduce the fracture safety factor by 10 times. For the load conditions and crack sizes considered, the fracture behaviors were quite unexpected. The smallest crack size, in a presence of a residual stress field only (no service load), was the most critical situation. REFERENCES [1] E.Gonçalves, Analysis of Fracture Welded Structures, PhD Thesis, Massachusetts Institute of Technology, USA, 1981. [2] E. Hornbogen, Werkstoffe, 5 th Edition, Springverlang, Berlin, 1991. [3] M. A. Calle Gonzales, E. Gonçalves, Design of a low-cost residual stresses measurement system for mechanical components, Proceeding of the 5th National Congress of Mechanical Engineering, Salvador, Brazil, 2008. [4] ANSYS 11.0, LS-DYNA, SAS IP, Inc, USA, 2007. [5] S.K.Chan et al, On the Finite Element Methods in Linear Fracture Mechanics, Eng.Fract.Mechanics, Vol 2, pp 1-17, Pergamon Press, 1970. Corresponding author: [email protected]

8