Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Crack growth resistance in metallic alloys: the role of isotropic versus kinematic hardening

Crack growth resistance in metallic alloys: the role of isotropic versus kinematic hardening The sensitivity of crack growth resistance to the choice of isotropic or kine- matic hardening is investigated. Monotonic mode I crack advance under small scale yielding conditions is modelled via a cohesive zone formulation endowed with a traction-separation law. R-curves are computed for materials that exhibit linear or power law hardening. Kinematic hardening leads to an enhanced crack growth resistance relative to isotropic hardening. Moreover, kinematic hardening requires greater crack extension to achieve the steady state. These differences are traced to the non-proportional loading of mate- rial elements near the crack tip as the crack advances. The sensitivity of the R-curve to the cohesive zone properties and to the level of material strain hardening is explored for both isotropic and kinematic hardening. Keywords: Kinematic hardening, Isotropic hardening, Cohesive zone modelling, Finite element analysis, Fracture 1. Introduction It is well established that material elements near a mode I crack tip undergo non-proportional straining due to crack advance, see for example the early analysis of crack growth by Rice and Sorensen [1]. The degree of hysteresis associated with this non-proportional loading is sensitive to the nature of the hardening law of the solid. For example, it is to be expected that kinematic hardening leads to greater hysteresis than isotropic hardening. Corresponding author. Tel: +44 1223 748525; fax: +44 1223 332662. Email address: mail@empaneda.com (Emilio Mart´ınez-Pan ˜eda) Preprint submitted to Journal of Applied Mechanics June 26, 2018 arXiv:1806.08986v1 [cond-mat.mtrl-sci] 23 Jun 2018 Consequently, one might expect that the choice of plastic hardening law will influence the stress intensity factor K versus crack extension Δa response, widely known as the R-curve. However, little attention has been paid to the effect of the hardening law upon crack growth resistance and no clear picture emerges from the literature. Lam and McMeeking [2] analysed steady state crack tip fields and concluded from a crack opening displacement criterion that isotropic hardening augments crack growth resistance. Carpinteri [3] performed finite element analyses of crack propagation by means of a strain- based criterion and observed a greater amount of crack extension in the kinematic hardening case for a given remote load; this also suggests that isotropic hardening increases crack growth resistance. In contrast, we shall demonstrate that kinematic hardening significantly raises the level of plastic dissipation and, thereby, elevates the R-curve along with the steady state fracture toughness K . SS 2. Numerical model We consider the small scale yielding problem of a plane strain mode I crack subjected to a remote stress intensity factor K. The elasto-plastic solid is isotropic with a Young’s modulus E, a Poisson’s ratio ν and an initial yield strength σ . Throughout our study we shall take ν = 0.3 and σ /E = 0.003. 0 0 We denote the Cauchy stress by σ and define s as its deviatoric part, such ij ij that s = σ − δ σ /3. The plastic response involves either isotropic or ij ij ij kk kinematic hardening, as follows. Isotropic hardening: The yield condition reads Φ = σ − σ = 0 (1) e Y where σ is the von Mises effective stress and the current yield strength σ is e Y a function of the accumulated von Mises plastic strain ε . Accordingly, the effective stress σ in J2 plasticity is defined as σ = s s (2) ij ij The increment in plastic strain ε˙ is computed from the normality hy- ij pothesis, ∂Φ 3 s p ij ε˙ = ε˙ = ε˙ (3) e e ij ∂σ 2 σ ij Y 2 in terms of the increment in effective plastic strain ε˙ . The relation between σ and ε is given by the uniaxial tensile response, such that the true tensile Y e stress σ is related to the true tensile plastic strain ε by Eε σ = σ 1 + (4) where N is the strain hardening exponent. In addition to this power law description we also consider the case of linear hardening by taking N = 1 and by replacing E in (4) with the tangent modulus E . Kinematic hardening: Assume that the centre of the yield surface is located at the point α in deviatoric stress space. We shall refer to α as a back- ij ij stress, and assume that the Armstrong-Frederick non-linear rule [4] defines the evolution of this backstress, such that (s − α ) ij ij α˙ = c ε˙ − γα ε˙ (5) ij e ij e where c and γ are material constants. This rule reproduces ratchetting when a material element is subjected to a non-zero mean stress and cyclic loading but predicts a particular shape of the stress-strain curve (see for example [5]). In order to model a more general shape of the uniaxial tensile response the constitutive statement (5) has been extended by Chaboche [6]. He replaced the single backstress α by a finite number n of backstresses α , such that ij ij α = α (6) ij ij k=1 Each backstress α evolves with ε˙ according to the independent hardening ij ij rule, (σ − α ) ij ij k k k k α˙ = c ε˙ − γ α ε˙ (no sum on k) (7) e e ij ij k k in terms of the material constants c and γ . The resulting Chaboche- Armstrong-Frederick (CAF) model has been widely used to capture ratchet- ting effects and non-linear hardening under non-proportional cyclic loading [7, 8]. We consider here the case of power law hardening (4) and select the k k values of c , γ for k ∈ (1, n) such that the desired response in unaxial ten- sion is obtained. The choice of n = 10 brings the CAF model into alignment 3 with (4) to within 0.04% for the range of true tensile strain 0 ≤ ε ≤ 2.0. The uniaxial stress-strain response for cyclic loading is given in Fig. 1a for the case N = 0.2. In the case of linear hardening (5) is used instead of (6)-(7), with γ = 0 and c = E ; this is the familiar Ziegler formulation [9]. The hardening laws employed are shown in Fig. 1b; power law hardening for the choices N = 0.1 and N = 0.2, and linear hardening for E /σ = 5/3 and t 0 E /σ = 50/3. t 0 We model tensile fracture at the tip of a mode I crack by means of a cohesive zone model, following Tvergaard and Hutchinson [10] - see Fig. 2. Cohesive zone formulations have a long history back to Dugdale [11] and Barenblatt [12]: fracture is regarded as a gradual process in which separa- tion takes place across an extended cohesive zone, and is resisted by cohesive tractions. As shown in Fig. 3, we shall make use of a trapezoidal traction- separation law of strength σˆ, with its shape being defined by a critical cohe- sive separation δ and by two shape parameters δ = 0.15δ and δ = 0.5δ . c 1 c 2 c The work of fracture Γ is given by the area under the traction-separation curve, such that Γ = σˆ (δ + δ − δ ) (8) 0 c 2 1 A reference stress intensity factor for crack growth initiation follows im- mediately as EΓ K = (9) (1 − ν ) along with a reference length R , where 1 EΓ 1 K 0 0 R = = (10) 3π (1 − ν ) σ 3π σ The crack tip is placed at the origin and the crack plane is aligned with the negative x axis of the Cartesian reference frame (x , x ). A remote 1 1 2 K field is imposed by a boundary layer formulation, as follows. The outer periphery of the mesh is subjected to the mode I elastic K-field, 1/2 u = r f (θ, ν) (11) i i 4 where r and θ are polar coordinates centred at the crack tip and the functions f (θ, ν) are given by, 1 + ν θ f = √ (3 − 4ν − cos θ) cos (12) 2π 1 + ν f = (3 − 4ν − cos θ) sin (13) 2π Upon exploiting the symmetry about the crack plane, only half of the model is analysed, as shown in Fig. 4. The finite element model is imple- mented in the commercial finite element package Abaqus [13] and we solve the boundary value problem by an implicit Backward Euler integration scheme. Plane strain quadratic quadrilateral elements are employed, with the mesh comprising 267272 degrees of freedom. A refined mesh was used along the co- hesive zone in order to obtain a converged solution. The characteristic length of the elements in the crack propagation region is chosen to be equal to δ . Cohesive elements with 6 nodes and 12 integration points are implemented by means of a user element (UEL) subroutine [14]. A control algorithm is used to avoid convergence problems due to snap-back instabilities, see [15, 16] for details. Computations have been performed within an infinitesimal de- formation framework since strains remain small, as argued by Tvergaard and Hutchinson [10] in their finite strain analysis. Dimensional analysis shows that the solution, given in terms of the remote K = K (Δa), is a function I R F of the following dimensionless quantities, K Δa σ σˆ E δ δ R 0 t 1 2 = F , , , , ν, N, , (14) K R E σ E δ δ 0 0 0 c c We will conduct calculations until steady state crack growth at constant K is attained. Tvergaard and Hutchinson [10] showed that K is sensitive SS SS to the ratio of cohesive strength σˆ to material yield strength σ . For an elas- tic, perfectly plastic solid, K /K raises steeply as σˆ/σ approaches 3. The SS 0 0 interpretation is straightforward by considering a stationary crack in an elas- tic, ideally plastic solid absent of a cohesive zone. The tensile stress directly ahead of the crack tip equals 3σ as given by the Prandtl field. Consequently, if σˆ/σ exceeds 3, the crack tip blunts without advance as the cohesive zone strength is not overcome. 5 Tvergaard and Hutchinson [10] also considered the role of isotropic strain hardening on the R-curve. In this case, the stress field ahead of the crack tip exceeds 3σ due to the presence of strain hardening. For σˆ/σ < 3, a shallow 0 0 R-curve is exhibited and K /K is slightly above unity. In contrast, for SS 0 σˆ/σ > 3, a steeper R-curve is observed and K /K increases its sensitivity 0 SS 0 to σˆ. A major aim of the present study is to explore the sensitivity of the R-curve to the nature of the hardening law: isotropic versus kinematic hardening. 3. Results R-curves are shown in Fig. 5 for linear hardening and σˆ/σ = 3.5; this value of σˆ/σ is close to the limiting value of σˆ/σ = 3 for an elastic, per- 0 0 fectly plastic solid, as discussed in the previous section. Consider first the R-curve for a small level of strain hardening E /σ = 5/3. A steeply ris- t 0 ing R-curve is predicted, which will give rise to a large steady state fracture toughness K /K and a large value of the crack extension to achieve steady SS 0 state (Δa/R ) . The steep R-curve is a consequence of plastic dissipation SS with crack advance. Little difference is observed between the kinematic and isotropic hardening predictions since the degree of hardening is small. We note in passing that for the elastic, ideally plastic case, E /σ = 0, and t 0 σˆ/σ = 3.5 no crack advanced is observed: the tensile traction ahead of the crack tip is insufficient to overcome the cohesive zone strength. Now consider the case of a high strain hardening rate E /σ = 50/3. The strain level near t 0 the crack tip can now exceed the cohesive strength σˆ at a relatively low value of plastic strain. A shallow R-curve is predicted. Again, kinematic hardening elevates the R-curve compared to the isotropic hardening case. The predicted R-curves for the case of a power law hardening solid are shown in Fig. 6 for the choice of N = 0.1. We consider both isotropic and kinematic hardening, and selected values of the cohesive strength σˆ/σ = 3.2, 3.4 and 3.5. As expected, increasing σˆ/σ elevates the K versus Δa response for both hardening laws. However, the R-curves are more sensitive to the cohesive strength for the case of kinematic hardening. Also, as in the linear hardening study, the R-curves are steeper for kinematic hardening, implying a higher value of the steady state fracture toughness K . We note that SS there is no straightforward relationship between K , K and the crack ini- 0 SS tiation toughness K as defined in the standard test methods for fracture Ic 6 toughness, such as the ASTM E 1820 [17]. The standard defines K as the Ic value of K corresponding to a crack growth increment which is in the range of 0.2 - 0.5 mm, see [18] for a discussion. When the R-curve is steep, K /K Ic 0 may be large. The steady state toughness is attained when K reaches a plateau value. Fig. 7 shows the sensitivity of K to the strain hardening exponent and SS to the cohesive strength. The isotropic curves are in agreement with the results of Tvergaard and Hutchinson [10]. First, note that for N = 0 a single K /K versus σˆ/σ curve corresponds to the cases of isotropic and kine- SS 0 0 matic hardening. The value of the steady state toughness increases rapidly in the vicinity of the limiting value of σˆ/σ = 2.8. With increasing N, kinematic and isotropic hardening theories give increasingly divergent pre- dictions. Consistently, for N > 0, kinematic hardening leads to a higher value of K /K at a given σˆ/σ than does isotropic hardening. Also, the SS 0 0 value of the cohesive strength at which K /K increases rapidly is lower for SS 0 the case of kinematic hardening. It is instructive to consider the value of crack extension (Δa) that is SS required to achieve the steady state toughness. The dependence of (Δa) SS upon σˆ/σ is shown in Fig. 8 for the power law hardening solid, for both kinematic and isotropic hardening. Note that both K /K and (Δa) /R SS 0 0 SS depend upon σˆ/σ in a highly non-linear manner for both hardening laws, recall Figs. 7 and 8. Is there a simple relation between (Δa) /R and SS K /K ? This might be expected as the plastic zone size associated with SS 0 K = K , is of the order, SS SS R = (15) SS 3π σ Assume that (Δa) is proportional to R , SS SS (Δa) = CR (16) SS SS where the constant C is of order unity with some sensitivity to the choice of the hardening law and to N. It follows immediately that, (Δa) R K SS SS SS = C = C (17) R R K 0 0 0 7 The accuracy of this prediction is shown by a cross-plot of (Δa) /R SS versus K /K in Fig. 9, with σˆ/σ as the parameter trending variable. A SS 0 0 curve fit reveals that C increases from 0.574 to 1.703 as N goes from 0.1 to 0.2 for kinematic hardening, and C increases from 0.31 to 0.496 as N goes from 0.1 to 0.2 for isotropic hardening. Our numerical predictions show that the isotropic hardening idealization may significantly underestimate the degree of subcritical crack propagation before catastrophic failure. Note further that (17) can be re-expressed in the form SS Δa = (18) SS 3π σ What is the physical basis for the steeper R-curve observed in the case of kinematic hardening? We show in Fig. 10 that significant non-proportional loading occurs in the vicinity of the crack tip, as the crack advances. Consider a representative material point P at a distance for 2R ahead of the initial crack tip and slightly above the cracking plane (height of 0.1R ). Allow the crack to advance by Δa = 2R for both cases of isotropic and kinematic hardening. The active plastic zone is shown in Fig. 10a for Δa = 0 and Δa = 2R . The plastic zone at Δa = 0 for kinematic hardening is identical to that for isotropic hardening, whereas the plastic zone at Δa = 2R is much larger in the kinematic hardening case. Only the isotropic hardening active plastic zone is shown at Δa = 2R for the sake of clarity. The stress paths imposed on point P for isotropic and kinematic hardening are given in Fig. 10b. Differences between kinematic and isotropic stress paths arise soon after cracking initiates (Δa = 0 ), due to non-proportional straining in the neighbouring points. As the crack advances not only are the stress paths non-proportional but they also deviate from each other, with the greatest change in stress direction given by isotropic hardening. The stronger path dependence of kinematic hardening also plays an important role on localiza- tion in thin sheets and in shear localization. For example, Tvergaard [19] showed that the forming limit curves predicted by kinematic hardening are in better agreement with experimental results than isotropic hardening pre- dictions. The dependence of the critical strain for shear localization upon the local curvature of the yield surface has been investigated by Mear and Hutchinson [20]. In addition, we investigate the level of energy dissipation in the main plastic zone in the vicinity of the crack tip (denoted W ) and in the secondary 8 plastic zone that arises in the crack wake (denoted as W ). Here, the energy dissipated due to plastic deformation is computed for all material elements in the active plastic zone as, Z Z Δa W (Δa) = σ ε˙ dV da (19) ij ij The predictions are given in Table 1 for various stages of crack advance. Computations reveal that plastic work in the secondary plastic region is neg- ligible relative to the energy dissipated in the vicinity of the crack. We con- clude that differences between isotropic and kinematic hardening responses are mainly due to non-proportional deformation in the crack tip plastic zone. Also, Table 1 shows that kinematic hardening involves a much larger plas- tic dissipation energy than isotropic hardening, and is consistent with the steeper R-curve. Table 1: Plastic energy dissipation with crack advance in the primary plastic zone region at the crack tip W and the secondary plastic zone region at the crack wake W . Material 1 2 properties: δ /δ = 0.15, δ /δ = 0.5, σ /E = 0.003, ν = 0.3, N = 0.1 and σˆ = 3.5σ . 1 c 2 c 0 0 Isotropic Kinematic Δa/R W / (Γ Δa) W / (Γ Δa) W / (Γ Δa) W / (Γ Δa) 1 0 2 0 1 0 2 0 0.5 21.98 0.01 52.69 0.42 1 17.24 0.14 58.36 0.85 2 8.21 0.17 78.75 1.08 4. Conclusions We investigated how the isotropic or kinematic nature of strain hard- ening influences crack growth resistance. Finite element results show very significant differences between isotropic and kinematic hardening laws that yield the same response under uniaxial tension. We show that kinematic hardening notably enhances plastic dissipation and the steady state frac- ture toughness K . These differences persist over different hardening levels, SS cohesive strengths, and hardening profiles. 9 5. Acknowledgments The authors acknowledge valuable discussions with V.S. Deshpande (Uni- versity of Cambridge). E. Mart´ınez-Pan˜eda also acknowledges valuable in- sight from K. Juul and C.F. Niordson (Technical University of Denmark). The authors would like to acknowledge the funding and technical support from BP (ICAM02ex) through the BP International Centre for Advanced Ma- terials (BP-ICAM). E. Mart´ınez-Pan˜eda additionally acknowledges financial support from the Ministry of Economy and Competitiveness of Spain through grant MAT2014-58738-C3 and the People Programme (Marie Curie Actions) of the European Union’s Seventh Framework Programme (FP7/2007-2013) under REA grant agreement n 609405 (COFUNDPostdocDTU). References [1] J. R. Rice, E. P. Sorensen, Continuing crack-tip deformation and frac- ture for plane-strain crack growth in elastic-plastic solids, Journal of the Mechanics and Physics of Solids 26 (3) (1978) 163–186. [2] P. S. Lam, R. M. McMeeking, Analyses of steady quasistatic crack growth in plane strain tension in elastic-plastic materials with non- isotropic hardening, Journal of the Mechanics and Physics of Solids 32 (5) (1984) 395–414. [3] A. Carpinteri, Crack growth resistance in non-perfect plasticity: Isotropic versus kinematic hardening, Theoretical and Applied Fracture Mechanics 4 (2) (1985) 117–122. [4] C. O. Frederick, P. J. Armstrong, A mathematical representation of the multiaxial Bauschinger effect, Materials at High Temperatures 24 (1) (2007) 1–26. [5] A. F. Bower, Applied Mechanics of Solids, CRC Press, Taylor & Francis, Boca Raton, 2009. [6] J. L. Chaboche, On some modifications of kinematic hardening to im- prove the description of ratchetting effects, International Journal of Plas- ticity 7 (7) (1991) 661–678. [7] J. Lemaitre, J.-L. Chaboche, Mechanics of solid materials, Vol. 19, Cam- bridge University Press, New York, 1990. 10 [8] J. L. Chaboche, A review of some plasticity and viscoplasticity consti- tutive theories, International Journal of Plasticity 24 (10) (2008) 1642– [9] H. Ziegler, A Modification of Prager’s hardening rule, Quarterly of Ap- plied Mathematics 17 (1) (1959) 55–65. [10] V. Tvergaard, J. W. Hutchinson, The relation between crack growth re- sistance and fracture process parameters in elastic-plastic solids, Journal of the Mechanics and Physics of Solids 40 (6) (1992) 1377–1397. [11] D. S. Dugdale, Yielding of steel sheets containing slits, Journal of the Mechanics and Physics of Solids 8 (2) (1960) 100–104. [12] G. I. Barenblatt, The Mathematical Theory of Equilibrium of Crack in Brittle Fracture, Advances in Applied Mechanics 7 (1962) 55–129. [13] Abaqus/Standard 2017. Dassault Systemes SIMULIA, Providence, RI, USA. [14] S. del Busto, C. Betego´n, E. Mart´ınez-Pan˜eda, A cohesive zone frame- work for environmentally assisted fatigue, Engineering Fracture Mechan- ics 185 (2017) 210–226. [15] J. Segurado, J. LLorca, A new three-dimensional interface finite element to simulate fracture in composites, International Journal of Solids and Structures 41 (11-12) (2004) 2977–2993. [16] E. Mart´ınez-Pan˜eda, S. del Busto, C. Betego´n, Non-local plasticity ef- fects on notch fracture mechanics, Theoretical and Applied Fracture Mechanics 92 (2017) 276–287. [17] ASTM E 1820-01, Standard Test Method for Measurement of Fracture Toughness. [18] T. L. Anderson, Fracture Mechanics. Fundamentals and Applications, 3rd Edition, CRC Press, Taylor & Francis, Boca Raton, 2005. [19] V. Tvergaard, Effect of kinematic hardening on localized necking in biaxially stretched sheets, International Journal of Mechanical Sciences 20 (9) (1978) 651–658. 11 [20] M. E. Mear, J. W. Hutchinson, Influence of yield surface curvature on flow localization in dilatant plasticity, Mechanics of Materials 4 (1985) 395–407. 12 List of Figures 1 Uniaxial stress strain response for (a) cyclic loading of a non- linear hardening solid with N = 0.2, and (b) half-cycle for linear and non-linear hardening. Material properties: σ /E = 0.003. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2 Schematic representation of the cohesive zone model for fracture. 15 3 Cohesive traction T - separation δ law characterising the frac- ture process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 4 Finite element mesh and configuration of the boundary layer. . 17 5 Crack growth resistance curves for linear isotropic and kine- matic hardening plasticity and different hardening levels. Ma- terial properties: δ /δ = 0.15, δ /δ = 0.5, σ /E = 0.003, 1 c 2 c 0 ν = 0.3, and σˆ = 3.5σ . . . . . . . . . . . . . . . . . . . . . . . 18 6 Crack growth resistance curves for power law isotropic and kinematic hardening plasticity and different levels of the cohe- sive strength. Material properties: δ /δ = 0.15, δ /δ = 0.5, 1 c 2 c σ /E = 0.003, ν = 0.3, and N = 0.1. . . . . . . . . . . . . . . 19 7 Steady state toughness as a function of the cohesive strength for isotropic and kinematic hardening at different N levels. Material properties: δ /δ = 0.15, δ /δ = 0.5, σ /E = 0.003, 1 c 2 c 0 and ν = 0.3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 8 Crack extension at steady state as a function of the cohesive strength for isotropic and kinematic hardening at different N levels. Material properties: δ /δ = 0.15, δ /δ = 0.5, σ /E = 1 c 2 c 0 0.003, and ν = 0.3. . . . . . . . . . . . . . . . . . . . . . . . . 21 9 Log-log representation of the relation between the crack exten- sion at steady state and the steady state fracture toughness for isotropic and kinematic hardening at different N levels. Each data point corresponds to a value of the cohesive strength. Material properties: δ /δ = 0.15, δ /δ = 0.5, σ /E = 0.003, 1 c 2 c 0 and ν = 0.3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 10 Schematic insight into the effect of isotropic or kinematic hard- ening on a material point ahead of the initial crack (r = 2R ); (a) active plastic zone and evolution path, and (b) stress state on the π-plane. Material properties: δ /δ = 0.15, δ /δ = 0.5, 1 c 2 c σ /E = 0.003, ν = 0.3, σˆ/σ = 3.7, and N = 0.1. . . . . . . . 23 0 Y 13 3 Kinematic Isotropic -1 -2 -3 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 (a) 3.5 2.5 1.5 0.5 0 0.05 0.1 0.15 (b) Figure 1: Uniaxial stress strain response for (a) cyclic loading of a non-linear hardening solid with N = 0.2, and (b) half-cycle for linear and non-linear hardening. Material properties: σ /E = 0.003. 14 K Continuum Crack Plastic zone Figure 2: Schematic representation of the cohesive zone model for fracture. 15 Figure 3: Cohesive traction T - separation δ law characterising the fracture process. 16 Figure 4: Finite element mesh and configuration of the boundary layer. 17 5 Isotropic hardening 4.5 Kinematic hardening 3.5 2.5 1.5 0 0.5 1 1.5 2 2.5 3 Figure 5: Crack growth resistance curves for linear isotropic and kinematic hardening plasticity and different hardening levels. Material properties: δ /δ = 0.15, δ /δ = 0.5, 1 c 2 c σ /E = 0.003, ν = 0.3, and σ ˆ = 3.5σ . 0 0 18 5 4.5 3.5 2.5 Isotropic hardening 1.5 Kinematic hardening 0 0.5 1 1.5 2 2.5 3 Figure 6: Crack growth resistance curves for power law isotropic and kinematic hardening plasticity and different levels of the cohesive strength. Material properties: δ /δ = 0.15, 1 c δ /δ = 0.5, σ /E = 0.003, ν = 0.3, and N = 0.1. 2 c 0 19 6 Isotropic hardening Kinematic hardening 2 2.5 3 3.5 4 4.5 5 Figure 7: Steady state toughness as a function of the cohesive strength for isotropic and kinematic hardening at different N levels. Material properties: δ /δ = 0.15, δ /δ = 0.5, 1 c 2 c σ /E = 0.003, and ν = 0.3. 20 50 3 3.2 3.4 3.6 3.8 Figure 8: Crack extension at steady state as a function of the cohesive strength for isotropic and kinematic hardening at different N levels. Material properties: δ /δ = 0.15, δ /δ = 1 c 2 c 0.5, σ /E = 0.003, and ν = 0.3. 21 1 -1 0 1 10 10 Figure 9: Log-log representation of the relation between the crack extension at steady state and the steady state fracture toughness for isotropic and kinematic hardening at different N levels. Each data point corresponds to a value of the cohesive strength. Material properties: δ /δ = 0.15, δ /δ = 0.5, σ /E = 0.003, and ν = 0.3. 1 c 2 c 0 22 (a) a = a = 0 2R (b) = 2R = 3R Initial yield surface Isotropic Kinematic Figure 10: Schematic insight into the effect of isotropic or kinematic hardening on a material point ahead of the initial crack (r = 2R ); (a) active plastic zone and evolution path, and (b) stress state on the π-plane. Material properties: δ /δ = 0.15, δ /δ = 0.5, 1 c 2 c σ /E = 0.003, ν = 0.3, σˆ/σ = 3.7, and N = 0.1. 0 Y http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Condensed Matter arXiv (Cornell University)

Crack growth resistance in metallic alloys: the role of isotropic versus kinematic hardening

Martínez-Pañeda, Emilio; Fleck, Norman A.
Condensed Matter , Volume 2018 (1806) – Jun 23, 2018
Download PDF
23 pages

Loading next page...
 
/lp/arxiv-cornell-university/crack-growth-resistance-in-metallic-alloys-the-role-of-isotropic-Sqn4uAbE0h
ISSN
0021-8936
eISSN
ARCH-3331
DOI
10.1115/1.4040696
Publisher site
See Article on Publisher Site

Abstract

The sensitivity of crack growth resistance to the choice of isotropic or kine- matic hardening is investigated. Monotonic mode I crack advance under small scale yielding conditions is modelled via a cohesive zone formulation endowed with a traction-separation law. R-curves are computed for materials that exhibit linear or power law hardening. Kinematic hardening leads to an enhanced crack growth resistance relative to isotropic hardening. Moreover, kinematic hardening requires greater crack extension to achieve the steady state. These differences are traced to the non-proportional loading of mate- rial elements near the crack tip as the crack advances. The sensitivity of the R-curve to the cohesive zone properties and to the level of material strain hardening is explored for both isotropic and kinematic hardening. Keywords: Kinematic hardening, Isotropic hardening, Cohesive zone modelling, Finite element analysis, Fracture 1. Introduction It is well established that material elements near a mode I crack tip undergo non-proportional straining due to crack advance, see for example the early analysis of crack growth by Rice and Sorensen [1]. The degree of hysteresis associated with this non-proportional loading is sensitive to the nature of the hardening law of the solid. For example, it is to be expected that kinematic hardening leads to greater hysteresis than isotropic hardening. Corresponding author. Tel: +44 1223 748525; fax: +44 1223 332662. Email address: mail@empaneda.com (Emilio Mart´ınez-Pan ˜eda) Preprint submitted to Journal of Applied Mechanics June 26, 2018 arXiv:1806.08986v1 [cond-mat.mtrl-sci] 23 Jun 2018 Consequently, one might expect that the choice of plastic hardening law will influence the stress intensity factor K versus crack extension Δa response, widely known as the R-curve. However, little attention has been paid to the effect of the hardening law upon crack growth resistance and no clear picture emerges from the literature. Lam and McMeeking [2] analysed steady state crack tip fields and concluded from a crack opening displacement criterion that isotropic hardening augments crack growth resistance. Carpinteri [3] performed finite element analyses of crack propagation by means of a strain- based criterion and observed a greater amount of crack extension in the kinematic hardening case for a given remote load; this also suggests that isotropic hardening increases crack growth resistance. In contrast, we shall demonstrate that kinematic hardening significantly raises the level of plastic dissipation and, thereby, elevates the R-curve along with the steady state fracture toughness K . SS 2. Numerical model We consider the small scale yielding problem of a plane strain mode I crack subjected to a remote stress intensity factor K. The elasto-plastic solid is isotropic with a Young’s modulus E, a Poisson’s ratio ν and an initial yield strength σ . Throughout our study we shall take ν = 0.3 and σ /E = 0.003. 0 0 We denote the Cauchy stress by σ and define s as its deviatoric part, such ij ij that s = σ − δ σ /3. The plastic response involves either isotropic or ij ij ij kk kinematic hardening, as follows. Isotropic hardening: The yield condition reads Φ = σ − σ = 0 (1) e Y where σ is the von Mises effective stress and the current yield strength σ is e Y a function of the accumulated von Mises plastic strain ε . Accordingly, the effective stress σ in J2 plasticity is defined as σ = s s (2) ij ij The increment in plastic strain ε˙ is computed from the normality hy- ij pothesis, ∂Φ 3 s p ij ε˙ = ε˙ = ε˙ (3) e e ij ∂σ 2 σ ij Y 2 in terms of the increment in effective plastic strain ε˙ . The relation between σ and ε is given by the uniaxial tensile response, such that the true tensile Y e stress σ is related to the true tensile plastic strain ε by Eε σ = σ 1 + (4) where N is the strain hardening exponent. In addition to this power law description we also consider the case of linear hardening by taking N = 1 and by replacing E in (4) with the tangent modulus E . Kinematic hardening: Assume that the centre of the yield surface is located at the point α in deviatoric stress space. We shall refer to α as a back- ij ij stress, and assume that the Armstrong-Frederick non-linear rule [4] defines the evolution of this backstress, such that (s − α ) ij ij α˙ = c ε˙ − γα ε˙ (5) ij e ij e where c and γ are material constants. This rule reproduces ratchetting when a material element is subjected to a non-zero mean stress and cyclic loading but predicts a particular shape of the stress-strain curve (see for example [5]). In order to model a more general shape of the uniaxial tensile response the constitutive statement (5) has been extended by Chaboche [6]. He replaced the single backstress α by a finite number n of backstresses α , such that ij ij α = α (6) ij ij k=1 Each backstress α evolves with ε˙ according to the independent hardening ij ij rule, (σ − α ) ij ij k k k k α˙ = c ε˙ − γ α ε˙ (no sum on k) (7) e e ij ij k k in terms of the material constants c and γ . The resulting Chaboche- Armstrong-Frederick (CAF) model has been widely used to capture ratchet- ting effects and non-linear hardening under non-proportional cyclic loading [7, 8]. We consider here the case of power law hardening (4) and select the k k values of c , γ for k ∈ (1, n) such that the desired response in unaxial ten- sion is obtained. The choice of n = 10 brings the CAF model into alignment 3 with (4) to within 0.04% for the range of true tensile strain 0 ≤ ε ≤ 2.0. The uniaxial stress-strain response for cyclic loading is given in Fig. 1a for the case N = 0.2. In the case of linear hardening (5) is used instead of (6)-(7), with γ = 0 and c = E ; this is the familiar Ziegler formulation [9]. The hardening laws employed are shown in Fig. 1b; power law hardening for the choices N = 0.1 and N = 0.2, and linear hardening for E /σ = 5/3 and t 0 E /σ = 50/3. t 0 We model tensile fracture at the tip of a mode I crack by means of a cohesive zone model, following Tvergaard and Hutchinson [10] - see Fig. 2. Cohesive zone formulations have a long history back to Dugdale [11] and Barenblatt [12]: fracture is regarded as a gradual process in which separa- tion takes place across an extended cohesive zone, and is resisted by cohesive tractions. As shown in Fig. 3, we shall make use of a trapezoidal traction- separation law of strength σˆ, with its shape being defined by a critical cohe- sive separation δ and by two shape parameters δ = 0.15δ and δ = 0.5δ . c 1 c 2 c The work of fracture Γ is given by the area under the traction-separation curve, such that Γ = σˆ (δ + δ − δ ) (8) 0 c 2 1 A reference stress intensity factor for crack growth initiation follows im- mediately as EΓ K = (9) (1 − ν ) along with a reference length R , where 1 EΓ 1 K 0 0 R = = (10) 3π (1 − ν ) σ 3π σ The crack tip is placed at the origin and the crack plane is aligned with the negative x axis of the Cartesian reference frame (x , x ). A remote 1 1 2 K field is imposed by a boundary layer formulation, as follows. The outer periphery of the mesh is subjected to the mode I elastic K-field, 1/2 u = r f (θ, ν) (11) i i 4 where r and θ are polar coordinates centred at the crack tip and the functions f (θ, ν) are given by, 1 + ν θ f = √ (3 − 4ν − cos θ) cos (12) 2π 1 + ν f = (3 − 4ν − cos θ) sin (13) 2π Upon exploiting the symmetry about the crack plane, only half of the model is analysed, as shown in Fig. 4. The finite element model is imple- mented in the commercial finite element package Abaqus [13] and we solve the boundary value problem by an implicit Backward Euler integration scheme. Plane strain quadratic quadrilateral elements are employed, with the mesh comprising 267272 degrees of freedom. A refined mesh was used along the co- hesive zone in order to obtain a converged solution. The characteristic length of the elements in the crack propagation region is chosen to be equal to δ . Cohesive elements with 6 nodes and 12 integration points are implemented by means of a user element (UEL) subroutine [14]. A control algorithm is used to avoid convergence problems due to snap-back instabilities, see [15, 16] for details. Computations have been performed within an infinitesimal de- formation framework since strains remain small, as argued by Tvergaard and Hutchinson [10] in their finite strain analysis. Dimensional analysis shows that the solution, given in terms of the remote K = K (Δa), is a function I R F of the following dimensionless quantities, K Δa σ σˆ E δ δ R 0 t 1 2 = F , , , , ν, N, , (14) K R E σ E δ δ 0 0 0 c c We will conduct calculations until steady state crack growth at constant K is attained. Tvergaard and Hutchinson [10] showed that K is sensitive SS SS to the ratio of cohesive strength σˆ to material yield strength σ . For an elas- tic, perfectly plastic solid, K /K raises steeply as σˆ/σ approaches 3. The SS 0 0 interpretation is straightforward by considering a stationary crack in an elas- tic, ideally plastic solid absent of a cohesive zone. The tensile stress directly ahead of the crack tip equals 3σ as given by the Prandtl field. Consequently, if σˆ/σ exceeds 3, the crack tip blunts without advance as the cohesive zone strength is not overcome. 5 Tvergaard and Hutchinson [10] also considered the role of isotropic strain hardening on the R-curve. In this case, the stress field ahead of the crack tip exceeds 3σ due to the presence of strain hardening. For σˆ/σ < 3, a shallow 0 0 R-curve is exhibited and K /K is slightly above unity. In contrast, for SS 0 σˆ/σ > 3, a steeper R-curve is observed and K /K increases its sensitivity 0 SS 0 to σˆ. A major aim of the present study is to explore the sensitivity of the R-curve to the nature of the hardening law: isotropic versus kinematic hardening. 3. Results R-curves are shown in Fig. 5 for linear hardening and σˆ/σ = 3.5; this value of σˆ/σ is close to the limiting value of σˆ/σ = 3 for an elastic, per- 0 0 fectly plastic solid, as discussed in the previous section. Consider first the R-curve for a small level of strain hardening E /σ = 5/3. A steeply ris- t 0 ing R-curve is predicted, which will give rise to a large steady state fracture toughness K /K and a large value of the crack extension to achieve steady SS 0 state (Δa/R ) . The steep R-curve is a consequence of plastic dissipation SS with crack advance. Little difference is observed between the kinematic and isotropic hardening predictions since the degree of hardening is small. We note in passing that for the elastic, ideally plastic case, E /σ = 0, and t 0 σˆ/σ = 3.5 no crack advanced is observed: the tensile traction ahead of the crack tip is insufficient to overcome the cohesive zone strength. Now consider the case of a high strain hardening rate E /σ = 50/3. The strain level near t 0 the crack tip can now exceed the cohesive strength σˆ at a relatively low value of plastic strain. A shallow R-curve is predicted. Again, kinematic hardening elevates the R-curve compared to the isotropic hardening case. The predicted R-curves for the case of a power law hardening solid are shown in Fig. 6 for the choice of N = 0.1. We consider both isotropic and kinematic hardening, and selected values of the cohesive strength σˆ/σ = 3.2, 3.4 and 3.5. As expected, increasing σˆ/σ elevates the K versus Δa response for both hardening laws. However, the R-curves are more sensitive to the cohesive strength for the case of kinematic hardening. Also, as in the linear hardening study, the R-curves are steeper for kinematic hardening, implying a higher value of the steady state fracture toughness K . We note that SS there is no straightforward relationship between K , K and the crack ini- 0 SS tiation toughness K as defined in the standard test methods for fracture Ic 6 toughness, such as the ASTM E 1820 [17]. The standard defines K as the Ic value of K corresponding to a crack growth increment which is in the range of 0.2 - 0.5 mm, see [18] for a discussion. When the R-curve is steep, K /K Ic 0 may be large. The steady state toughness is attained when K reaches a plateau value. Fig. 7 shows the sensitivity of K to the strain hardening exponent and SS to the cohesive strength. The isotropic curves are in agreement with the results of Tvergaard and Hutchinson [10]. First, note that for N = 0 a single K /K versus σˆ/σ curve corresponds to the cases of isotropic and kine- SS 0 0 matic hardening. The value of the steady state toughness increases rapidly in the vicinity of the limiting value of σˆ/σ = 2.8. With increasing N, kinematic and isotropic hardening theories give increasingly divergent pre- dictions. Consistently, for N > 0, kinematic hardening leads to a higher value of K /K at a given σˆ/σ than does isotropic hardening. Also, the SS 0 0 value of the cohesive strength at which K /K increases rapidly is lower for SS 0 the case of kinematic hardening. It is instructive to consider the value of crack extension (Δa) that is SS required to achieve the steady state toughness. The dependence of (Δa) SS upon σˆ/σ is shown in Fig. 8 for the power law hardening solid, for both kinematic and isotropic hardening. Note that both K /K and (Δa) /R SS 0 0 SS depend upon σˆ/σ in a highly non-linear manner for both hardening laws, recall Figs. 7 and 8. Is there a simple relation between (Δa) /R and SS K /K ? This might be expected as the plastic zone size associated with SS 0 K = K , is of the order, SS SS R = (15) SS 3π σ Assume that (Δa) is proportional to R , SS SS (Δa) = CR (16) SS SS where the constant C is of order unity with some sensitivity to the choice of the hardening law and to N. It follows immediately that, (Δa) R K SS SS SS = C = C (17) R R K 0 0 0 7 The accuracy of this prediction is shown by a cross-plot of (Δa) /R SS versus K /K in Fig. 9, with σˆ/σ as the parameter trending variable. A SS 0 0 curve fit reveals that C increases from 0.574 to 1.703 as N goes from 0.1 to 0.2 for kinematic hardening, and C increases from 0.31 to 0.496 as N goes from 0.1 to 0.2 for isotropic hardening. Our numerical predictions show that the isotropic hardening idealization may significantly underestimate the degree of subcritical crack propagation before catastrophic failure. Note further that (17) can be re-expressed in the form SS Δa = (18) SS 3π σ What is the physical basis for the steeper R-curve observed in the case of kinematic hardening? We show in Fig. 10 that significant non-proportional loading occurs in the vicinity of the crack tip, as the crack advances. Consider a representative material point P at a distance for 2R ahead of the initial crack tip and slightly above the cracking plane (height of 0.1R ). Allow the crack to advance by Δa = 2R for both cases of isotropic and kinematic hardening. The active plastic zone is shown in Fig. 10a for Δa = 0 and Δa = 2R . The plastic zone at Δa = 0 for kinematic hardening is identical to that for isotropic hardening, whereas the plastic zone at Δa = 2R is much larger in the kinematic hardening case. Only the isotropic hardening active plastic zone is shown at Δa = 2R for the sake of clarity. The stress paths imposed on point P for isotropic and kinematic hardening are given in Fig. 10b. Differences between kinematic and isotropic stress paths arise soon after cracking initiates (Δa = 0 ), due to non-proportional straining in the neighbouring points. As the crack advances not only are the stress paths non-proportional but they also deviate from each other, with the greatest change in stress direction given by isotropic hardening. The stronger path dependence of kinematic hardening also plays an important role on localiza- tion in thin sheets and in shear localization. For example, Tvergaard [19] showed that the forming limit curves predicted by kinematic hardening are in better agreement with experimental results than isotropic hardening pre- dictions. The dependence of the critical strain for shear localization upon the local curvature of the yield surface has been investigated by Mear and Hutchinson [20]. In addition, we investigate the level of energy dissipation in the main plastic zone in the vicinity of the crack tip (denoted W ) and in the secondary 8 plastic zone that arises in the crack wake (denoted as W ). Here, the energy dissipated due to plastic deformation is computed for all material elements in the active plastic zone as, Z Z Δa W (Δa) = σ ε˙ dV da (19) ij ij The predictions are given in Table 1 for various stages of crack advance. Computations reveal that plastic work in the secondary plastic region is neg- ligible relative to the energy dissipated in the vicinity of the crack. We con- clude that differences between isotropic and kinematic hardening responses are mainly due to non-proportional deformation in the crack tip plastic zone. Also, Table 1 shows that kinematic hardening involves a much larger plas- tic dissipation energy than isotropic hardening, and is consistent with the steeper R-curve. Table 1: Plastic energy dissipation with crack advance in the primary plastic zone region at the crack tip W and the secondary plastic zone region at the crack wake W . Material 1 2 properties: δ /δ = 0.15, δ /δ = 0.5, σ /E = 0.003, ν = 0.3, N = 0.1 and σˆ = 3.5σ . 1 c 2 c 0 0 Isotropic Kinematic Δa/R W / (Γ Δa) W / (Γ Δa) W / (Γ Δa) W / (Γ Δa) 1 0 2 0 1 0 2 0 0.5 21.98 0.01 52.69 0.42 1 17.24 0.14 58.36 0.85 2 8.21 0.17 78.75 1.08 4. Conclusions We investigated how the isotropic or kinematic nature of strain hard- ening influences crack growth resistance. Finite element results show very significant differences between isotropic and kinematic hardening laws that yield the same response under uniaxial tension. We show that kinematic hardening notably enhances plastic dissipation and the steady state frac- ture toughness K . These differences persist over different hardening levels, SS cohesive strengths, and hardening profiles. 9 5. Acknowledgments The authors acknowledge valuable discussions with V.S. Deshpande (Uni- versity of Cambridge). E. Mart´ınez-Pan˜eda also acknowledges valuable in- sight from K. Juul and C.F. Niordson (Technical University of Denmark). The authors would like to acknowledge the funding and technical support from BP (ICAM02ex) through the BP International Centre for Advanced Ma- terials (BP-ICAM). E. Mart´ınez-Pan˜eda additionally acknowledges financial support from the Ministry of Economy and Competitiveness of Spain through grant MAT2014-58738-C3 and the People Programme (Marie Curie Actions) of the European Union’s Seventh Framework Programme (FP7/2007-2013) under REA grant agreement n 609405 (COFUNDPostdocDTU). References [1] J. R. Rice, E. P. Sorensen, Continuing crack-tip deformation and frac- ture for plane-strain crack growth in elastic-plastic solids, Journal of the Mechanics and Physics of Solids 26 (3) (1978) 163–186. [2] P. S. Lam, R. M. McMeeking, Analyses of steady quasistatic crack growth in plane strain tension in elastic-plastic materials with non- isotropic hardening, Journal of the Mechanics and Physics of Solids 32 (5) (1984) 395–414. [3] A. Carpinteri, Crack growth resistance in non-perfect plasticity: Isotropic versus kinematic hardening, Theoretical and Applied Fracture Mechanics 4 (2) (1985) 117–122. [4] C. O. Frederick, P. J. Armstrong, A mathematical representation of the multiaxial Bauschinger effect, Materials at High Temperatures 24 (1) (2007) 1–26. [5] A. F. Bower, Applied Mechanics of Solids, CRC Press, Taylor & Francis, Boca Raton, 2009. [6] J. L. Chaboche, On some modifications of kinematic hardening to im- prove the description of ratchetting effects, International Journal of Plas- ticity 7 (7) (1991) 661–678. [7] J. Lemaitre, J.-L. Chaboche, Mechanics of solid materials, Vol. 19, Cam- bridge University Press, New York, 1990. 10 [8] J. L. Chaboche, A review of some plasticity and viscoplasticity consti- tutive theories, International Journal of Plasticity 24 (10) (2008) 1642– [9] H. Ziegler, A Modification of Prager’s hardening rule, Quarterly of Ap- plied Mathematics 17 (1) (1959) 55–65. [10] V. Tvergaard, J. W. Hutchinson, The relation between crack growth re- sistance and fracture process parameters in elastic-plastic solids, Journal of the Mechanics and Physics of Solids 40 (6) (1992) 1377–1397. [11] D. S. Dugdale, Yielding of steel sheets containing slits, Journal of the Mechanics and Physics of Solids 8 (2) (1960) 100–104. [12] G. I. Barenblatt, The Mathematical Theory of Equilibrium of Crack in Brittle Fracture, Advances in Applied Mechanics 7 (1962) 55–129. [13] Abaqus/Standard 2017. Dassault Systemes SIMULIA, Providence, RI, USA. [14] S. del Busto, C. Betego´n, E. Mart´ınez-Pan˜eda, A cohesive zone frame- work for environmentally assisted fatigue, Engineering Fracture Mechan- ics 185 (2017) 210–226. [15] J. Segurado, J. LLorca, A new three-dimensional interface finite element to simulate fracture in composites, International Journal of Solids and Structures 41 (11-12) (2004) 2977–2993. [16] E. Mart´ınez-Pan˜eda, S. del Busto, C. Betego´n, Non-local plasticity ef- fects on notch fracture mechanics, Theoretical and Applied Fracture Mechanics 92 (2017) 276–287. [17] ASTM E 1820-01, Standard Test Method for Measurement of Fracture Toughness. [18] T. L. Anderson, Fracture Mechanics. Fundamentals and Applications, 3rd Edition, CRC Press, Taylor & Francis, Boca Raton, 2005. [19] V. Tvergaard, Effect of kinematic hardening on localized necking in biaxially stretched sheets, International Journal of Mechanical Sciences 20 (9) (1978) 651–658. 11 [20] M. E. Mear, J. W. Hutchinson, Influence of yield surface curvature on flow localization in dilatant plasticity, Mechanics of Materials 4 (1985) 395–407. 12 List of Figures 1 Uniaxial stress strain response for (a) cyclic loading of a non- linear hardening solid with N = 0.2, and (b) half-cycle for linear and non-linear hardening. Material properties: σ /E = 0.003. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2 Schematic representation of the cohesive zone model for fracture. 15 3 Cohesive traction T - separation δ law characterising the frac- ture process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 4 Finite element mesh and configuration of the boundary layer. . 17 5 Crack growth resistance curves for linear isotropic and kine- matic hardening plasticity and different hardening levels. Ma- terial properties: δ /δ = 0.15, δ /δ = 0.5, σ /E = 0.003, 1 c 2 c 0 ν = 0.3, and σˆ = 3.5σ . . . . . . . . . . . . . . . . . . . . . . . 18 6 Crack growth resistance curves for power law isotropic and kinematic hardening plasticity and different levels of the cohe- sive strength. Material properties: δ /δ = 0.15, δ /δ = 0.5, 1 c 2 c σ /E = 0.003, ν = 0.3, and N = 0.1. . . . . . . . . . . . . . . 19 7 Steady state toughness as a function of the cohesive strength for isotropic and kinematic hardening at different N levels. Material properties: δ /δ = 0.15, δ /δ = 0.5, σ /E = 0.003, 1 c 2 c 0 and ν = 0.3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 8 Crack extension at steady state as a function of the cohesive strength for isotropic and kinematic hardening at different N levels. Material properties: δ /δ = 0.15, δ /δ = 0.5, σ /E = 1 c 2 c 0 0.003, and ν = 0.3. . . . . . . . . . . . . . . . . . . . . . . . . 21 9 Log-log representation of the relation between the crack exten- sion at steady state and the steady state fracture toughness for isotropic and kinematic hardening at different N levels. Each data point corresponds to a value of the cohesive strength. Material properties: δ /δ = 0.15, δ /δ = 0.5, σ /E = 0.003, 1 c 2 c 0 and ν = 0.3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 10 Schematic insight into the effect of isotropic or kinematic hard- ening on a material point ahead of the initial crack (r = 2R ); (a) active plastic zone and evolution path, and (b) stress state on the π-plane. Material properties: δ /δ = 0.15, δ /δ = 0.5, 1 c 2 c σ /E = 0.003, ν = 0.3, σˆ/σ = 3.7, and N = 0.1. . . . . . . . 23 0 Y 13 3 Kinematic Isotropic -1 -2 -3 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 (a) 3.5 2.5 1.5 0.5 0 0.05 0.1 0.15 (b) Figure 1: Uniaxial stress strain response for (a) cyclic loading of a non-linear hardening solid with N = 0.2, and (b) half-cycle for linear and non-linear hardening. Material properties: σ /E = 0.003. 14 K Continuum Crack Plastic zone Figure 2: Schematic representation of the cohesive zone model for fracture. 15 Figure 3: Cohesive traction T - separation δ law characterising the fracture process. 16 Figure 4: Finite element mesh and configuration of the boundary layer. 17 5 Isotropic hardening 4.5 Kinematic hardening 3.5 2.5 1.5 0 0.5 1 1.5 2 2.5 3 Figure 5: Crack growth resistance curves for linear isotropic and kinematic hardening plasticity and different hardening levels. Material properties: δ /δ = 0.15, δ /δ = 0.5, 1 c 2 c σ /E = 0.003, ν = 0.3, and σ ˆ = 3.5σ . 0 0 18 5 4.5 3.5 2.5 Isotropic hardening 1.5 Kinematic hardening 0 0.5 1 1.5 2 2.5 3 Figure 6: Crack growth resistance curves for power law isotropic and kinematic hardening plasticity and different levels of the cohesive strength. Material properties: δ /δ = 0.15, 1 c δ /δ = 0.5, σ /E = 0.003, ν = 0.3, and N = 0.1. 2 c 0 19 6 Isotropic hardening Kinematic hardening 2 2.5 3 3.5 4 4.5 5 Figure 7: Steady state toughness as a function of the cohesive strength for isotropic and kinematic hardening at different N levels. Material properties: δ /δ = 0.15, δ /δ = 0.5, 1 c 2 c σ /E = 0.003, and ν = 0.3. 20 50 3 3.2 3.4 3.6 3.8 Figure 8: Crack extension at steady state as a function of the cohesive strength for isotropic and kinematic hardening at different N levels. Material properties: δ /δ = 0.15, δ /δ = 1 c 2 c 0.5, σ /E = 0.003, and ν = 0.3. 21 1 -1 0 1 10 10 Figure 9: Log-log representation of the relation between the crack extension at steady state and the steady state fracture toughness for isotropic and kinematic hardening at different N levels. Each data point corresponds to a value of the cohesive strength. Material properties: δ /δ = 0.15, δ /δ = 0.5, σ /E = 0.003, and ν = 0.3. 1 c 2 c 0 22 (a) a = a = 0 2R (b) = 2R = 3R Initial yield surface Isotropic Kinematic Figure 10: Schematic insight into the effect of isotropic or kinematic hardening on a material point ahead of the initial crack (r = 2R ); (a) active plastic zone and evolution path, and (b) stress state on the π-plane. Material properties: δ /δ = 0.15, δ /δ = 0.5, 1 c 2 c σ /E = 0.003, ν = 0.3, σˆ/σ = 3.7, and N = 0.1. 0 Y

Journal

Condensed Matter – arXiv (Cornell University)

Published: Jun 23, 2018

References

  • Continuing Crack-Tip Deformation and Fracture for Plane-Strain Crack Growth in Elastic-Plastic Solids
  • Analyses of Steady Quasistatic Crack Growth in Plane Strain Tension in Elastic-Plastic Materials With Non-Isotropic Hardening
  • Crack Growth Resistance in Non-Perfect Plasticity: Isotropic Versus Kinematic Hardening
  • A Mathematical Representation of the Multiaxial Bauschinger Effect