Search

Slow crack growth during SCC of 7075 aluminum has been shown to comprise both an electrochemical and a mechanical component. These findings prompted a review of several possible mechanical models, and seven possible controlling thermally-activated processes. Since no existing interpretation could satisfy all of the observations, an empirical model was developed. The conclusion is that slight modification of many existing proposed mechanisms could explain the general features of SCC but that any theoretical model must contain some aspect of the mechanical rupture process.

A detailed study of rapid stress-corrosion-cracking (SCC) in a 7075 aluminum alloy has allowed separation of the mechanical and chemical contributions. This was accomplished by combining scanning electron microscopy, stress-wave emission and crack growth rate observations as a function of test temperature. These established an activation energy of 11.2 kcal/mol, a stress-intensity squared dependence of crack growth, and a range of 20 to 80 pct dimpled rupture on the fracture surfaces. Thus a two-step crack growth mechanism is proposed combining a thermally activated electrochemical process and a discontinuous mechanical jumping process.

The fatigue crack propagation behavior of a class of metastable austenitic steels called TRIP steels has been investigated. The alloy composition was chosen to have theMs well below room temperature and theMD above room temperature after thermomechanical processing. A simple theoretical model of fatigue crack propagation (FCP) based on fracture mechanics was developed. Fatigue crack propagation tests on SEN specimens at various stress intensity ranges (ΔK) were carried out, and two stage plastic-carbon replica were used to observe the fracture surface of the FCP specimens. To a first approximation, both the experimental and theoretical results followed the usual relationship between ΔK and FCP rates;i.e. da/dn ∝ (ΔK).4 The fatigue fracture surface contained fatigue striations, quasicleavage and elongated dimples; a reflection of the complex structure of TRIP steels. A beneficial effect of strain induced martensite transformation with regards to fatigue crack propagation was found. TRIP steels showed better FCP properties than a number of alloy steels of similar strength levels and compared favorably with mar aging steels in the low ΔK range.

A technique for measuring stress waves emitted during slow crack growth is briefly described. The number and size of the stress waves generated appears to bear a unique relationship to the amount of crack growth. From elasticity theory, a semi-empirical relationship is developed. Δ ≈ (Σg)2 E/K2, where ΔA is the incremental area swept out by the crack, Σg, is the sum of the stress-wave amplitudes associated with that increment of growth, E is the elastic modulus and K is the applied stress-intensity factor. This relationship is shown to be valid for about 100 data points obtained from crack growth in high-strength aluminum, titanium and steel alloys.

Several fracture criteria are reviewed with respect to ductile fracture. It is suggested that both critical crack-tip displacement, 2Vc*, and critical fracture strain,∃*, criteria may describe the fracture of a ductile second phase rod in a ductile matrix. As a first approximation, this is experimentally verified by observations of ductile stainless steel fibres fracturing in an age-hardened aluminium matrix. For 0.05, 0.10 and 0.20 volume fraction composites, the average fracture strains are calculated to be 1.15 as compared to a measured average of 0.93 while the average critical crack-tip displacement is calculated to be 0.50 mm as compared to an “observed” average of 0.40 mm. The statistical variation in the fracture strain was not sufficiently small to allow any choice between these proposed criteria. In fact, both the experimental and theoretical evidence point to the equivalency of these criteria as given by 2Vc*=π/*∃* where /* is the microstructural unit in front of the crack over which the strain is greater than or equal to∃*.

The complex material behavior associated with multiple pop-in phenomena has been treated with some degree of success. For pop-in length, the analysis considers an effective load reduction at the crack front due to plastic deformation. The effective stress intensity factor is decreased by the plastic deformation while it is increased by the length of pop-in, l . Due to these competing effects, the stress intensity, K , first increases and then decreases with increasing l . The point at which the effective K reduces back to the initial K defines l . For a second or third pop-in, the stress intensity increment, ΔK , between pop-ins is interpreted in terms of a changing fracture criterion (plane strain to plane stress) and the crack-tip blunting that occurs with increasing stress intensity. Although semi-quantitative in nature, these analyses do predict the experimentally observed increase in l and decrease in ΔK associated with increasing K . The limits of these models provide additional insight into fracture behavior. (a) As l → 0 : A geometrical lower bound for pop-in behavior is hypothesized; (b) As ΔK → 0 : The critical stress intensity factor at any particular thickness is given thus allowing establishment of the fracture-mode transition curve.

A technique is presented which allows separation of the principal plastic strains within an enclave at a crack tip in relatively thin plates. For plates with a number of crack lengths, several degrees of tensile loading are applied to five ductile materials with widely varying mechanical properties. Principal plastic strain distributions, plastic distortional strain energy density and total plastic energy are determined within the resulting plastic encalves. It is shown that the degree of loading and strain hardening greatly affect the principal plastic strain and energy density distributions and that this is reflected in the amount of strain energy that can be plastically absorbed at a crack tip.

An experimental stress-analysis technique using a birefringent coating is reported for determining the stress distribution about a slowly growing crack. The maximum error of the test method for a large strain gradient is found to be less than 10 percent. For a plate with an internal crack, the experimentally determined stress distribution compares favorably with two numerical solutions. Comparison of stresses about an internal or double-edge crack to those about a single-edge crack indicates that the isochromatics bend over to about 45 deg with the plane of the crack in the former and are inclined at about 60 deg in the latter. Also, the stresses for a single-crack tip vary as the inverse square root of the radius, while the stresses for a double-crack tip follow anr−1/4 law more closely.

F racture mechanics of a metal-matrix composite containing ductile-metallic fibers is described. Experimental verification of the proposed descriptions has been established with an aluminum-base composite containing uni-directional stainless-steel fibers. Theoretical description of the plastic energy dissipation shows that crack propagation across fibers is very difficult because of the high energy density represented by the fiber. The fiber contribution is an increasing function of volume fraction which results in the critical stress-intensity factor increasing with fiber content. On the other hand, crack propagation between fibers is very easy because the inter-fiber spacing limits the plastic energy dissipation in the matrix. The critical stress intensity for crack propagation between fibers is a decreasing function of volume fraction.

External test variables such as rate and temperature, and changes in alloy composition are shown to have a number of effects on the fracture of high-strength, metastable austenitic steels. One rate-dependent phenomenon is an unusual fracture mode transition wherein a flat mode changes to a shear mode when the amount of transformation product in the vicinity of the crack tip is reduced by adiabatic heating. The point at which this happens in any one test is dependent upon the velocity of the slowly growing crack which in turn is dependent upon the crosshead rate. Because of this rate effect, the plane stress fracture toughness decreases by as much as 30 pct at higher crosshead rates. Fractographically, it was ascertained that at room temperature, both phases failed in a ductile manner, but at −196°C, martensite containing greater than about 0.27 wt pct C would cleave. This resulted in a “ductile-brittle” transition in metastable austenites at −196°C as a function of carbon content. Other compositional variations change the austenite stability which controls the amount of strain-induced marteniste occurring at the crack tip. It is shown that a plane stress fracture toughness (KC) approaching 500,000 psi-in.1/2 may be achieved by decreasing the stability of the austenite. The variation ofKc with austenite stability agrees qualitatively with a theoretical model for the invariant shear contribution to the fracture toughness of metastable austenites.

Most high strength steel and titanium alloys have been observed to have a threshold stress intensity, below which no environmentally-induced cracking occurs. Recent models for the concentration of atomic hydrogen in the vicinity of a dilatational stress field can be used to predict such threshold stress intensities. These models can predict state of stress effects which can be changed by yield strength, thickness and loading mode. The following analysis demonstrates how an externally applied hydrostatic pressure should increase the threshold stress intensity. A predicted increase of 5000 psi-in 1 2 in an alpha-beta titanium alloy is in reasonable agreement with experimental observations of Dehart and Liebowitz.

The fatigue of welded tubular joints, as found in offshore structures, is explored by developing a qualitative model. This model considers the influence exerted on the initiation and growth of fatigue cracks by secondary effects of the joints fabrication and behavior. The model is then used, along with basic concepts from fracture mechanics, to develop an analytical model for the prediction of fatigue failure. The behavior of the tubular joint undergoing fatigue is divided into three phases: primary phase—no noticeable effect from fatigue; transition phase—decaying behavior; and secondary phase—distinctly new behavior. Fatigue failure is then defined as a distinct change in the joint’s load transfer characteristics (the end of the transition phase). Experimental data is presented to support both this definition and the theoretical prediction of failure.