The effect of the stress ratio on fatigue crack growth in a 3% NaCl solution

Corrosion fatigue crack growth tests have been carried out at various stress ratios for a low alloy steel SNCM 2 and type 304 stainless steel.

Measurements of the effective stress intensity factor range ratio U were performed to explain the effect of stress ratio R.

The corrosive environment decreased da/dN at R = 0.1, 0.4 and little affected da/dN at R = 0.9 for SNCM 2 and increased da/dN at all R ratios for SUS 304.

It was confirmed that there exists a threshold stress intensity factor ΔK_thCF in 3% NaCl solution for both materials tested.

The corrosive environment decreased ΔK_thCF for all conditions tested except at R = 0.1 and 0.4 for SNCM 2, where ΔK_thCF-values were nearly equal to ΔK_th-values in air. ΔK_thCF/ΔK_th was 0.6 at R = 0.9 for SNCM 2 and 0.8, 0.5 and 0.7 at R = 0.1, 0.7 and 0.9 for SUS 304, respectively.

It was shown that the complicated effect of stress ratios on crack growth for SNCM 2 can be explained using effective stress intensity factor ΔK_eff.     

A microcomputer control system for the measurement of fatigue crack growth data

A microcomputer-based system for the measurement of fatigue crack growth da/dn versus cyclic stress intensity factor ΔK data using compact-tension test specimens is described. The procedure has been developed to allow automatic measurement of crack growth rate under any specified combination and sequence of load conditions, i.e. ΔK and R (stress ratio) and includes the capability of establishing the threshold cyclic stress intensity factor ΔK₀. Crack extension measurement is effected from the elastic compliance evaluated from the AC component of the load and displacement signals to an accuracy of -3 μm every 1000 load cycles. Results from a typical low-alloy-steel rotor forging are presented to illustrate the use of the system.     

Study of crack tip opening under cyclic loading taking into account the environment and R ratio

This study involves the R effect and environment effect on crack closure mode, in 7175 T 651 aluminium alloy. To obtain one of the selected objectives, it was necessary to use a clip gauge located at the notch of the compact specimen and a C.T.O.D. gauge located at the crack tip. The crack opening phenomena observed in our tests depends on the applied method which accounts for the differences in ΔK_eff found in the literature. The systematic use of the two methods allowed us to bring to light common features permitting the calculation of ΔK_eff according to Elber's criteria. The concept of ΔK_eff does not fully explain the influence of R ratio and the environment effect.     

Modelling and measurement of crack closure and crack growth following overloads and underloads

Ignoring crack growth retardation following overloads can result in overly conservative life predictions in structures subjected to variable amplitude fatigue loading. Crack closure is believed to contribute to the crack growth retardation, although the specific closure mechanism is debatable. The delay period and corresponding crack growth rate transients following overload and overload/underload cycles were systematically measured as a function of load ratio (R) and overload magnitude. These responses are correlated in terms of the local “driving force” for crack growth, i.e. the effective stress intensity factor range (ΔK_eff). Experimental results are compared with the predictions of a Dugdale-type crack closure model and improvements in the model are suggested.     

Understanding fatigue crack propagation in AISI 316 (N) weld using Elber’s crack closure concept: Experimental results from GCMOD and acoustic emission techniques

An experimental study of fatigue crack propagation and crack closure behaviour, in compact tension specimens of AISI 316 (N) weld has been conducted. The crack closure load was determined from the changes in the slope of the load–displacement curves using global crack mouth opening displacement (GCMOD) type gauge. The results were compared with those measured by acoustic emission technique which showed good agreement with each other. The experimental data bear clear evidence of fatigue crack closure. The crack opening force was found to increase moderately with crack length and increasing R-ratio, under a constant P_max of 5 kN. Above a critical R-ratio of 0.45 (approximately), the crack closure load is smaller than the minimum applied load. A good correlation was obtained for ΔK_eff/ΔK = 0.6684 – 2.4135R + 7.0077R² in the range 0 R  0.5. The magnitude of crack closure is used to interpret observed crack growth behaviour at different R-ratios.     

The effect of load ratio on the threshold stresses for fatigue crack growth in medium carbon steels

A study of the effect of load ratio, R, on the low crack growth rates and the threshold stress conditions exhibited by five medium carbon steels has been conducted. It was found that decreased values of R retarded crack growth to an increasing degree as a defined threshold for crack growth was approached, such that this threshold showed a marked dependence on K max rather than ΔK. Comparisons of the five materials showed that the threshold increased as yield strength increased for a given R, but this effect could be normalised in terms of a constant value of the maximum plastic zone size at the crack tip.     

Fatigue crack propagation rate at low ΔK of two aluminum sheet alloys, 2024-T3 and 7075-T6

For transport aircraft with long lifetimes, crack growth data to implement the durability life requirements in the low ΔK range are needed. This is the region comprising most of the lifetime for the cracks of interest, and it is also the region where there is little data available. Crack growth data in the form of da/dN vs ΔK vs R at constant amplitude for two primary aircraft aluminum sheet alloys, 2024-T3 and 7075-T6 (clad) were measured in laboratory air and 140°F in the low da/dN region 10⁻⁸ to 10⁻⁵ in./cycle. Crack growth rates were correlated with stress state and with fractographic features.     

The growth of fatigue cracks through particulate sic reinforced aluminum alloys

Crack growth rates for large fatigue cracks in 12 variations of particulate silicon carbide reinforced aluminum alloy composites have been measured. Composites with seven different matrix alloys were tested, four of which were of precipitation hardening compositions, and those were tested in both as-extruded and peak aged conditions. Five of the materials were made by casting, ingot metallurgical methods and two of the alloys by mechanical alloying, powder metallurgical methods. For both manufacturing methods, primary fabrication was followed by hot extrusion. The fatigue crack growth curves exhibited an approximately linear, or Paris law, region, fitting the function da/dN = BΔK^s, and a threshold stress intensity factor, ΔK_th. As has been found for other materials, the coefficients B and s are correlated; for these composites In B= −16.4−2.1s. A correlation was also found between ΔK_th and s, and it was found possible to compute the magnitude of ΔK_th using a simple model for the threshold together with yield stress and SiC size and volume fraction. These results were explained using a relationship between ΔK_th and crack closure determined previously for unreinforced aluminum alloys. The path of fatigue crack growth is through the matrix for these composites, and SiC has the effect of altering the slip distance, therefore, the plasticity accompanying fatigue cracks. It was shown that all the crack growth rate curves were reduced to one equation having the form da/dN = B'ΔK_eff^s' where B' = 6.5 × 10^-9m/cy and s' = 1.7. A partly theoretical method for predicting fatigue crack growth rates for untested composites is given. Fatigue crack surface roughness was measured and found to be described by a fractal dimension, but no correlation could be obtained between surface roughness parameters and ΔK_th.     

Fatigue properties of ultrafine grained low carbon steel produced by equal channel angular pressing

Ultrafine grained low carbon (0.15 wt.% C) steel produced by equal channel angular pressing (ECAP) was tested for investigating fatigue properties, including cyclic softening and crack growth rate. Emphasis was placed on investigating the effect of load ratio on the fatigue crack growth rates of ultrafine grained microstructure. The ECAPed steel exhibited cyclic softening. After the first cycle, the tension and compression peak stresses decreased gradually with the number of cycles. Fatigue crack growth resistance and the threshold of ECAPed ultrafine grained steel were lower than that of an as-received coarse grained steel. This was attributed to a less tortuous crack path. The ECAPed steel exhibited slightly higher crack growth rates and a lower ΔK_th with an increase in R ratio. The R ratio effect on growth rates and ΔK_th was basically indistinguishable at a lower load ratio (R>0.3) compared with other alloys, indicating that the contribution of the crack closure vanished. This was explained by the fact that finer grained materials produce a lower opening load P_op due to a relatively less serrated crack path. Consequently, K_min can reach K_op readily with a smaller increment of load ratio. The crack growth rate curve for the ECAPed ultrafine grained steel exhibited a linear extension to the lower growth rate regime than that for the coarse grained as-received steel. This behavior can be explained by a reverse crack tip plastic zone size (r_p) that is always larger than the grain size.     

The use of ΔK when examining microstructural effects on small fatigue crack growth

The behaviour of small fatigue cracks has been studied in the Al---Li---Cu---Mg---Zr alloy 8090. It was found that the crack inclination normal to the surface of the specimen made crack deflections and kinking in the plane of the specimen surface irrelevant to the crack driving force. The low closure levels associated with small fatigue cracks reduce the effect of microstructure on crack growth but this does not affect the ability of ΔK (stress intensity factor range) to detect microstructural influences. The use of ΔJ (J-integral range) as a correlating parameter reduced the differences between the data for long and short fatigue cracks. However, there was no evidence that ΔJ was superior at identifying microstructural effects. Similarly the effect of the higher-order terms on the value of ΔK was found to be minor. It is concluded that the use of ΔK is not likely to bias the microstructural effects and so ΔK may be used when examining microstructural effects on small fatigue crack growth.     

A study of the effects of mechanical and environmental variables on fatigue crack closure

Using the potential drop technique, fatigue crack closure has been monitored in pin loaded SEN specimens of -titanium, a titanium alloy and EN24 steel. The specimens were tested in tension-tension under conditions closely approximating to plane strain, and closure was only detected in vacua of better than 133 mN m⁻² (10⁻³ torr). No significant closure was detected in air. The extent of the crack area closed at minimum load varied with air pressure, applied stress, R ratio (R = L_min/L_max), crack length, material, and loading mode. Additional experiments made with a dip gauge showed that the COD/applied load response of the crack was non-linear in vacuum above minimum load in the fatigue cycle confirming that crack closure was occcurring. It is shown that for a given material, loading mode and air pressure, the effect of loading and crack length variables on crack area closed at minimum load can be characterised in terms of the parameter (K²_min-ΔK², this being proportional to the calculated COD at minimum load. The extent of closure in vacuum is influenced extensively by this parameter.     

Effects of welding residual stresses on fatigue crack growth behaviour in butt welds of a pipeline steel

In the present test the fatigue crack growth rate in the parent plate, weld and cross-bond regions was measured and the results were correlated with the stress intensity range ΔK and the effective stress intensity range ΔK_eff. It is indicated that the welding residual stresses strongly affect the crack growth rate. For the weld metal and cross-bond compact tension specimens in which crack growth is along the weld line the fatigue crack growth rate increases as the crack grows. However, for the T compact tension specimen in which crack growth is perpendicular to the weld line at a constant value of applied ΔK the crack growth rate initially decreases as the crack grows. Particularly, at a low constant value of applied ΔK the crack growth rate obviously decreases and the crack fails to grow after short crack growth. When the crack grows to intersect the welded zone, the fatigue crack growth rate gradually increases as the crack grows further. It is clear that the effect of welding residual stresses on the crack growth rate is related to the position of the crack and its orientation with respect to the weld line. Finally, the models of welding residual stress redistribution in the compact tension specimens with the growing crack and its influence on the fatigue crack closure are discussed. It appears that for a butt-welded joint one of the crack closure mechanisms may be considered by the bend or rotation deformation of crack faces due to the welding residual stress redistribution as the fatigue crack grows in the welded joint.     

A mechanistic model for the influence of stress ratio on the LEFM fatigue crack growth behavior of metals and alloys—I. Crack-ductile materials

Concentrating on local behavior of a highly stressed zone ahead of the crack tip, a recent mechanistic approach to analyse LEFM fatigue crack growth behavior in three stages at stress ratio R = 0 is extended here to include the effect of a positive stress ratio. This paper is limited to analysing primarily the stages I and II of “crack-ductile” materials, characterised by a purely “reversed shear” (or ductile “striation”) growth mechanism in stage II. It is shown that in these materials stage I is R-sensitive and stage II is insensitive, and these can, without invoking crack closure arguments, be rationalised alternatively by considering the dominance of a K_max-controlled “Submicroscopic Cleavage” and a ΔK-controlled “ reversed shear ” fracture mechanism, respectively. Assuming Paris type power relations to hold, a predictive model is developed that contains separate growth equations with R-effect for stages I and II and shows the existence of a characteristic “master shear-curve” and a “moving pivot-point” on this curve for a class of materials. Good agreement was found between quantitatively predicted growth curves at selected R-values and a relatively large volume of available experimental data for low strength steels, aluminum alloys and titanium alloys. Besides providing more physical explanations for the observed growth behavior, the model may also be useful as a convenient alternative to crack closure for obtaining fairly accurate and conservative estimates of fatigue life for design applications.     

Fatigue crack-growth characteristics of two magnesium alloys

Magnesium alloys are being increasingly used for engineering applications. Fatigue crack-growth data have therefore been obtained for a high strength magnesium-Zr alloy and a medium strength, weldable magnesium-Mn alloy. The results of tests on sheet material are presented in terms of the range of stress intensity factor ΔK. Critical values of ΔK necessary for fatigue crack growth ΔK_c were also obtained. The behaviour of the two alloys was similar; both rates of crack growth and ΔK_c were sensitive to mean stress. Fatigue crack growth was entirely on a 90° plane with no sign of the transition to crack growth or 45° planes usually observed in sheet materials. This was ascribed to the effects of preferred orientation of the crystal structure.     

Fatigue crack growth characteristics of rotor steels

Room temperature fatigue crack growth rate data were generated for Ni-Mo-V (ASTM A469, Cl-4), Cr-Mo-V (ASTM A470, Cl-8) and Ni-Cr-Mo-V (ASTM A471, Cl-4 and a 156,000 psi yield strength grade) rotor forging steels. Testing was conducted with WOL type compact toughness specimens and the results presented in terms of fracture mechanics parameters. Data show that the Ni-Cr-Mo-V steels exhibit slower fatigue crack growth rates at a given stress intensity range (ΔK) than do the Ni-Mo-V steels. In addition, the Cr-Mo-V steel was found to exhibit slower growth rates than the other alloys at ΔK levels below 40 ksi √in but somewhat foster rates at ΔK levels in excess of 45 ksi √in. The fatigue crack growth rate properties of the alloys studied conform to the generalized fracture mechanics crack growth rate law where da/dN = C₀ΔK^R. It was noted that the fatigue crack growth rate parameters n and C₀ tend to decrease and increase, respectively, with increasing material toughness, K_ic.     

Fatigue crack initiation and growth under mixed mode loading in aluminum alloys 2017-T3 and 7075-T6

Fatigue crack initiation and growth characteristics under mixed mode loading have been investigated on aluminum alloys 2017-T3 and 7075-T6, using a newly developed apparatus for mixed mode loading tests. In 2017-T3, the fatigue crack initiation and growth characteristics from a precrack under mixed mode loading are divided into three regions—shear mode growth, tensile mode growth and no growth—on the ΔK_I-ΔK_II plane. The shear mode growth is observed in the region expressed approximately by ΔK_II > 3MPa√m and ΔK_II/ΔK_I > 1.6. In 7075-T6, the condition of shear mode crack initiation is expressed by ΔK_II > 8 MPa√m and ΔK_II/ΔK_I > 1.6, and continuous crack growth in shear mode is observed only in the case of ΔK_I/ΔK_II, 0. The threshold condition of fatigue crack growth in tensile mode is described by the maximum tensile stress criterion, which is given by Δσ_θmax √2πr 1.6MPa√m, in both aluminum alloys. The direction of shear mode crack growth approaches the plane in which K_I decreases and K_II increases towards the maximum with crack growth. da/dN-ΔK_II relations of the curved cracks growing in shear mode under mixed mode loading agree well with the da/dN-ΔK_II relation of a straight crack under pure mode II loading.     

Flexural fatigue characteristics of sandwich structures at different loading frequencies

The effects of frequency on the fatigue behavior of S2 glass fiber–vinylester reinforced sandwich composites with two different PVC cores have been investigated. Flexural fatigue tests were performed on sandwich beams with core densities of 130 and 260 kg/m³ at frequencies of 3 and 15 Hz, at a stress ratio, R=0.1 and at four different load levels viz. 90%, 85%, 80% and 75% of the ultimate load. S–N diagrams were generated and it was observed that the fatigue strength increased with core density, and the number of cycles to failure, N_f, increased with increase in frequency. In all cases failure was dominated by a primary shear crack in the core however, the crack path and crack propagation rates varied with frequency. The fatigue crack growth rate (FCGR) in the core of the H130 sandwich beams was subsequently investigated and the relationship between the crack growth rate, da/dN, and the cyclic stress intensity range, ΔK, was determined. It was found that crack growth rate decreased with increase in loading frequency.     

Fatigue crack closure

A loading stage for the scanning electron microscope has been used in conjunction with the stereoimaging technique to study fatigue crack closure for a center notched specimen of an aluminum alloy. The results are compared to similar measurements from single edge notched specimens of the same alloy and other materials. The magnitudes of closure loads were found to be different for these two specimen designs, but the dependence on stress intensity factor (ΔK) is the same. Comparison of experiments is made with a finite element model simulation of crack opening, a simple model is developed which simulates the observed opening behavior, and a method is given for estimating threshold ΔK from materials parameters and crack opening behavior in the near threshold region.     

On the calculation of closure-free fatigue crack propagation data in monolithic metal alloys

A computational method is described for the determination of ΔK_b, corresponding to a fatigue crack growth rate of b/cyc, where b is the Burgers vector for a monolithic metal alloy. ΔK_b is found to be numerically equal to E√b for the case of closure-free crack growth behavior. Given that the closure-free FCP rate of many monolithic metals varies with ΔK³, the growth rate of metal alloys at ΔK ΔK_b is given by da/dN = (ΔK/E)³(1/√b. Excellent agreement is found between experimental and computed FCP data for the case of monolithic metal alloys. The limits of these relations for metal-matrix composites and ceramics are discussed.     

Near-threshold fatigue crack growth at room temperature and an elevated temperature in Al---Li alloy 8090

The effects of slip distribution and crack tip shielding mechanisms on the near-threshold fatigue crack growth of the Al---Li alloy 8090 have been studied at both room temperature and an elevated temperature. The slip distribution has been varied by changing the distribution of the S phase, through prior stretching or by means of a duplex heat treatment. Fatigue crack growth (FCG) tests were conducted at a high stress ratio to reduce possible effects due to crack closure.

At room temperature the changes in FCG rates are interpreted as arising from the changes in the degree of planarity of slip in the materials.

At 150°C, the microstructural changes due to the long exposure to elevated temperature appear to dominate the effects observed. At lower ΔK, where the time at temperature is greatest, lower ΔK thresholds than those found at room temperature are obtained. These have been attributed to increased slip homogenization due to the increased precipitation and coarsening of the incoherent S phase together with loss of toughness due to the growth of coarse grain boundary phases and the formation of the associated δ′ precipitate free zone.

At higher ΔK, where the time at temperature is low and microstructural changes are minimal, slower FCG rates than those found at room temperature are obtaine. These are explained in terms of increased crack tip shielding which arises because of the increase in tortuosity of the crack path, the increased slip homogenization and the climb and cross-slip within the crack tip plastic zone.