Showing papers in "Fatigue & Fracture of Engineering Materials & Structures in 1993"

A path-independent parameter for fatigue under proportional and non-proportional loading

Abstract: -A path-independent multiaxial fatigue damage criterion is proposed based on critical plane concepts: fatigue crack growth is controlled by the maximum shear strain, and an important secondary effect is due to the normal strain excursion over one reversal of the maximum shear strain. The effect of loading path on fatigue endurance is quantified by the normal strain excursion. Only one multiaxial material constant is required in the model which can be determined from uniaxial test data plus one torsional result. The parameter can be easily integrated with a shear strain-life relationship to predict low cycle fatigue lifetime. Experimental data of four different materials: En1 5R steel, I% Cr-Mo-V steel, 304 stainless steel, and 316 stainless steel at two temperatures were used to verify the criterion. It is shown that the proposed parameter can satisfactorily correlate test results for various proportional and non-proportional straining paths. NOMENCLATURE

The two thresholds of fatigue behaviour

Abstract: Two limiting thresholds to fatigue crack propagation are discussed. The first threshold is related to the microstructural texture and this threshold may therefore be deemed a material-based threshold. The second threshold is mechanically-based, and is related only to the stress state at the tip of a substantial defect. The material-based threshold is characterized in terms of Microstructural Fracture Mechanics (MFM) and the mechanically-based threshold is characterized in terms of Linear Elastic Fracture Mechanics (LEFM). The former condition is important when considering the fatigue limit of materials and components, while the latter is more applicable to the fatigue limit of structures. The different factors which affect the two threshold conditions are briefly presented. Finally, this paper discusses aspects of MFM relevant to the fatigue resistance of metals and components.

Note on cycle counts in irregular loads

Abstract: In this paper we discuss cycle counting methods, such as rainflow-, crest-to-trough-, positive peak-count and different damage accumulation rules, for irregular random loads, which have an infinite number of local extremes in finite intervals, e.g. the fourth spectral moment is infinite. We present conservative bounds for the expected damage for such loads from the upcrossing intensity. These results are illustrated by examples of Gaussian, xz-, and Morison-loads. NOMENCLATURE E(X) = mathematical expectation of the random variable X E(X I Y = y) = conditional expectation of X given Y = y Var(X) = variance of the random variable X a E A = a is an element of the set A {f; . } = the set of points t fulfilling the condition . # { . } = the number of elements in the set { . } 1 {. ,(x) = the indicator function of the set { . } sup{ . } = the supremum (least upper bound) of elements in the set { . } inf{ } = the infimum (greatest lower bound) of elements in the set { . 1

Elevated temperature fatigue crack growth in alloy 718—part ii: effects of environmental and material variables

Abstract: Observations concerning the effects of the environment and material variables on the crack growth process in alloy 718 are reviewed and analyzed on the basis of deformation characteristics in the crack tip region. The review of the role of material variables has focused on the effects of chemical composition and microstructure parameters including precipitate size and morphology as well as grain size and morphology. These analyses have suggested that the governing mechanism at the crack tip is the degree of homogeneity of plastic deformation and associated slip density. For conditions promoting homogeneous plastic deformation, with a high degree of slip density, the environmental damage contribution is shown to be limited, thus permitting the dominance of cyclic damage effects which are characterized by a transgranular crack growth mode and a lower crack growth rate. Under conditions leading to inhomogeneous plastic deformation and lower slip density the crack tip damage is described in terms of grain boundary oxidation and related intergranular fracture mode. Considering that the crack growth damage mechanism in alloy 718 ranges from fully cycle dependent to fully environment dependent, conflicting experimental observations under different operating conditions are examined and a sensitizing approach is suggested to increase the alloy resistance to environmental effects.

Repair efficiency in fatigue‐cracked aluminium components reinforced with boron/epoxy patches

Abstract: This paper describes studies on fatigue crack propagation in cracked aluminium alloy (2024 T3) panels repaired with boron/epoxy patches, adhesively bonded with either an epoxy-nitrile film adhesive or an acrylic adhesive. Studies were undertaken to assess the effect on patching efficiency of (a) disbonding of the patch system and (b) test temperature. A simple model is proposed for estimating the reduction of patching efficiency due to cyclic disbonding of the reinforcement. In the elevated-temperature tests it was found, unexpectedly, that patching efficiency in panels patched using the film adhesive was unaffected by temperatures up to 100°C.

Modeling the long‐life fatigue behavior of a cast aluminum alloy

Abstract: As-cast specimens and smooth specimens of a AA 319 cast aluminum alloy containing casting porosity were fatigue tested with special attention given to the long-life region (N 1.25 × 108 cycles). Fatigue cracks were observed to initiate from the near-surface casting pores or from discontinuities resulting from the as-cast surface texture. The observed fatigue lives were strongly dependent on the size (√area) of these casting defects. The effect of casting defects on the fatigue life was modeled assuming the fatigue life to be the sum of the crack nucleation and the crack propagation life (including both the growth of short and long cracks). The crack growth behavior of (mechanically) short cracks was considered in detail by a developed crack-closure-at-a-notch (CCN) model. The CCN model predicted the fatigue lives for both as-cast and machine-notched specimens. Extension of the CCN model to reliability-based design was attempted using the measured size distribution of the fatigue-initiating casting pores.

Elevated temperature fatigue crack growth in alloy 718—part i: effects of mechanical variables

Abstract: — In this paper observations concerning the effects of mechanical variables on the crack growth process in alloy 718 are reviewed and analyzed on the basis of the related deformation characteristics in the crack tip region. The variables included temperature, frequency, wave shape, hold time, load ratio and load interaction. These analyses have suggested that the role of each parameter in the acceleration of crack tip damage is governed mainly by their relative influence on the nature of the corresponding plastic deformation and associated slip line density. On the basis of this view (which assumes crack growth damage covers the range from cyclic- to fully time-dependent processes), the interactive effects of loading parameters are discussed when considering the corresponding fracture mode. Conflicting experimental observations under different operating conditions are examined.

Determination of predominantly mode ii stress intensity factors from isochromatic data

Abstract: — A new approach to the solution of stress intensity factors from isochromatic data has been developed and is based on complex Fourier analysis. Solutions are shown to be accurate when the loading of the crack tip is predominantly Mode II and also for cases that are generally mixed-mode. The new method and its application to a four-point-bend test is described in this paper. Tests on angled-edge-cracked specimens have also been performed to show the validity of the method for a range of mixed-mode loadings.

Limit load and j‐estimates for idealised problems of deeply cracked welded joints in plane‐strain bending and tension

Abstract: Using simple assumed deformation fields, approximate solutions have been obtained for tension and bending specimens containing welds for both limit loads and for fully plastic proportionality coefficients between the singular field amplitude J and the strain energy. The solutions allow for the degree of over or undermatch in the material tensile properties of the weld metal relative to the parent steel, and for the size of the weld region relative to the remaining ligament ahead of the crack. Detailed finite-element analyses have been performed for particular values of under/over-match and size of weld region. These refine the approximate analytical solutions for the particular cases examined, and show broad agreement with the trends predicted by the analytical models. The results have been used to provide guidance for testing weldments using standard, bend-type geometries. For small-specimen testing, cracks should be sufficiently deep for the remaining ligament ahead of a centrally located crack to be less than the total width of the weld. For large specimens, the weld region should be less than 20% of the size of the remaining ligament. If these guidelines are followed then standard relationships may be used to derive J from the area under the load-displacement curve. Common advice that the tensile properties of the weaker material in a weldment should be used in J-estimation techniques has been shown to be appropriate in many cases. However, the advice is likely to be overconservative when plastic deformation is predominantly in the weld even for overmatched weld properties, or predominantly in base metal even for undermatched weld properties. The results in the paper enable such cases to be identified.

Size effects and a mean‐strength criterion for ceramics

Abstract: On the basis of a large amount of research on the fracture behaviour of ceramics, a new failure criterion is advanced. It is pointed out that the rupture of structural ceramics is governed by the mean stress in a specific zone, rather than by the peak stress value. The size of the process zone was derived as a material constant, and reflects the relation between the toughness and the strength of the material. The results show that the critical maximum stress varies with the stress gradient: the steeper the stress gradient, the higher the critical peak stress. Different experiments concerning the size effect have been carried out and the reason for size effects has been explained by a mean strength theory.

Effective stresses and microstructure in cyclically deformed 316l austenitic stainless steel: effect of temperature and nitrogen content

Abstract: — This paper reports a study on the factors influencing dislocation slip during cyclic deformation of 316L austenitic stainless steel. TEM investigations show that low temperature and interstitial nitrogen content favour planar slip and lead to higher erective stress values. Measurements of effective and internal stresses with the Handfield-Dickson technique indicate that the contribution of nitrogen in the effective component is more important than that of temperature. It is deduced that nitrogen acts through a pinning effect, while low temperature exerts an effect on friction stress. The results also suggest that cyclic plasticity could modify the short range order leading to a redistribution of nitrogen.

A study of superimposed fracture modes i, ii and iii on pmma

Abstract: A test apparatus has been developed to study the fracture behaviour of engineering materials subjected to superimposed tensile and shear (Mode I and II) loads using a single edge notch specimen. Stress intensity factors were calculated using finite element analysis. Test results for PMMA are reported. Results from circumferentially notched round bar specimens, subjected to combined tension and torsion (Mode I and III) loading are also reported. The Mode I/II results are consistent with the mixed mode fracture response of a wide range of brittle materials, although there is some evidence of non-linear behaviour. The fracture behaviour for superimposed Modes I and III indicates the strong influence of non-linear deformation which causes the mixed mode toughness to be dependent on the sequence and type of loading.

Prediction of the first spinning cylinder test using ductile damage theory

Abstract: Using the standard Rousselier ductile damage theory, a prediction is made of crack growth in the First Spinning Cylinder Test carried out by AEA Technology, Risley. Only data from small-scale mechanical tests and metallography have been used. The prediction has been “blind” in the sense that, although we were aware of the observed high J-R -curve (hitherto defying detailed interpretation by any of the standard fracture mechanics techniques), we restrained ourselves from any attempt to use the detailed results of the test itself in making the prediction. Indeed, we were in some confusion about the reported J-R-curves until a final meeting when our predictions were discussed with the experimental workers. Furthermore, crack growth versus rotation speed was predicted before sight of the experimental results. Originally, reasonable estimates of the size scale determining the averaging involved in the damage theory gave predictions of crack growth versus rotation speed which agreed very well with the cylinder test, except near initiation, where the apparent increase in initiation toughness was not predicted. These predictions involved the parameter J only through the use of the compact tension specimen results in fitting the damage theory parameters and were quite independent of any complications which might arise from the use of J when centrifugal or body forces are present. The J-R -curve predictions were not quite so good, but were still well above the small specimen results. Again, there was discrepancy near initiation. Subsequently, we have learned that metallography which we were led to believe applied to a section perpendicular to the line of the mean crack front in fact referred to a section parallel to the mean crack plane. A rough correction using limited additional data improves our “blind” prediction slightly, but other limitations of the available small-scale data mean that, had some suggestions in the literature been adopted, a result more in accord with that expected from the small-scale tests might well have been obtained. It is concluded that further development of the modelling and the associated large- and small-scale testing is required. Some comments on this further development and the potential of the method as a predictive engineering tool for the transfer of results from specimens to structures are made.

THE EFFECT OF MICROSTRUCTURE AND FRACTURE SURFACE ROUGHNESS ON FATIGUE CRACK PROPAGATION IN A Ti‐6A1‐4V ALLOY

Abstract: Characteristics of fatigue crack propagation (FCP) have been studied on materials with three different microstructures of a Ti-6A1-4V alloy, prepared with different heat treatments. The effect of microstructure on the FCP behaviour was attributed to the development of crack tip shielding, primarily resulting from the role of crack path morphology in inducing crack closure and crack deflection. Roughness-induced crack closure played an important role on the near-threshold FCP behaviour at a stress ratio of 0.05, but the FCP data plotted in terms of the effective stress intensity factor range, δKeff (allowing for crack closure), still exhibited the effect of microstructure. Fractographic examinations were performed, using a scanning electron microscope (SEM) with the aid of image processing, which enabled a three-dimensional reconstruction of the fracture surface using a stereo pair of SEM micrographs. Fracture surface roughness was evaluated quantitatively by the ratio of the real area of the reconstructed fracture surface to its projected area. As fracture surface roughness was taken into account in evaluating the FCP data in addition to crack closure, the effect of microstructure disappeared, indicating that the intrinsic FCP resistance was the same in all the materials. Thus, it was concluded that fracture surface roughness was a dominating parameter in controlling the FCP of the Ti-6A1-4V alloy.

A crack closure model for predicting the threshold stresses of notches

Abstract: A model is proposed to estimate the threshold stress range of a notched component. The model considers the variation of crack closure with crack length in the presence of a notch. The threshold stress range was found from the condition that the minimum value of effective threshold stress intensity range of a crack emanating from a notch equals the effective threshold stress intensity range of a “long” crack. The effects of notch depth, of notch acuity, of notch and specimen type, of load ratio, and of material properties on the threshold stresses were considered. Experimental data reported in the literature were used to assess the validity of the model. It was found that the model correctly predicts the behavior of cracks in notched components.

The use of a failure assessment diagram for initiation and propagation of defects at high temperatures

Abstract: The use of a failure assessment diagram, of the type in the R 6 defect assessment procedure, is investigated for creep crack growth under steady loading While a detailed approach based on C* remains attractive for appreciable creep crack growth, it is shown that a simplified approach can be formulated for limited crack extension A failure assessment diagram is derived based on the option 2 curve of R 6 using isochronous stress-strain data The inclusion of elastic strains in the isochronous data covers stress redistribution effects Equations are given which enable a toughness, Kmat, for assessments at temperatures in the creep range to be evaluated from creep crack incubation and growth data presented in terms of C* The toughness, Kmat, replaces the fracture toughness used in low temperature R 6 assessments Thermal stresses can be included in assessments by evaluating the stress intensity factor for the combined thermal and mechanical loading A formula is given which enables the effect of thermal stresses to be reduced when creep strains are sufficient to relax out part or all of the thermal stress

A normal distribution or a weibull distribution for fatigue lives

Abstract: The results of six series of 18 to 30 similar fatigue tests on three types of specimens are used to check the applicability of the normal distribution function and the three-parameter Weibull distribution function. A least-square procedure is adopted to obtain the three parameters of the Weibull function. Comments are made on the results obtained.

A probabilistic model for prediction of lcf surface crack initiation in pm alloys

Abstract: The Low Cycle Fatigue (LCF) life of PM Ni-base superalloys is commonly reduced by surface crack initiation at ceramic inclusions. For this reason, a probabilistic model has been developed that predicts the size of surface crack initiation sites from the inclusion size distribution. For the experimental correlation of the model two sets of alloys were examined: a “standard” (i.e. as-received) alloy, and a second material of identical composition to which a known distribution of ceramic inclusions was incorporated (or seeded). Model predictions were found to be in excellent agreement with the results obtained from the seeded materials in which the defect size distribution is larger and better characterized, and were satisfactory for the unseeded material in which two types of surface defects (pores and ceramic inclusions) initiate LCF cracks. The results of these experiments were employed in LCF simulations of both test specimens and full scale components. These indicated that differences exist between the site preference for LCF crack initiation in small test specimens and large scale components due to a scale effect. Such results demonstrate the utility of seeding experiments for generation of LCF test data used in component design.

Crack‐line and edge green's functions for stress intensity factors of inclined edge cracks

Abstract: — Fretting loads on the surfaces of structural components can cause accelerated growth of short cracks. The rate of growth will depend on the combined stress intensity factor resulting from both remote and local loading. Many stress intensity factor solutions are available for remote loading, but solutions for arbitrary fretting loads are not readily accessible. In this paper accurate crack-line Green's functions are obtained from a boundary element analysis and then used to develop the Green's functions for loads on the edge of a half-plane containing a slant crack at various angles to the edge. These latter Green's functions can be used to obtain stress intensity factors for arbitrary stresses (normal or shear) on the edge of the half-plane without further stress analysis; simple integration procedures are all that is required.

Aspects of high-temperature low-cycle thermomechanical fatigue of a single crystal nickel-base superalloy

Abstract: High-temperature thermomechanical fatigue (TMF) tests were performed on monocrystalline specimens of the γ′-precipitate-strengthened nickel-base superalloy CMSX-6. The specimens had axial orientations near [001]. In-phase, out-of-phase, clockwise-diamond and counter-clockwise-diamond TMF tests were conducted in total strain control in symmetric push-pull. The temperature interval extended from 873 to 1373 K, the total strain amplitude was 0.5% with a cycle time of 300 s. The stress response, the plastic strain amplitudes and the fatigue lives were measured. In addition, fatigue-induced micro-structural changes were investigated. The shapes of the hysteresis loops and the mechanical TMF behaviour varied markedly with the shape of the strain-temperature cycles. The TMF life of the alloy was found to correlate with the stress amplitude in the tensile half-cycle. The evolution of the microstructure shows the formation of dislocation networks around the precipitates and in particular, severe changes in the size and the morphology of the γ'-precipitate structure. These changes vary with the strain-temperature cycle shape. Failure occurred exclusively by fatigue (and not creep) in the form of localized shear.

Prediction of the lifetime distribution of high‐strength components subjected to fatigue loading

Abstract: A probabilistic model is developed for predicting the lifetime distribution of a high-strength component subjected to low cycle fatigue loading The damage process considered consists of two consecutive stages First, cracks are initiated from microstructural inhomogeneities such as non-metallic inclusions These cracks will then extend stably until final failure occurs or the component is removed from service The probabilistic model is used to determine the maximum tolerable inclusion size for a model turbine disc This quantity is defined as the size of the initial defect for which the number of load cycles to failure at a given reliability level exceeds the design life of the component This maximum tolerable inclusion size is compared with the results of a deterministic lifetime analysis

The effect of artificial fatigue‐crack closure on fatigue‐crack growth

Abstract: An investigation of the effects of artificial fatigue-crack closure on the subsequent growth of fatigue cracks, at constant range of stress-intensity factor, was carried out for the following closure materials: (i) electroplated nickel; (ii) electroless nickel; (iii) ethylcyanoacrylate (ECA) adhesive; (iv) low-melting point (LMP) solder. It was observed that (a) the wedge thickness and area of crack surface covered by the artificial-closure material and (b) the distance of the tip of the closure wedge from the crack tip (i.e. the penetration of the closure material towards the crack tip) were the main parameters controlling the reduction in the subsequent growth rate of the fatigue crack. The results suggest that if the crack thickness (opening) is completely filled with a dense wedge of artificial-closure material, the wedge becomes effectively rigid and the precise mechanical properties of the closure material will have only a secondary influence on subsequent fatigue-crack growth in comparison to (a) and (b) above. This follows from the fact that the closure material is effectively a very thin strip compressed between the rough (high friction) sides of a crack. A rigid-wedge and elastic-crack model was found to successfully predict the changes in fatigue-crack growth rate following artificial crack-closure and gave a lower bound to the measured growth rates.

Characterization of ductile tearing resistance by energy dissipation rate

Abstract: The J integral, which is widely used in elastic-plastic fracture mechanics, is not the true driving force any more if the crack is propagating. This leads to some inconsistencies when ductile tearing resistance is characterized in terms of J, especially for large crack extensions. Instead, Turner has proposed the energy dissipation rate as a physically more meaningful quantity. His concept is discussed and more evidence is given that will provide a better understanding of ductile tearing. It is shown how this quantity can be determined by measuring the heat production ahead of a fast running crack, or calculated in a finite element analysis, or re-evaluated from J-R test records of bend and tensile specimens. The energy dissipation rate is decreasing with crack extension in gross plasticity and approaches a stationary state. From these relations, the shapes of the cumulative J-R curves can be derived for different specimen geometries.

Strategy for characterizing fracture toughness in the ductile to brittle transition regime

Abstract: A tentative method is proposed for treating the scatter and size effects of fracture toughness of ferritic steels in the ductile-to-brittle transition regime. The method is aimed at the determination of the probability function for initiation of fast growing cleavage cracks under the plane strain condition. The practical relevance of this method has been checked with toughness data sets obtained from various CT-specimen sizes and CCT-specimens of a pressure vessel steel tested at two different temperatures in the transition regime

Effect of surface texture on fatigue life in a squeeze-cast 6082 aluminium alloy

Abstract: The effect of surface texture on fatigue life in a squeeze-cast Al-Mg-Si alloy has been investigated in terms of initiation and subsequent propagation of small fatigue cracks. Small cracks nucleated from defects on both as-cast and polished surfaces. However, the large (∼90μm in diameter) surface defects found on as-cast surfaces caused an approximate 20% reduction in fatigue life when compared to polished surfaces where cracks initiated from smaller (∼ 12 μm in diameter) microporosity. Integration of averaged small fatigue crack growth data enables the fatigue lives of both types of specimen to be predicted at stresses above the fatigue limit, showing that the difference in fatigue behaviour is principally due to the extra period of crack growth in the polished surfaces.

Effects of crack length, notch root radius and grain size on fracture toughness of fine ceramics

Abstract: Generally, fracture toughness and fracture stress of ceramics depend on crack length, notch root radius and grain size. These three parameters are most important when assessing the integrity of structural ceramic members and developing high-performance ceramics. A new failure criterion called the process zone size failure criterion, has been proposed based on the existence of a crack-tip process zone. Using this criterion, it is shown that theoretical values are in good agreement with many test results quoted from many papers. It is concluded that this failure criterion is useful when evaluating crack length and notch root radius problems. The effect of grain size on both the fracture toughness and on the toughening mechanism is also considered.

A mechanistic-based thermomechanical fatigue life prediction model for metal matrix composites

Abstract: The framework for developing a mechanistic-based life prediction model for metal matrix composites is described. For a composite consisting of unidirectional silicon carbide fibers in a titanium aluminide matrix, SCS-6/Ti-24A1-1INb (at%) [0]8, three dominant damage mechanisms were identified: (1) matrix fatigue damage, (2) surface-initiated environmental damage, and (3) fiber-dominated damage. Damage expressions were developed for each mechanism along with a method for determining the constants. The damage is summed to obtain the total life. The model is capable of making predictions for a wide range of histories, including isothermal fatigue at different frequencies and stress-ratios, thermomechanical fatigue (TMF) under in-phase and out-of-phase cycling conditions, thermal cycling at constant stress, and stress holds at either maximum or minimum stress. Considering the wide range of cyclic conditions, the predictions compare favorably with experiments. In addition, the controlling damage mechanism for each history is predicted.

Fatigue crack growth in several ceramic materials

Abstract: Fatigue crack growth tests of five ceramic materials were carried out in order to investigate the general characteristics of cyclic and static fatigue crack growth in ceramics. A cyclic effect was found in Si 3 N 4 , Al 2 O 3 and TiB 2 , where the main fracture mechanism was intergranular separation, and stress shielding due to grain bridging occurred near the crack tip. On the other hand, no cyclic effect was observed in SiC and ZrO 2 , where the main fracture mechanism was transgranular fracture. The fatigue crack growth curves reduced to a unique curve by using the parameter K tip /E, regardless of the kind of ceramics studied

The initiation and propagation behaviour of short fatigue cracks in Waspaloy subjected to bending

Abstract: The fatigue crack growth behaviour of a nickel base superalloy, Waspaloy, has been studied using four point bending specimens in laboratory air. Two groups of tests, for which the span/width ratios were 1 and 2.5, were conducted and the results compared. Surface crack lengths were measured from plastic replicas of the surface. Equations which describe both short and long cracks have been derived and used to predict the fatigue life for the Waspaloy specimens. From plastic replication studies and scanning electron microscope examinations, a physical understanding of the relationship between crack growth and microstructural features was enhanced.

Fatigue behaviour of steel castings containing near‐surface defects

Abstract: Manufacturing defects in the near-surface region of 2¼% Cr 1% Mo castings were investigated in a joint research programme. After ultrasonic testing and X-ray inspection, large push-pull specimens (cross section 1000 mm2) were taken from castings in regions with indications of defects and tested under fully reversed loading in the elastic and elastic plastic region up to predetermined levels of defect growth. The defects are micropores, pin holes, slag and oxide layers or inclusions, hot tears and small micro cracks in welds. The initiation of crack growth was analysed by two methods, namely (a) the local strain concept with idealization of defects as three-dimensional elliptical notches and (b) by the fracture mechanics concept with idealization of defects as two-dimensional elliptical cracks. The main results can be summarized as follows: 1. The local strain concept describes the crack initiation potential of the defects more accurately than the fracture mechanics concept. 2. Interpretation of casting defects as cracks produces very conservative estimates in many cases. 3. Analytical evaluation of crack growth behaviour using the fracture mechanics approach for defects experiencing crack growth of at least 1 mm is much more accurate than when defects having a smaller extent of crack growth are included in the analysis. Additional evaluations of crack growth behaviour of hot tears by fracture mechanics show that crack growth starting from open and partially closed hot tears must be subjected to a larger number of cycles than normal fatigue cracks of sound material. Partially closed hot tears in the vicinity of the crack front additionally delay crack propagation.