Fatigue damage of low amplitude cycles in low carbon steel

The influences of low load cycles on fatigue damage in 0.15% C steel (C15E, No. 1.1141) are investigated in the very high cycle fatigue regime using ultrasonic fatigue testing equipment. Constant amplitude (CA) endurance limits at limiting lifetime of 10⁹ cycles are determined in cyclic tension–compression and cyclic torsion tests. Non-propagating fatigue cracks are found in specimens subjected to cyclic torsion loading at the endurance limit. The endurance limit is considered as maximum stress amplitude where possibly initiated fatigue cracks do not propagate to failure. Two-step variable amplitude (VA) tension–compression endurance tests are performed with repeat sequences consisting of high stress amplitudes above the endurance limit and far greater number of cycles below. The measured lifetimes are compared with linear damage accumulation calculations (Miner calculations). If the high stress amplitude is more than approximately 13% above the CA endurance limit, detrimental influences of low load cycles and failures at low damage sums are found. If the high stress is less than 13% above the CA endurance limit, numerous low load cycles cause prolonged fatigue lifetimes and specimens can sustain large damage sums without failure. Two-step VA fatigue crack growth investigations show that load cycles below the threshold stress intensity accelerate crack growth, if the high stress intensity is 18% or more above the CA threshold stress intensity. In repeat sequences with high stress intensities 14% above threshold stress intensity, low load cycles decelerated and stopped fatigue crack growth. Low load cycles can reduce or prolong fatigue lifetimes of low carbon steel and one reason is the accelerated or retarded fatigue crack growth due to numerous low amplitudes, and the maximum load amplitude of a VA load sequence determines whether detrimental or beneficial effects prevail.     

Fatigue analysis of unidirectional GFRP composites under combined bending and torsional loads

Fatigue behavior of unidirectional glass fiber reinforced polyester (GFRP) composites at room temperature under in-phase combined torsion/bending loading was investigated. All fatigue tests were carried out on constant-deflection fatigue machine with frequency of 25 Hz. A 30% reduction from the initial applied moments was taken as a failure criterion in the combined torsion/bending fatigue tests of the composite materials. A series of pure torsional fatigue tests were conducted to construct the failure contour of GFRP composites using different failure theories. The obtained S–N curves from combined torsion/bending tests were compared with both, pure torsion fatigue test results and published results of pure bending fatigue tests of GFRP rods. Pictures by scanning electron microscope were used to closely examine the failure mode of the tested specimens under combined torsion/bending loading.

The results showed that, the unidirectional glass fiber reinforced polyester composites have poor torsional fatigue strength compared with the published results of pure bending fatigue strength. Endurance limit value (calculated from S–N equation at N = 10⁷ cycles) of GFRP specimens tested under combined torsion/bending loading equals 8.5 times the endurance limit of pure torsion fatigue. On the other hand the endurance limit of combined torsion/bending fatigue strength approximately half the fatigue limit of pure bending fatigue strength. The predicted values of combined torsion/bending fatigue strength at different number of cycles, using the published failure theory are in good agreement with the experimental data. For the investigated range of fiber volume fractions (V_f) it was found that higher stress levels are needed to produce fatigue failure after the same number of cycles as V_f increases.     

The effect of stress transients on the HCF endurance limit

Constant amplitude fatigue of a material at a fixed stress ratio, R, and at some limiting stress level, may produce high cycle fatigue (HCF) lives in excess of some large number, typically 10⁷ or higher, which can be treated as an endurance limit. Under vibratory loading, stress transients can exceed this endurance limit amplitude and cause damage that accumulates with repeated transient loading. These HCF transients normally occur at lower stress amplitudes than those needed to cause low cycle fatigue (LCF) where lives, N, are typically in the range N < 10⁴–10⁵. Therefore, the HCF transient stresses produce cycles to failure beyond the normal LCF regime but correspond to amplitudes that are above the fatigue limit stress. In this investigation, a titanium alloy, Ti-6Al-4V, is subjected to HCF stress transients while being cycled under constant amplitude HCF. The HCF transients correspond to blocks of loading above the fatigue limit stress applied for a specified fraction of their expected life. A step-loading procedure is used to determine the fatigue limit stress at a frequency of 420 Hz. Stress transients applied at stresses up to 40% above the endurance limit for cycle counts up to 25% of expected life are found to have little or no effect on the fatigue limit stress. Simple calculations of the propagation life in a test specimen show that most of the life at these transient stress levels is spent in the nucleation phase. Fractography, aided by heat tinting, was unable to detect any prior cracks due to the HCF stress transients on the fractured specimens.     

Einfluss von Gussfehlern auf die Dauerfestigkeit von Aluminium‐ und Magnesiumgusslegierungen

Influence of casting defects on the endurance limit of aluminium and magnesium cast alloys The influence of porosity (voids and shrinkage) on the fatigue properties at very high numbers of cycles is shown for the alloys AZ91 hp, AM60 hp, AE42 hp, AS21 hp and AlSi9Cu3 produced by high pressure die casting. Fatigue tests performed with ultrasonic equipment up to 10⁹ cycles show that these alloys exhibit a fatigue limit. The mean endurance limits (50% failure probability) of the magnesium alloys are 8–50 MPa and of the aluminium alloy 75 MPa. Fatigue cracks initiate at porosity, and whether a specimen fractures or not depends on the stress amplitude and the area and the site of the defect. Regarding the cast defect as an initial crack, a critical stress intensity value (K_cr) may be found to propagate a crack until final failure. K_cr of the magnesium alloys is 0,80–1,05 MPa√m, and 1,80 MPa√m was found for AlSi9Cu3. Using K_cr it is possible to correlate the probability of different defect sizes and the failure probability at different stress amplitudes. Additionally, predictions of the influence of rare large casting defects on the endurance limit are possible.     

Tensile fatigue behaviour of ultra-high performance fibre reinforced concrete (UHPFRC)

The tensile fatigue behaviour of ultra-high performance fibre reinforced concrete (UHPFRC) under constant amplitude fatigue cycles is presented. Three series of uniaxial tensile fatigue tests up to a maximum of 10 million cycles were conducted with the objective to determine the endurance limit of UHPFRC that was supposed to exist for this material. The fatigue tests reveal that an endurance limit exists in all three domains of UHPFRC tensile behaviour at S-ratios ranging from 0.70 to 0.45 with S being the ratio of the maximum fatigue stress to the elastic limit strength of UHPFRC. Rather large variation in local specimen deformations indicates significant stress and deformation redistribution capacity of the UHPFRC bulk material enhancing the fatigue behaviour. The fatigue fracture surface of UHPFRC shows features of the fatigue fracture surfaces of steel, i.e. fatigue crack propagation is identified by a smooth surface while final fracture leads to rather rough surface. Various fatigue damaging mechanisms due to fretting and grinding as well as tribocorrosion are identified.     

Estimation of the endurance fatigue limit for structural steel in load increasing tests at low temperature

The endurance fatigue strength of structural steel S355 was investigated in fatigue tests according to the method of increasing stress amplitude. The so‐called ‘load increasing test’ is based on the direct correlation between the fatigue limit and the temperature changes caused by local plastic deformation ahead of the tip of a micro crack, which was initiated as a result of cyclic loading. In the present work the fatigue limit for testing temperatures 40°C and ?20°C was estimated not only from the temperature measurements but also from the electrical potential measurements. Further, the obtained results were validated in standard fatigue tests with constant stress amplitude and a very good agreement was found.     

Initiation of corrosion fatigue cracks at defined grain boundaries in bicrystals

The fatigue tests in air show that persistent slip bands (PSB's) and cracks nucleate very early at special grain boundaries. At stress amplitudes for which no persistent slip band nucleation was observed in single crystals and which were oriented for single slip, cracks still nucleate at grain boundaries. The endurance limit for special bicrystals lay 50% below the endurance limit of polycrystals. In air, the cracks nucleate at the boundary but propagate within the PSB. From the stress amplitudes at which PSB's nucleate in single crystals and in bicrystals, which have one grain with the same orientation as the single crystals, the additional shear stress due to elastic anisotropy was measured to be 55 MPa. These tests yield an understanding of the behavior of short cracks, which nucleate at special boundaries but cannot propagate further, if they hit an unfavorably oriented boundary for which higher local stresses for propagation were needed.With the same bicrystals, corrosion fatigue tests were carried out in ammonium carbonate solutions. In the solution, crack nucleation was found to depend on frequency and amplitude. Contrary to the behavior in air, the cracks nucleate at and propagate along the grain boundary. Specimens which last for 10⁵ cycles in air only survive 6·10³ cycles at the lowest frequencies tested. In addition, if the stress amplitude is reduced by 27%, it was observed that, for a given frequency, the fatigue life is reduced by more than 90% relative to the fatigue life in air at the same stress level. The susceptibility of special boundaries against corrosion fatigue combined with the observed dependencies on stress amplitude and frequency could be understood on the basis of the slip step dissolution model for SCC.     

LONG LIFE SPECTRUM FATIGUE OF CARBON AND STAINLESS STEEL WELDS

Abstract— Fatigue tests of non-load carrying carbon and stainless steel fillet welds have been performed using spectrum loading typical for rail vehicles. The proportion of spectrum cycles exceeding the constant amplitude fatigue limit ranged between 0.86% and 100% and cycles to failure ranged from 4.2 ± 10⁵ to 2.1 ± 10⁷. For the longest tests, the majority of fatigue damage was contributed by cycles with stress ranges less than the constant amplitude fatigue limit. For the carbon steel welds a significant portion of fatigue damage was produced by cycles with stress ranges less than 50% of the fatigue limit but only a small fraction of damage was produced by cycles of this size for the stainless steel welds. The carbon steel welds had slightly better fatigue strength at lives less than 10⁷ cycles but results suggest that stainless steels may have superior long-life variable amplitude fatigue strength when a greater portion of life is spent in the early stages of crack nucleation and growth.     

Beitrag niedriger Lastamplituden zur Ermüdungsschädigung von 0,15 %C Stahl

Contribution of low load cycles to fatigue damage in 0.15 %C steel The S‐N curve of 0.15 %C steel shows an endurance limit. Two‐step variable amplitude loading experiments serve to investigate the influence of numerous cycles below the endurance limit on fatigue damage. If high stress amplitudes of the loading sequences are more than approx. 15 % above the endurance limit, low load cycles contribute significantly to fatigue damage. Investigations of fatigue crack propagation under two‐step variable amplitude loading show accelerated crack growth caused by low load cycles. If high stress amplitudes of the two‐step sequences are less than 15 % above the endurance limit, beneficial influences of numerous low load cycles are found. Under these conditions, the material can sustain far greater numbers of load cycles than predicted by Miner damage accumulation calculation. Fatigue crack growth studies show that under these conditions for the high load numerous low load cycles lead to stop of the crack propagation.     

Fatigue failure of sandwich beams with face sheet wrinkle defects

This paper presents experimental fatigue results for GFRP face sheet/balsa core sandwich beams with face sheet wrinkle defects, subjected to fully reversed in-plane fatigue loading. An estimate of the fatigue design limit is presented, based on static test results, finite element analyses and application of the Northwestern University failure criteria. The presence of a wrinkle defect reduced the fatigue life by approximately 66%, compared to that of an unnotched reference laminate. Furthermore, the results from the fatigue tests revealed that the design limit was initially overestimated, as the specimens loaded close to the predicted design limit typically failed before reaching the target life, or reached test run-out with visible face sheet damage indicating imminent final failure in the worst case. It was found that specimens would reach target life with no visible or otherwise detectable damage by lowering the fatigue load amplitude below 80% of the predicted design limit. By extrapolating the test results it appears that the undamaged specimens would reach a fatigue life of 10⁷–10⁸ load cycles and would thus be safe for design of wind turbine blades.     

Prediction of low-cycle fatigue-life by acoustic emission—1: 2024-T3 aluminum alloy

Low-cycle fatigue tests were conducted by tension-tension until rupture, on a 2024-T3 aluminum alloy sheet. Initial crack sizes and orientations in the fatigue specimens were found to be randomly distributed. Acoustic emission was continuously monitored during the tests. Every few hundred cycles, the acoustic signal having the highest peak-amplitude, was recorded as an extremal event for the elapsed period. This high peak-amplitude is related to a fast crack propagation rate through a phenomenological relationship. The extremal peak amplitudes are shown by an ordered statistics treatment, to be extremally distributed. The statistical treatment enables the prediction of the number of cycles left until failure. Predictions performed a posteriori based on results gained early in each fatigue test are in good agreement with actual fatigue lives. The amplitude distribution analysis of the acoustic signals emitted during cyclic stress appears to be a promising nondestructive method of predicting fatigue life.     

Fatigue failure prediction of a rear axle housing prototype by using finite element analysis

A premature failure that occurs prior to the expected load cycles during the vertical fatigue tests of a rear axle housing prototype is studied. In these tests, crack mainly originated from the same region on test samples. To determine the reason of the failure, a detailed CAD model of the housing was developed. Mechanical properties of the housing material were determined via tensile tests. Using these data, stress and fatigue analyses were performed by finite element method. Fatigue crack initiation locations and minimum number of load cycles before failure initiation were determined. Results provided from tests were compared with the analyses. Design enhancement solutions were proposed to increase the fatigue life of the housing.     

A fracture mechanics approach to estimate the fatigue endurance of welded t-joints including residual stress effects

Residual stresses and weld defects play a major role in the fatigue behaviour of welded structures, so these effects need to be accounted for in a theoretical analysis. A simplified engineering procedure based on linear‐elastic fracture mechanics is applied to estimate the fatigue behaviour, particularly the limit of endurance. Local geometrical irregularities and pre‐existing flaws, which are typical for this kind of weld, are covered by an overall notch intensity factor instead of a specific stress intensity factor, so the initial flaw size is not needed explicitly in the analysis. The effect of residual stresses can be easily included. The cut‐compliance method was applied to measure the residual stress distribution on the cross‐section of the weld. A welded T‐joint was used as a benchmark. Unexpectedly, compressive residual stresses were found to prevail in the root region. According to the analysis, they contribute to the endurance limit of the considered joint by about 50%. This result was confirmed by fatigue tests where a significant decrease in the fatigue strength after a post‐weld stress relieving heat treatment was observed.     

Fatigue behaviour of clinched joints in a steel sheet

Clinch joining has been used in sheet metal work owing to its simplicity and because it facilitates the joining dissimilar metal sheets. In this study, monotonic and fatigue tests were conducted using coach‐peel and cross‐tension type specimens to evaluate the fatigue strength of clinch joints in a cold‐rolled mild steel sheet. The monotonic experimental results reveal that the coach‐peel specimen exhibits the lowest monotonic strength among the three specimen configurations. The coach‐peel and cross‐tension specimen geometries exhibit very low fatigue ratios, compared with the tensile‐shear specimen. The maximum von Mises and principal stresses at the fatigue endurance limit are much higher than the engineering tensile strengths of the steel sheet used to determine the three specimen geometries. Compared with the effective stress and maximum principal stress, the Smith–Watson–Topper fatigue parameter can be used for an appropriate prediction of the current experimental fatigue life. With regard to the coach‐peel specimen geometry, all samples exhibit pull‐out failure mode in the fatigue testing range. However, for the cross‐tension specimen geometry, mixed (pull‐out and interface) and interface failure modes occurred, depending on the number of cycles to failure.     

Very high cycle fatigue of VDSiCr spring steel under torsional and axial loading

Very high cycle fatigue (VHCF) properties of VDSiCr spring steel are investigated with ultrasonic equipment under fully reversed cyclic torsion loading and under cyclic axial loading at load ratios R = –1, R = 0.1 and R = 0.5. Shot‐peened specimens with surface finish similar to valve springs in combustion engines are tested until limiting lifetimes of 10¹⁰ cycles. Under cyclic torsion loading, specimens either fail below 10⁶ cycles with crack initiation at the surface or they do not fail. Under cyclic axial loading, failures above 10⁹ cycles were found for all load ratios with crack initiation at the surface or at internal inclusions. Ratio of mean endurance limit (50% failure probability at 10¹⁰ cycles) under fully reversed cyclic torsion and cyclic tension‐compression loading is 0.86. Cyclic torsion loading slightly below the endurance limit leads to cyclic softening first followed by cyclic hardening whereas cyclic stability is found for tension‐compression loading. Cyclic torsion reduces surface compression stresses whereas they are hardly affected by cyclic tension‐compression loading. Mean endurance limit at 10¹⁰ cycles for R = 0.1 is 61% of the endurance stress amplitude at load ratio R = –1, and for R = 0.5 it is 44% of the tension‐compression endurance limit. Endurance limits for cyclic torsion and cyclic tension‐compression loading are comparable, if effective stress amplitude is used that considers cyclic normal stresses and residual compression stresses at the surface.     

Fatigue limit estimated using finite lives

For metallic material one is interested in estimating the fatigue limit although it is impossible to test if a specimen has infinite life. In this paper two models are compared. One model estimates the fatigue limit and another estimates the endurance limit. The model which estimates the endurance limit is a model which only uses the information that a specimen has broken or not broken at a certain stress level. The model which estimates the fatigue limit also uses the information about when the specimen is broken, i.e., after how many cycles the specimen breaks. The two models are compared, both on real and simulated staircase tests. It is shown that it is possible to estimate the distribution of the fatigue limit although the fatigue limit itself is impossible to observe. When the two models give very different estimates it indicates that the chosen run‐out level is too low.     

Fatigue behavior of the aeronautical Al–Li (2198) aluminum alloy under constant amplitude loading

Tensile and fatigue mechanical behavior of wrought aluminum alloy 2198-T351 is examined and compared against 2024-T3 that is currently used in aerostructures. Experimental fatigue tests were carried out under constant amplitude stress ratio R = 0.1 and respective stress–life (S–N) diagrams were constructed for both alloys. Fatigue behavior of both alloys is described with varying parameters being the percentage of fatigue life as well as the effect of maximum applied stress as a function of ultimate tensile strength. It was found that fatigue endurance limit of AA2024-T3 is approximately 40% below its yield stress, while only 9% below for the AA2198-T351. The latter was found to be superior in the high cycle fatigue and fatigue endurance limit regimes, especially when considering specific mechanical properties. Absorbed energies per fatigue cycle as well as dynamic stiffness of the fatigue hysteresis loop were calculated and plotted against the number of fatigue cycles and with varying maximum applied stress; both parameters are continuously decreasing due to the combination of hardening effect and micro-cracking in AA2024-T3, while this was the case only for the high applied stresses regime in AA2198-T351. Cyclic stress strain (CSS) curves were constructed and proved that work hardening exponent of AA2198-T351 is substantially decreasing with increasing fatigue life.     

The influence of stress-controlled tensile fatigue loading on the stress–strain characteristics of AISI 1045 steel

This study analyses the influence of fatigue loading on the residual tensile properties of AISI 1045 steel. The fatigue tests were carried out under stress-controlled tensile loadings at a stress ratio equal to 0. The maximum applied stresses were within the range from 550 MPa to 790 MPa. An analysis of ratcheting strain and plastic strain amplitude evolution due to fatigue loading was performed on the experimental data. In the next stage of this study, the initial fatigue loadings were introduced. Two maximum stresses, 550 MPa and 750 MPa, and three cycle lengths, 25%, 50% and 75% of the total number of cycles required to fracture the material at a given stress, were used. The pre-fatigued specimens were subjected to tensile testing at strain rates from 10⁻⁴ to 10⁰ s⁻¹. A large number of fatigue cycles, equal to 75% of the fatigue life, induces material softening as well as a drop in elongation and a reduction of area. Pre-fatigue at maximum stress equal to 550 MPa results in the increase of the elastic limit and offset yield point as well. Both parameters reach almost constant value after number of cycles equal to 25 % of the fatigue life. The further increase in the number of cycles does not affect elastic limit and offset yield point in a clearly visible way. The increase of maximum stress of the initial fatigue loadings up to 750 MPa induces similar but stronger effect i.e. increase and stabilization of elastic limit and offset yield point values, however decrease of both parameters value is observed at large number of pre-fatigue cycles corresponding to 75% of the fatigue life.     

Fatigue behaviour of 3-d SiC/SiC Composites

A tension–tension fatigue damage analysis was performed using 3-d silicon carbide fibre reinforced (orthogonal) silicon carbide matrix (SiC/SiC) composites. Two groups of SiC/SiC specimens were tested. The first group consisted of samples without any oxidation protective top layer coating, whilst the latter one contained samples covered with a well fitting, chemical vapour deposited (CVD) SiC system. This coating is necessary for the material to sustain high temperatures. Both the coated and uncoated material had a fibre volume fraction of about 36% equally distributed in three rectangular directions. Load control fatigue tests were conducted at room temperature. The fatigue life was found to decrease by increasing the cyclic stress level. A power-law equation is proposed, which correlates the applied maximum stress during the fatigue test with the number of cycles to failure. In general, the presence of the coating layer decreases the static strength of the material. However, the nominal maximum cyclic stress for which the endurance fatigue limit appeared, remained unaffected by the presence of the oxidation protective SiC coating. Microstructural examination has also been performed on the fractured specimens and it reveals some of the failure mechanisms of the composite that appeared under quasi-static and dynamic loading.     

Experimental investigation of the fatigue strength of plain concrete under high compressive loading

Fatigue tests are conducted on plain concrete cylinders subjected to axial cyclic compression. The upper level of the cyclic stress ranges between 60% and 90% of the static compressive strength. A method based on volume strain measurements is used to predict the individual static strength of each tested sample. The results of fatigue tests are presented in a load versus cycles to failure diagram. The recorded values of the longitudinal strains are, plotted in terms of number of cycles. The test figures are discussed assuming that the failure is a result of both time and cycle dependent damage. For loading levels ranging up to 80% of the static strength, the number of cycles to failure shows hardly any variation with the effects of time. Above this level, the fatigue strength proves to be more sensitive to time dependent effects, as the level of loading increases.