Search

Ductilities generated after four types of creep test are compared: (i) constant load, (ii) constant stress (calculated); (iii) constant stress (from direct observation of the minimum section) and (...

It is shown that the assumption of unit (negative) slope in the well known Monkman–Grant plot of time to failure against minimum creep rate is too restrictive. By acknowledging observed slopes in the range 0.8–1, a ductility–strain-rate relation is deduced where ductility decreases with reducing strain rate. This in turn has implications for the ductility exhaustion method as applied during stress relaxation in the dwell period of low cycle fatigue tests of austenitic steels at elevated temperature. The simple method is used to calculate the cyclic creep damage in typical tests on austenitic steels in the region 550–650 °C and is compared to other calculations as employed in the R5 high temperature assessment procedure. The assumption of a uniform nucleation rate of grain boundary voids with creep strain goes some way to predicting the slope of the ductility–strain-rate relation. Both the ‘unconstrained’ and ‘constrained’ (lower shelf) regions of void growth are discussed.

Strain-controlled (low-cycle fatigue type) experiments were performed on 12 different high temperature alloys (ferritic, austenitic, superalloys, single-crystal material, a directionally solidified alloy and two titanium alloys) in the range 550–950°C. Symmetrical cycling conditions were suddenly changed to offset strain cycling at the same strain range, but in the tension or compression direction. Up to 50 cycles were allowed to determine the subsequent approach to symmetrical stress limits for each material (elastoplastic shakedown). Comparison is made with limited reported tests and related load-control tests involving strain ratchetting . Shakedown behaviour is shown to depend both on the path of approach, defined by stress–strain properties, but overall trends of stress versus cycle plots are best made by assuming an effective strain range which is less than the imposed experimental strain range.

Given total endurances of low cycle fatigue specimens, several empirical formulae enable an estimate to be made of ‘initiation’ cycles to a specific crack depth in a zone ranging from 20 to 400 μm from the surface. For greater depths, the cyclic growth behaviour of cracks can be calculated directly from relations using established constants (including the effects of creep dwell) for a wide range of alloys. However, below a critical depth of 200 μm or thereabouts, these empirical growth laws, expressed in terms of total strain range, break down. Early crack growth can take place discontinuously, as the crack tip encounters grain boundaries or other obstacles. Such information arises from replica takings, or in situ microscopic observations of surfaces during continuous cycling and creep–fatigue tests. Striation spacings on the fracture surface and the first observed striation spacing provide valuable indications of early crack growth. In this region, the cyclic rate of growth depends on (i) crack s...

Reversed-load (low-cycle-fatigue type) experiments were carried out on 13 different high temperature alloys (ferritic, austenitic and superalloys, single crystal materials, a directionally solidified alloy and two titanium alloys) in the range 400–1025°C. By means of interrupted tests in the steady state, the asymmetrical nature of the Bauschinger effect was examined, i.e., having regard to the magnitude and direction of the (prior) stress reversal during the course of the hysteresis loop. The cyclic stress–strain curve was characterised using the Ramberg–Osgood power deformation relation. The yield stress in a given direction during the course of a cycle could be made to vary according to the reversal point in the interrupted tests. A simple interpretation is provided in terms of the prevailing difference between the back stress and a friction stress. Comparison of the back-stress behaviour is made with that calculated from the widely used Chaboche relation. Predictions of yield stress behaviour ...

This paper attempts to link the energies to failure involved in low-cycle fatigue at elevated temperature with those involved in forward creep. These energy levels are connected by a suitable sliding strain-rate scale. The principle is then applied to (i) dwell tests on 1CrMoV, 304 and 20Cr25Ni/Nb steels at 550, 650 and 750°C respectively and (ii) service failure of a thick-walled ferritic steam vessel at 570°C. Calculated creep-fatigue endurance using the linear damage summation rule is then plotted on Damage Diagrams. Sensitivity analyses are carried out as regards the linking of energy levels in fatigue (upper-shelf) with those in creep (lower-shelf). These involve knowledge of creep laws or stress relaxation relations and the methods of connecting energy levels (power law or those giving an S-shaped curve). Within this paper, sufficient information has been provided for investigators to undertake their own analyses on relevant laboratory tests or service assessments.

This paper considers the likely strain concentration effects generated in the shoulder/gauge intersection region of typical specimens employed in low cycle fatigue tests at high temperatures. The starting point is taken as the well known elastic stress concentration factor curves of Peterson which are extrapolated using data in the literature and other supporting (finite-element) calculations. Elastic–plastic values of strain concentration factors are then derived using the relations due to (a) Neuber, (b) Stowell–Hardrath–Ohman and (c) an energy method. It is shown in each case that these factors decrease with evolutionary cyclic hardening and increase with cyclic softening in a systematic way, as the material behaves according to the Ramberg–Osgood deformation law. A sensitivity study is carried out for each method, both for a ‘model material’ and real alloys and the implications of the shoulder radius/gauge diameter ratio is considered against the recommendations of various Testing Standards an...

Cyclic stress-strain tests were undertaken at 550 °C on plain specimens and notched specimens of different notch acuities in several low- and high-alloy ferritic steels (1Cr-Mo-V, NF616, TB12M, and HCM12A). Integrated axial strains were measured between the minimum sections of the notches using a longitudinal extensometer, while surface hoop strains were measured by means of a diametral extensometer with probes located across the notch root. The same extensometry was employed in plain specimens. Over a period of 100 cycles, softening occurred in all plain specimens. These effects were also demonstrated in notched specimens in both axial and diametral directions, although to a lesser degree. From hysteresis loops determined in the notch tests, the cyclic deformation response of the notched regions was expressed in terms of an ‘equivalent gauge length’. Deviations from elastic-plastic behaviour in plain specimens were noted in that the commonly used ‘effective Poisson's ratiO' was greater than calculated. The effect was investigated further by exploring the characteristics of a very shallow notch, induced by straining a plain specimen to the onset of necking and beyond. The implications of observed behaviour in strain-control low-cycle fatigue tests is discussed.

This paper investigates the importance of multiaxial stress states by considering several distinct testing techniques used in assessing both creep deformation and creep damage accumulation. The requirements of testing programmes to determine the necessary data are discussed in respect of sensitivity and interdependence of the principal and hydrostatic stress ratios.

This paper is concerned with the laboratory isothermal (low cycle) fatigue crack growth properties (generally at 550°C) of eleven ferritic steel samples taken from power station components which have seen between 63,000 h and 175,000 h service exposure to high temperatures. Full chemical analyses and abbreviated service histories of the materials are presented. Both the short crack (LCF) and deep crack (LEFM) regimes are examined. In some cases it was possible to continue the growth of actual service cracks under laboratory conditions, by suitable machining of samples into specimens; in other cases it was necessary to start the crack artificially. It is intended that the results can be used in direct application to remanent life assessments. In many tests the creep-fatigue interaction was simulated by the addition of a 0.5 h tensile dwell, it being assumed that major damage was already present in the extracted samples. The short crack growth tests at 550°C with dwell have demonstrated that even wh...

Reheat cracking in the heat affected zones of austenitic stainless steels can occur during high temperature service and is thought to arise from the relaxation of welding residual stresses, resulting in creep damage. Key features of reheat cracking are the magnitude of the residual stresses, degree of triaxiality, extent of stress relaxation behaviour and the influence of triaxial stress on creep ductility. A thermo-mechanical pre-conditioning technique has been developed for testing specimens to represent the welding cycles experienced between 200 and 1100 °C by the ‘strain affected zone’ of a weldment in a steam header during manufacture. Subsequent creep tests have been carried out at 550 °C on pre-conditioned plain and notched specimens to identify the behaviour as regards creep rate and ductility in both the uniaxial and triaxial stress states. The results have been compared with data in the literature and with models of the influence of multiaxial stress on creep ductility. It has been found that certain types of pre-conditioning give a marked reduction in creep ductility as strain rate is reduced, which helps to explain the reheat cracking process.

This paper is concerned with the laboratory isothermal (low cycle) fatigue response of sixteen ferritic steel samples taken from power station components which have seen between 70,000 h and 170,000 h service exposure to high temperatures. After suitable machining into specimens the cyclic stress-strain properties between 300 and 600°C at strain rates between 6 × 10−4/s and 5 × 10−6/s were established using the multiple step technique (which is discussed in some detail). Ferritic steels generally cyclically soften, and by evolutionary tests undertaken at 550°C it is demonstrated that start/up and shut/down procedures in service have accelerated this process in material subjected to temperature transients. The steady-state cyclic stress-strain response has shown a marked loss in strength compared with data on a selection of unexposed steels e.g. material trepanned from turbine casings at the manufacturing stage or from specially cast test blocks. The data have been fitted to the Ramberg–Osgood law ...

Cyclic stress–strain tests were undertaken at temperatures between 550°C and 820°C on plain and doubly-notched specimens of three alloy steels: 20Cr/25Ni/Nb, 316 and 1CrMoV respectively. Integrated axial ‘strains’ were measured between the minimum sections of the (semi-circular) notches using a longitudinal extensometer, while surface hoop strains were measured on one notch by means of a diametral extensometer. Cyclic hardening occurred in plain specimens of 316 steel, cyclic softening occurred in the 1CrMoV steel and the 20Cr/25Ni/Nb alloy showed stable behaviour. These effects were also demonstrated in notched specimens, although to a lesser degree. From hysteresis loops determined during multiple step tests, the cyclic deformation response of the notched regions was expressed in terms of an ‘equivalent gauge length’. Comparison was also made between the equivalent (von Mises) stress–strain curves deduced at a representative (skeletal) point in the minimum section with stress–strain data obtaine...

Comparisons have been made on evolutionary and stabilised cyclic stress-strain behaviour following different prior ageing treatments in a 316 steel. Prior stress-free ageing at a temperature higher than the test temperature reduces the ability of this steel to cyclically harden and consequently the material does not achieve maximum strength. An apparent saturation of cyclic hardening at room temperature can however, be continued by further cyclic hardening at 625°C. The rate of precipitation during low cycle fatigue can be two orders of magnitude greater than that occurring during stress-free ageing at 625°C. Quantitative measurements based on an extraction technique are presented and compared with particle size measurements in the literature. Prior high temperature low cycle fatigue exposure has been shown to reduce subsequent creep rates with a concomitant reduction in ductility: the same effect is found by prestrain at room temperature. Neither cyclic stress-strain properties nor creep behaviou...

Data from a previous round robin low cycle fatigue exercise on 316L steel, 9Cr1Mo steel and IN718 alloy at 550°C and Nimonic 101 at 850°C are re-examined, emphasis being placed on material stress-s...

Strain-controlled total endurance tests have been carried out on smooth specimens of 2 1/4Cr-1Mo steel at 550°C in continuous cycling, and also using interspersed blocks of major and minor cycling. Similar low-cycle fatigue crack growth tests were also carried out on cylindrical specimens of 1/2CrMoV steel containing a 0.2 mm deep starter defect using interspersed blocks of major, intermediate and minor cycling, their order being changed in separate tests. Damage summation in the smooth specimen tests at failure was found to be less than unity (irrespective of the failure criterion) when Miner's rule was employed, and the results are compared with limited previous data. Damage was also summed on an energy-expended basis and this was found to be (i) independent of strain range, and (ii) independent of cycle type (i.e. continuous loading or block loading). The nearer the chosen failure criterion to crack initiation, the closer was the agreement in accumulated energies. Similarly, in the crack growth tests, the number of complete loops to failure (10% load drop) was independent of block order over a wide range of applied strain ranges, but was affected by the number of cycles in each block. However, the energy criterion was again able to rationalise all the data.

Isothermal uniaxial cyclic loading experiments have been conducted on specimens of the nickel-based superalloy IN100 at selected temperatures in the range 600–1000°C. Power-law stress—strain relations based on plastic strain (Ramberg-Osgood), total strain and their bilinear equivalents have been used to describe the behaviour. In addition, uniaxial thermomechanical fatigue (TMF) tests have been performed on tubular specimens of the material over the temperature range 650–1050°C. The aim of these tests was to model the response during thermal shock experiments involving rapid heating and cooling cycles between 200°C and 1100°C. Some transient tests have also been carried out on blade-shaped single-wedge specimens of IN 100 to simulate the type of thermal fatigue (TF) cycle experienced by turbine blades during service. A finite element sensitivity analysis has been undertaken to predict the results of both the TMF and TF tests using the two different bilinear descriptions of the isothermal cyclic stress-str...

High strain fatigue experiments were carried out on thirteen different high temperature alloys at 550°C. By means of multiple step tests and interrupted tests in the steady state, the asymmetrical nature of the Bauschinger effect was examined, i.e. having regard to the magnitude and direction of the stress reversal point during the course of the hysteresis loop. The cyclic stress–strain curve was characterised using the well known Ramberg—Osgood deformation relationship, and distinction was made between actual curve shape and the locus approximation. The yield stress during the course of a cycle could be made to vary according to the reversal point in the interrupted tests, but even in steady state cycling, the cyclic yield point was found to decrease as the total strain range was increased. Limited investigations were also performed during evolutionary (hardening or softening) cycling and all results were interpreted in terms of the prevailing back stress and friction stress. A formula to predic...

— Thermo-mechanical fatigue stress-strain data on ferritic/austenitic steels and superalloys from a variety of sources are analysed with regard to hysteresis loop stress asymmetry. This arises from a decoupling of the thermal and mechanical strain signals in the test technique so that many tension-compression load combinations are possible. Data from simplified isothermal and bithermal tests are also examined. Taking a typical example of an “out-of-phase” thermo-mechanical loop on a 1/2 CrMoV steel cycled between 200 and 550°C, isothermal stress-strain data were generated at 50°C intervals on material from the same cast and, used in conjunction with the elastic characteristics of the apparatus, an attempt was made to re-create this loop. The methods employed were (i) a graphical construction between appropriate isothermal yield contours (ii) a tangent modulus calculation (iii) a secant modulus calculation. Method (i) appeared to give the closest agreement in the present case.

An earlier modification of the Paris law for the growth of deep cracks in the linear elastic fracture mechanics regime is extended to include a term enabling the prediction of cyclic crack growth rates at low frequencies. The relation requires (i) a reference growth law under continuous cycling at the appropriate elevated temperature and (ii) a specified, dimensionless degradation term, defined as Dc, the creep/oxidation damage per cycle, which increases as the applied frequency decreases or as the dwell time at peak load is prolonged. The relationship is validated against data from the previous analysis on low alloy ferritic and austenitic steels in the range 538–650°C and against further published results on Ni-based alloys at temperatures up to 700°C. It appears that for the former series oxidation is the dominant damaging mode, whereas a linear creep damaging mechanism is manifest in the Ni-based alloys. Moreover, levels of cyclic damage in terms of Dc are higher in the latter, ranging between 10−3 an...

High temperature total endurance tests were conducted on two alloys (316L steel and 9Cr–1Mo (wt-%) steel) at 550°C in high strain fatigue under continuous cycling at total strains in the range 0·6–2·0%. The austenitic steel cyclically hardened, whereas the ferritic steel cyclically softened. From striation spacing measurements on fracture surfaces (and to a lesser extent from observations of tensile load drop), it has been possible to correlate progress of the major crack with strategic points on the hardening/softening curves at each strain range. The minimum depth at which striations were measured with confidence was ~0·04 mm and by establishing growth laws it is demonstrated that this depth was typically achieved by the ‘plateau’ regions for both alloys. By extrapolation of the growth relationships it is concluded that cracking ~4–20 μm deep has occurred by the end of the earlier rapid hardening and softening phases.MST/1813

High temperature total endurance tests were conducted on three alloys (316L steel at 550°C; 9Cr–1Mo steel at 550°C; Nimonic 101 at 850°C) in low cycle fatigue under continuous cycling. The austenitic steel was observed to cyclically harden, whereas the ferritic steel and the superalloy both cyclically softened. Methods for rationalising evolutionary behaviour were variation of percentage hardening/softening with strain range, cumulative ductility criterion, variation of secant modulus in the hysteresis loops, and cumulative energy criterion. The energy expended per cycle appears to be the most attractive. For a given strain range this was approximately constant at any stage of evolution. More significantly, the cumulative energy at saturation was constant, approximately independent of total strain range, and about 0·5–3·0 J mm−3 according to the material used. The Palmgren–Miner hypothesis (which is widely used in summing fatigue damage) was originally derived on an energy argument and it is propo...

A simple TMF facility is described which is capable of 'in-phase', 'out of phase' and 'diamond' type strain-temperature cycling. Typical TMF loops are presented for the 'diamond' type cycle for the nickel-base alloys IN100, Nimonic 90, IN738 and directionally solidified CM247LC-DS over the temperature range 400 °C-1000 °C. Alongside these results, isothermal cyclic stress-strain data are provided for ascending and descending strain amplitudes at discrete temperatures encompassing the strain and temperature range of the TMF tests. Testing on any individual specimen was terminated after a fixed energy consumption. By superimposing the symmetrical isothermal loops about a common origin, the isothermal data are then used to predict the observed (asymmetrical) TMF behaviour. The method works well for materials which show history independence such as IN100. However, for some of the materials, slight modification to the procedure is required. The sensitivity of the method to the form of isothermal data is demonstrated.

This chapter follows the course of a crack in a typical component from the initiation and short crack growth stage, through to the deep crack growth stage and on to the possibility of complete penetration across the wall thickness. The causes of such growth are considered, such as thermal shock and other constraints against expansion or contraction and the many ways of simulating propagation behaviour in the laboratory are discussed, where a cyclic event in service is identified with a fatigue cycle performed in the laboratory. Parameters which are used to describe crack growth in the various regions are explained, together with methods of accounting for internal structural damage in the material (‘creep-fatigue interaction’) which is observed to enhance crack growth rates. Many worked examples are given, either to illustrate a technical point or based on service experience. Finally, a complete case study (retrospective analysis) of crack propagation across a component in power plant is undertaken together with a validation of the calculations.

J. Ginsztler: Lifetime Assessment at High Temperatures.- R.P. Skelton: Crack Initiation and Growth during Thermal Transients.- E. Czoboly: Basic Mechanisms of Creep and the Testing Methods.- R.A. Ainsworth: High Temperature Defect Assessment Procedures.- K.M. Nikbin: The Fracture Mechanics Concepts of Creep and Creep/Fatigue Crack Growth.

Low cycle fatigue (LCF) endurance data have a valuable part to play in the lifetime assessment of components and structures. These data comprise the initiation and growth stages, but the growth relations themselves and their practical use are not as familiar as those employed for deeper cracks. Early work modelled continuous-cycling fatigue crack growth by assuming a succession of miniature LCF specimens at the crack tip, the field then being extended by investigators examining behaviour at high temperatures. Models were developed allowing for the concomitant contribution of creep damage for comparison with continuous-cycling properties where striation spacings recorded cyclic crack progress. Alongside such modelling, empirical laws were deduced describing the progress of short cracks. Expressions may be derived linking LCF with linear-elastic fracture-mechanics (LEFM) crack growth, using the parameter ΔJ (equivalent stress-intensity parameter). However, the purpose of this review is to compare and contrast those models which employ an easily measured surface parameter (such as total or plastic strain range) as the governing variable. Crack growth normally adopts an exponential form so that the rate of growth per cycle accelerates as the crack deepens. The distinguishing feature is that the process zone at the crack tip is itself surrounded by cyclically yielding material, in contrast with LEFM. Energy methods may also be employed, where the process zone at the crack tip fails when the accumulated energy density reaches a critical value. An upper bound relation is provided, accounting for the deleterious effects of creep-fatigue-oxidation interaction, if empirical data are not to hand. A conservative assessment may thus be made of cyclic crack growth rate at a specified depth. This review examines the capability of each model to allow for such creep-fatigue effects.

This paper attempts to assess the contribution of oxidation to crack growth during high strain fatigue of a Cr-Mo-V steel at 550°C from three separate oxidation studies. Stress-free data are clearly inapplicable to the highly strained crack tip and so it is first shown that internal stresses, as in bainitic material, promote enhanced oxidation. Weight gain experiments on oxide-free fracture surfaces are next described. It is shown that the stored energy of fatigue fracture likewise causes an increase in oxidation rate but that it is difficult to simulate the process at the tip of an advancing crack. Similarly, metallography of oxide in fatigue cracks does not reveal propagation. history. Finally, a dynamically worked surface, provided. by high-strain fatigue specimens deforming at 550°C, causes increasing oxidation with increasing total strain above a threshold value of oxide strain. It is shown that this is due to layering producing at least a tenfold increase compared with stress-free specimens....

— A method of calculating crack growth rates in high strain fatigue from total endurance data is presented. Endurances are known to be affected by dwell periods and environment and the expression derived takes these into account. The basic argument is that crack propagation may be regarded as increments of successive reinitiation and so the deeper a crack grows, with an accompanying increase in strain concentration at the tip, the fewer fatigue cycles are required for the next step. Previously derived expressions for strain concentration are employed which require a knowledge of the cyclic strain hardening properties of the material. These are produced in detail for a cast 1/2 Cr-Mo-V steel at 550°C. Two constants in the resulting propagation expression are semi-empirical, but if the surface (plastic or total) strain range is known then cyclic crack growth rates may be calculated for any crack depth. It is also shown that in the limiting case of zero surface plastic strain, the expression reduces to that observed for crack growth under linear elastic conditions.

In this investigation small, cylindrical specimens of 316, 9CrlMo and 12CrMoV steels and alloy 718 were subjected to a ‘ranking test’ whereby some 30 000 thermal quench cycles were imposed and ultimate crack penetration determined metallographically. Further tests on large specimens of 316 and 1/2CrMoV steels are reported where fewer cycles were involved. Crack progress was also assessed from striation measurements which furnished evidence ofacceleration away from a notch and ofsubsequent arrest. It is concluded that under the conditions oftesting, 316 steel shows the least resistance to crack penetration, the ferritic steels demonstrate intermediate behaviour while alloy 718 is most resistant to crack initiation.

The effect of R (minimum load/maximum load) ratio on fatigue crack growth rates in tests on cast CrMoV steels at 550°C was investigated in the nominally LEFM range. A tensile dwell of 1/2 h was introduced in each cycle, some tests being load controlled while others were displacement controlled. Most tests began at an R value of -1, which thereafter varied with the circumstances of the test (i. e. remained cpnstant in load control, but decreased in displacement control). Load/displacement loops also indicated that the crack could remain open part way into compression. For these reasons the crack propagation results could not be described adequately by using just the tensile loading to calculate the stress intensity amplitude ∆K.It is shown that by calculating the stress range for which the crack is open (which also varied during the test) and making further corrections for (a) ratchetting, which occurred under load control (b) reversible plastic displacement, which sometimes occurred later in a tes...

The variation of the grain boundary self-diffusion coefficient with applied stress and the production of vacancies in the lattice are applied to the case of cavity growth in high temperature fatigue with particular reference to the case of magnesium at 400°c. It is shown that cavity growth should increase with increasing testing frequency and that there is a critical condition for growth to take place. An outline is given of a theoretical estimate of times to fracture and it is shown that the time of testing, rather than the total number of cycles, becomes the important endurance parameter. Predictions of the model are compared with experimental findings.

The effect of grain size on the high-temperature fatigue properties of an Mg–0·8% Al alloy has been studied. A reduction in rupture life occurred when the grain size exceeded the specimen thickness. It is shown that this effect is related to an increase in the initial plastic strain amplitude, caused by an increase in the rate of grain-boundary sliding. It is also established that this deformation mode controls cavity growth and that failure occurs at a critical mean value of the sliding displacement, independent of applied stress and structure.

High-frequency (10–1700 c/s) push/pull fatigue tests on a magnesium alloy at 0.75 of the melting point show that the rate of fatigue-hardening generally increases with increasing frequency and stress level. The time to onset of major crack propagation is virtually independent of frequency, though decreasing slightly at the higher frequencies. Beyond this point, the rate of crack propagation may be either time- or cycle-dependent, according to stress level. Fatigue specimens show marked grain growth at a rate that increases with increasing stress level. High-frequency tests are associated with less growth and enhanced boundary cavitation. Smaller grains give rise to less overall plastic strain than large grains and evidence is presented of grain-boundary sliding during fatigue.

On the basis of a progressive nucleation of voids with secondary creep strain, an expression is derived for the elongation to failure for specimens experiencing grain boundary cavitation during creep. The effect of temperature and strain rate on ductility is discussed.

High strain fatigue endurances of a 20%Cr/25%Ni/Nb stabilized stainless steel at 1025 K (750° C) in CO2, air, and vacuum at strain rates 10−3–10−5s−1 are presented. In vacuum tests, bulk density decreases, cavitation and internal cracking, and general structural changes were observed only at high plastic strain ranges (∼0·08). At low strain ranges (

Some fatigue lives of a magnesium-0.6 wt.-% zirconium alloy at 430°C and an alternating stress of 1400 lb/in2 are presented. Homogenizing and/or precipitation heat-treatments carried out on hot-worked material are shown to be generally deleterious to fatigue endurances. The effects of heat-treatments are assessed in terms of observed microstructure. It is revealed that high fatigue strengths are incompatible with high creep strengths in certain circumstances. The likelihood of solute diffusion being enhanced by fatigue and the growth of grain-boundary denuded zones are discussed.

A recently proposed model of diffusional creep strain caused by atoms ‘plating-out’ along grain boundaries during void growth is extended by assuming that voids nucleate uniformly with time. Calculated strains at fracture are compared with those resulting from a constant number of voids.

This chapter reviews an important aspect of high strain fatigue testing, namely the cyclic stresses generated when experiments are conducted at a given total strain range. It has in mind the design engineer, who may wish to use more sophisticated data in an analysis, and the materials scientist who has before him a wide open experimental field at elevated temperature. After an introduction to the basic concepts of cyclic hardening or softening, there follows a section on definitions of parameters used throughout the chapter, particularly the Ramberg-Osgood relation as modified for the cyclic case. Testing methods are then briefly examined and alternatives such as constant plastic strain range experiments considered. The multifarious ways of presenting cyclic stress-strain data are next investigated and a ‘best’ (i. e. least confusing) method recommended.

The smooth cylindrical fatigue specimen provides the starting point for most investigations in high strain fatigue. If it is to be used for a determination of cyclic stress-strain properties or a metallurgical investigation of microstructural changes or for crack morphology studies at elevated temperatures, one need enquire no further. However, if specimen endurance is to be related to component behaviour then size effects become important. Further, the traditional laboratory simplification of testing between fixed strain limits warrants closer examination when certain stresses encountered in service are being simulated.

(1985). High Temperature Materials — Modelling Service Behaviour. High Temperature Technology: Vol. 3, No. 4, pp. 225-226.

When a relation for void growth by vacancy diffusion during creep is combined with experimental data on bulk density changes, it is possible to assess void population at any stage of creep. The analysis is extended to estimate fracture times and comparison is made with previously published experimental results.

The method of estimating activation energies of void; growth during creep from the plot of ln (rupture time) vs. reciprocal temperature is critically discussed.