Abstract
EPRI has been providing technical support to key Global partners in the Electricity Supply Industry for over 40 years. In the Generation Sector, a key research imperative is knowledge creation and technology transfer linked to reliable, safe and economic flexible operation of power plants. The present paper outlines achievements related to assessment of transients in modern generating plant, highlights particular forms of damage and outlines considerations for additional research. In many cases, this future work will consider the behavior of creep strength enhanced steels. These steels, typically based on 9 to 12% Cr, offer significant benefits to the design and fabrication of components in high efficiency fossil fuelled plants because, when properly processed, tempered martensitic steels offer an excellent combination of strength and toughness. EPRI collaboration has been focused on root cause assessment so that lessons learned can be used to establish best practice. EPRI facilitated activities have included annual creep–fatigue discussions, publication of summary documents and additional research. Excellence in science and engineering is necessary to underpin technology which will help to meet challenges associated with safe and reliable operation of plant.
Introduction
The increasing need for flexible electricity generation means that most components must perform under cyclic operating conditions [1–3]. The multiplicity of generating options also means that defining a ‘typical’ cycle is difficult. Thus, in addition to the traditional hot and cold starts and stops, there are requirements for units to rapidly change load, on many occasions followed by operation at generation levels of 30% of rated capacity [3]. These transients, and the operation at low loads, result in complex changes in local pressure, temperature and flow at different parts of a boiler system. Indeed, these differences, in combination with the use of metallurgically complex steels, are such that the range of potential damage mechanisms and ‘at risk’ locations has increased. It is also clear that as cyclic operation increases the number of failures and forced outage rates increase, Fig. 1.
Trend curves showing that there is an increase in the rate of failures of boiler components as the numbers of unit starts increases [1]. (Colors are visible in the online version of the article;
A photomicrograph showing an example of creep and fatigue damage which developed independently in a low alloy steel pressure boundary weld.
Understanding the factors affecting the mechanisms involved in all of the failures is not straightforward. Furthermore, over the last ten years there has been a marked reduction in published results for post service failure investigations and root cause analyses. It is therefore increasingly difficult to obtain full information regarding the factors affecting damage and thus comprehensive analysis of problems and solutions is impossible.
As utilities continue to adopt generating practices that involve flexible operation the number, magnitude, and complexity of the cycles associated with transient type operation will all increase. Thus, in addition to starts and stops, changes in generating output can lead to problems associated with thermal and/or mechanical loading as well as potential issues with water chemistry and corrosion. For components which operate at high temperature, damage associated with transient operation is frequently called creep–fatigue. However, even within this group of components, specifics of damage mechanisms in various components can vary widely. Thus, review of cracking in different components demonstrates that fatigue dominated or creep dominated damage can occur depending on specific operating conditions, component geometries and steel types. As shown in the photomicrograph in Fig. 2, it is apparent that both of these mechanisms can take place without interaction. Thus, the fatigue crack which initiated at a stress concentration at the inside surface has been growing in a transgranular manner. However, the transgranular defect had extended to the point where further propagation will occur in a region which has sustained significant intergranular creep. Thus, early in the damage process the mechanisms were independent (or even competitive) but later in life fracture can be accelerated due to the presence of microdamage. The influence of the role of creep microdamage on reducing fracture toughness and promoting burst rather than leak type fractures has been widely discussed. Moreover, the susceptibility of a steel to the nucleation, growth and link up of creep voids will be a critical in assessment of the performance of high energy components and analysis of expected fracture behaviour. The present paper outlines achievements from EPRI collaborative work on creep–fatigue which has been on-going for more than ten years and illustrates some of the most recent trends and results of high temperature tests on tempered martensitic steels. The implications of this latest analysis on the design, material specifications and in-service performance of high energy components are discussed.
The long-range research program at EPRI (known as Technology Innovation) initiated a set of activities in 2006 related to assessment of the behaviour of high energy components operating under cyclic conditions. The initial goals of the work were to:
Review international approaches for Creep–Fatigue assessment,
Identify deficiencies with current knowledge, materials data and methods for testing and analysis, and
Recommend improvements in the application of the available technology and identify future Research and Development needs.
A key achievement of these discussions was to agree an overall approach to establishing solutions for assessing creep–fatigue performance and to formalize collaborative activities through:
facilitating annual expert workshops,
the preparation of critical reports and publications,
drafting and review of new ASTM standards for laboratory testing linked to creep–fatigue crack initiation and for crack growth, and
establishing a knowledge base of materials behavior directly relevant to assessing behavior.
The overall success of this effort is the direct results of international experts being prepared to collaborate with ideas, resources and expertise for the benefit of the overall group and the industry as a whole. Participants have also been involved in the preparation and discussion of review documents which are important to efforts for modernization of Codes such as ASME. Important elements of these achievements are outlined here with appropriate references which should allow access to the original source documents.
Expert workshops
Annual expert creep–fatigue workshops
Annual expert creep–fatigue workshops
A key topic for discussion at these expert meetings has been consideration of different methods used in the evaluation of creep crack initiation for various engineering alloys. The most widely used overall approaches in industrial assessments are ASME documents, Rules for the Construction of Nuclear Facility Components, Class 1 Components in Elevated Temperature Service, Boiler and Pressure Code, Section III, Division 1 – Subsection NH [4], publications linked to TRD 301, Annex I – Design: Calculation for Cyclic Loading due to Pulsating Internal Pressure or Combined Changes of Internal Pressure and Temperature [5], The RCC-MR documents, Design and Construction Rules for Mechanical Components of FBR Nuclear Islands, Section I [6] and The British Energy procedures known as R5, An Assessment Procedure for the High Temperature Response of Structures [7].
The different approaches can lead to very different estimates of creep–fatigue behaviour of components [8]. For example, the recommended interaction line for Grade 91 steel originally proposed in ASME – NH [4] was very conservative. At least some of this excessive conservatism was directly linked to the fact that Grade 91 steel exhibits complex behavior and creep softening during cyclic tests at high temperature. Moreover, use of time fraction rules without proper consideration of stress–relaxation and strain–softening effects gave a conservative outcome. The level of conservatism using the standard ASME NH approach is demonstrated by consideration of Fig. 3(a).
The lives calculated using the original ASME NH methodology [4] are less than the observed behavior by more than an order of magnitude. Application of an improved methodology, such as that developed in the French Code RCCMR [6], indicates that by accounting for at least some of the metallurgical complexities, the prediction comes into reasonable agreement with the observed behavior, see Fig. 3(b). Current research shows that the creep cavitation susceptibility varies considerable for Grade 91 steels depending on factors such as the levels of impurity elements present and details of the steel making and component processing methods [9]. The results shown in Fig. 4 are a compilation of data showing the relationship between minimum creep rate and reduction of area for different Grade 91 steels, tested at temperatures in the range 500°C to 650°C. It is immediately apparent that for creep samples which exhibited the same minimum creep rate, fracture took place with very different ductility values.
Comparison of trends between the creep rate and reduction of area for creep tests on a wide range of Grade 91 steels. (Colors are visible in the online version of the article;
The EPRI initiatives related to transient performance of high energy components have focused effort on different specific topics. A summary of the Titles of the Milestone reports produced as part of these activities is provided in Table 2.
Other reports have documented the results of specific experimental research projects which have evaluated cyclic performance in selected alloy systems. These include: Investigation of the Creep–Fatigue Behavior of P91 Steel, reference 1024718 Program on Technology Innovation: Evaluation of the Creep–Fatigue Behavior of Grade 92 Steel and Its Predictability, reference 1024588 EPRI Milestone Reports
The results of these specialist tests provide important direct information on how specific metallurgical factors influence key aspects of performance. A particularly important activity is carrying out post-test characterization of damage. In many cases, the differences in creep–fatigue damage behaviour observed for similar tests on Grade 91 steels can be explained differences in the creep ductility shown in Fig. 4. As previously discussed, Grade 91 steels which are supplied as fully compliant to specifications can exhibit dramatically different fracture behavior [9]. These differences, which stem directly from variations in the susceptibility to form creep microvoids, influence the level creep–fatigue interaction observed [10].
Variations in damage processes for specific Grade 91 steel casts further complicates analytical assessments of creep–fatigue interaction as in most cases the analysis performed seeks to consider only average material behavior. It is apparent that the two main methods used to asses creep damage are:
Time fraction
Strain fraction (ductility exhaustion)
Whether time summation or strain summation is chosen as the basis for analysis, investigators usually seek to demonstrate an agreement between predictions and experimental data to within a factor of two. Thus, although for some steels metallurgical variations result in significantly greater differences in actual high temperature behavior, the analysis methods usually ignore these ‘complications’. Indeed, because post-test metallographic characterization is often ignored [11], judgements over relevance of expressions are impossible.
While these general terms describe the overall methodologies, there are many variations in the detailed applications. The variations, and the assumptions that are often required to support the more detailed analysis, frequently make definitive engineering judgments on accuracy difficult. However, requirements for additional complexity means that the models begin to lose their attractiveness – the most robust have proved to be those which are the easiest to use.
One particular challenge when using strain fraction approaches relates to knowledge of ductility. One basic question must be asked – which is the correct ductility to take for the given test conditions? As shown in Fig. 4, although ductility as measured by Reduction in Area for Grade 91 steel decreases with decreases in strain rate there is very large scatter in the results. In many cases, however, the time fraction rule is often applied because of the paucity of accurate creep ductility data.
Guidelines and procedures for creep–fatigue testing have been established in many countries. However, there are differences in detail between many of the recommendations. Moreover, where tests are carried out on welds, it appears that despite the heterogeneity of structures, samples are generally dimensionally similar to those used for parent. This decision is normally taken to accommodate test rig requirements etc. Some general points regarding creep–fatigue testing include the following:
Ensure the accuracy, stability and calibration of equipment,
To reduce the scatter of the data, the tests should preferably be carried out under axial strain control, isothermal conditions, and using cylindrical smooth specimens,
Tension hold time tests are useful for ferritic steels as well as austenitic steels to simulate the thermal fatigue, and
Ductility normalized strain range partitioning methods are useful because the lives depend on the ductility and the strength of the materials.
A key activity facilitated by the EPRI support was the development of specific ASTM Creep–Fatigue Standards, namely:
ASTM Standard: Test Method for Creep–Fatigue Testing, ASTM E2714-09 [12]. This method covers the determination of mechanical properties pertaining to creep–fatigue crack formation in nominally homogeneous materials. It is primarily aimed at providing the material properties required for assessment of defect-free engineering structures containing features that are subject to cyclic loading at temperatures that are sufficiently high to cause creep deformation.
ASTM Standard: Test Method for Creep–Fatigue Crack Growth Testing, ASTM E2760-10. This test method is concerned with developing creep-crack growth data under cyclic conditions which is used in some more sophisticated assessments of in-service materials when large flaws may be present.
As part of the review and acceptance process of the provisional standards listed work was required to evaluate for precision and bias statements. This work involves conducting round robin test programs. The round robin testing program is complete for ASTM E2714-09 [12] (details below) and the precision and bias statement will be added in the next revision of the standard. Planning, preliminary testing, and specimen blank fabrication are now complete for ASTM E2760-10 with round robin testing expected in 2014. Results will be published in 2015.
The round robin test program for ASTM E2714-09 utilized Grade 91 steel test blanks provided by EPRI. Sixteen laboratories around the world agreed to participate in the study, with 13 eventually reporting their test results. These results were considered by the ASTM Task Group on Creep–fatigue Crack Formation (E08.05.08). Strain controlled creep–fatigue tests were conducted at 625°C and at three stain amplitudes. Each laboratory followed the provisional standard, but variations to accommodate local procedures in specimen geometry, heating methods, and numbers of tests were acceptable. Statistical analysis of the inter- and intra-laboratory variability was conducted. Initial analysis of the results found the 95% confidence interval bands increased at longer hold times and lower strains, Fig. 5(a).
Comparison of data produced as part of the creep–fatigue round robin test program, (a) showing all the data, and (b) showing only data for tests which met the criteria for validity. The 95% confidence interval bands decreased significantly when post-test metallographic evaluation was used to eliminate invalid tests. (Colors are visible in the online version of the article;
One significant finding from the round robin tests was that post-test laboratory examination of specimens was necessary to determine whether or not a test was valid. Uneven heating, due to the use of induction heating methods, or failures due to bending could not be unambiguously identified simply by the laboratory results. However, these influences were clearly established by post-test inspection and metallographic examination. Indeed, results of the post-test examination resulted in some tests being shown to be invalid. Thus, the results were excluded from the dataset, Fig. 5(b). Performing analysis only on relevant test results was very important to meaningful assessment. The round robin program resulted in the following recommendations regarding the standard E2714:
The original precision and bias statements should be modified (this work is underway).
A recommendation to mandate post-test metallographic analysis should be added. This form of examination is essential to ensuring that the dominant crack(s) are formed within the gage length of the specimen and whether geometric instabilities, such as bending and/or bulging, was present thus warranting rejection of the data.
A cautionary note should be added to emphasize the possible deleterious effects of heating methods on measured lives.
In addition to the ASTM standards, the EPRI group produced a code of practice for short-crack growth under creep–fatigue conditions [13].
It is clear that updated methodologies for estimating damage due to cyclic operation in high energy components should distinguish between sequential damage, e.g. steady operation leading to creep followed by cyclic performance resulting in fatigue, and interactive damage, i.e. under conditions where the damage processes lead to synergistic, rapid damage development. This is particularly important for alloys where time and/or cyclic microstructural changes occur. Damage accumulation in creep–fatigue should be described by a formulation that includes terms giving the influence of creep on fatigue and vice versa. In other words, where damage is truly interactive, the capacity for creep must be reduced due to the fatigue, and the capacity for fatigue must be reduced by creep.
Other challenges associated with performing assessments include the fact that, despite repeated discussion of the need to utilize analysis methods which consider damage mechanisms, even recent work (e.g. [14]) continues to focus on attempting to judge the usefulness of curve fitting approaches on how well (or otherwise) limited laboratory results are described. In these cases, the emphasis is on accumulating selected data and then comparing the accuracy of fits with published examples for data descriptions [14]. When seeking to extrapolate results. The relevance of the expressions used for parametric fitting of experimental data must be considered. This issue is very important because although reasonable fits can be obtained to an existing data set, the accuracy of parametric data extrapolations may not be guaranteed even by a reasonable fit to a limited set of results. In all cases, there is an absolute requirement for the results of experimental and analytical programs to be assessed by application of meaningful post-test metallographic examination. Thus, post-test examination should be performed to document the type, density and character of damage present. At the very least, this should record whether the primary damage is intergranular or transgranular.
An example of accumulated information for CSEF steels in general, and Grade 91 in particular, has demonstrated that there is a very clear “Metallurgical Risk Factor” which must be considered with respect to assessment of the performance of components which operate in the creep range [9]. In-service experience of CSEF steel components [15] has shown that damage susceptibility can be linked to composition and heat treatment. This is true both for base steel and when creep or creep–fatigue damage is found in weld heat affected zones (HAZs). Information regarding the creep life of Grade 91 steel cross weld samples has been compiled into an integrated data set. The results for cross weld creep tests, at applied stresses below about 160 MPa for temperatures in the 600°C to 650°C, are shown in Fig. 6. The results are presented as the variation of Larson–Miller parameter (LMP) with applied stress. These results show the expected trend that lives increase with decreases in temperature and stress. However, recent results have shown that post exposure tests on ex-service girth welds from a header which experienced in-service cracking [15], recorded creep lives below the minimum of previous results, Fig. 6. EPRI is carrying out additional research to further assess the reasons for this relatively poor creep performance. It is apparent that results of cross weld tests and root cause analysis of in-service cracking of components should continue to be undertaken. These activities are critical to identification of the influence of specific fabrication variables on creep behavior. Work to establish and, if possible quantify, metallurgical risk factors associated with damage susceptibility has already indicated that the tendency for low ductility fracture in CSEF steels can be linked to steel making methods and the level of trace elements present (e.g., [9]).
Variation in the creep lives of cross weld tests for Grade 91 CSEF steel shown as a function of the Larson–Miller parameter. The lower inset (a) shows stub tube welds where local grinding was carried out to remediate in-service developed cracking. The upper inset (b) illustrates the results of component stress analysis undertaken as part of the cause analysis. (Colors are visible in the online version of the article;
The practicality of using complex constitutive equations with many variables also needs to be considered. While design approaches, particularly for turbine components, may be based on detailed knowledge, it is apparent that in-service assessment methods that are slightly conservative and can be applied without the need for extensive materials testing and/or complex stress analysis are required. This is particularly true for boilers and piping. These issues are illustrated by consideration of Fig. 7. Here different analytical approaches have been applied to creep–fatigue results from Grade 92 steel [16]. The relatively simple strain fraction method results in a consistently conservative prediction of behaviour, Fig. 7(a). In contrast, a more detailed modified strain fraction approach shows a much less conservative outcome, Fig. 7(b).
Comparison of experimental and predicted creep–fatigue lives for Grade 92 steel tested in the range 600 to 650°C. Analysis was performed using a strain fraction rule in (a) and a more complex modified strain fraction rule in (b) [16].
With the modified strain fraction analysis, experimental results and predictions are scattered on both sides of the line showing matching agreement, and all data are well within ±2. However, the knowledge necessary to apply this method to components is relatively complex. The continued validation of simple, sensibly conservative methods for at-risk plant components is necessary because of the implementation of new alloys and operating regimes.
There are presently few examples where comprehensive data sets have been established to allow a comprehensive assessment of creep–fatigue performance without the need for assumptions and/or extrapolations. Typically there are often challenges with application of the strain fraction/ductility exhaustion methodology because detailed measurements of strain: time behavior are not available. Agreement regarding the correct parameter to describe creep ductility is another parameter which can introduce uncertainty. This variability arises as there no agreed definition for the appropriate value of strain to fracture to use in the analysis. Suggested methods for the most appropriate ductility value include:
Data averaged from the range of available information for the particular alloy,
Simplified methods to estimate total strain, for example, using the product of the minimum creep rate and time to failure,
Estimates based on relationships describing creep strain with time for applicable stress/temperature combinations,
Strain to failure, elongation/reduction of area, or other measured value for different conditions which must then be extrapolated to the conditions of interest.
To expand the database of knowledge on the high temperature performance of Grade 92 steel a major research project is in progress. This research involves nearly 20 partners drawn from all sectors of the Electricity Supply Industry. This work involves round bar and notch bar creep tests on base metal from different sources, creep–fatigue tests on notched bar specimens and creep tests on cross-weld specimens. The mechanical test results are supported by a comprehensive program of microstructural characterization.
Very low creep ductilities have been reported in Grade 92 steel base metal samples using test conditions near the typical operating conditions of advanced boilers. The rupture life reported for different test temperatures is shown, with selected ranges in reduction of area (R of A) designated by different symbols in Fig. 8. The general key to Fig. 8 is that samples with an R of A greater than 50% are shown as an open circle, with tests with an R of A below 50% are shown as a solid square. The difference between open and solid symbols facilitates comparison of the results.
Reduction of area measured in creep tests on Grade 92 steel base metal samples as a function of test temperature and duration (a) and a micrograph showing the high density of creep voids present in tests where the R of A was less than 50% (b). (Colors are visible in the online version of the article;
Greater definition of the specific ductility ranges are provided through the use of different colors. The tests included in Fig. 8 cover data from many different sources, yet a general trend in the rupture behaviour is apparent [10]. For tests at 650°C and durations near to, or above 10,000 hours the creep ductility is below 50%. In contrast, for tests at 550°C even with durations approaching 100,000 hours the reported ductilities are >50%. At 600°C, i.e. near to the design temperature for many Grade 92 steel components, tests with durations above 10,000 hours show mixed behaviour. Thus, some steel casts show relatively low ductility at lives around 10,000 hours; yet others show creep ductilities above 50% even at rupture lives very close to 100,000 hours.
Post-test metallographic characterization has been performed to document the damage present and to identify the microstructural features linked to void nucleation [10]. It is now clear that a key factor associated with the development of creep voids in Grade 92 steel is the existence of high number densities of relatively large non-metallic inclusions.
Creep testing of notched bar specimens has demonstrated that trends in ductility reduction occur at shorter times than has been observed in plain bar cylindrical samples. These trends are typically explained by the enhanced effects of triaxial stresses in the notched samples. As noted previously the low ductility failures in the notched bars are a direct consequence of the formation and interlink up of creep voids. Detailed assessment has been performed to characterize the distribution of the creep voids in the notched region. A general photograph showing the distribution of damage at a notch within the last 5% of life is presented in Fig. 9. In this binarized image the contrast has been selected so that the creep voids are shown as bright spots. Even at a relatively low magnification it is apparent that creep voids extend across the net section between the notches. Detailed quantification of the number density of creep voids is in progress and will be complimented by the results of damage mechanics computer analysis.
Post-test binarized micrograph showing the development of creep voids and damage in the notch region of a Grade 92 steel sample.
The advantages of determining cyclic stress relaxation in notched samples include:
Testing at elevated temperatures compared to service so that results which are relevant to plant can be achieved in a shorter times,
Triaxial states of stress are present and these distributions can be varied by using different notch geometries, and
Stress relaxation occurs with elastic follow up; this is achieved by allowing the stress to relax during a displacement controlled creep dwell.
Previous work on AISI 316H steel [17] demonstrated that a multiaxial creep damage model based on ductility exhaustion provided a good prediction of failure at levels of triaxiality relevant to reheat cracking in plant welds. In the tests on Grade 92 steel, displacement was carefully controlled over the centre, parallel region of the gauge length. Four different levels of cyclic strain were used, namely 0.3%, 0.24%, 0.15% and 0.12%. It is apparent that the strain accumulation in the notch region will be significantly above these values and detailed analysis is in progress to link the different hold levels to local strain accumulation.
The relatively simple analysis that has been performed to date shows that for three different casts of Grade 92 steel (all complying with applicable specifications) the behaviour with hold time is different. As shown in Fig. 10, the behaviour with increasing hold time varies for the three steels which were fabricated using different procedures. For the steel designated as BM C (
Relationship showing the trend in number of cycles to failure with hold time for cyclic stress relaxation tests on three different Grade 92 steels [16]. (Colors are visible in the online version of the article;
The examples discussed demonstrate that a comprehensive assessment of creep, fatigue and creep–fatigue behaviour can only be obtained through the performance of relevant testing supported by detailed microstructural characterization and data analysis. EPRI continues working with members of the expert group to establish comprehensive data compilations on the most widely used alloys. It is planned that analysis of the data sets will result in published material data sheets suitable for base line type analysis. Alloys for which ‘Creep–fatigue data sheets’ are planned include key steels used in Boiler Headers and Piping, for example P22, P91 (X10CrMoV9-1), P92, E911, as well as the non-ferrous alloys IN718, IN706, IN617, etc.
It is apparent that more components in advanced power generating plants are operating under conditions of elevated temperature and increasingly under creep–fatigue conditions. This paper has outlines achievements from on-going EPRI collaborative work and illustrates recent results from high temperature creep–fatigue tests on tempered martensitic Grade 92 steel. It is clear that this work is critical to establishing sound technical solutions which support design for flexible operation and in-service assessment.
In many cases, consideration of the results of the Root Cause Assessment of cracking in CSEF steel components indicates that the use of these metallurgically complex steels provides challenges for the use of traditional Design by Rule methods. These problems are particularly focussed on issues linked to:
Cyclic operation,
Connections (and the associated reinforcement), and
Welds.
Thus, in-service cracking has often been found at welded connections where cyclic operation was involved.
An alternative viewpoint to the use of simple design methods is to consider the performance of a structure in terms of damage tolerance. This approach has been applied in many nuclear plant related circumstances where Design by Analysis is an established approach. However, damage tolerance has not been widely used in the design of components for fossil plants. The performance of feature test samples considering the extent of excavation on creep behavior has shown that repair welds in Grade 91 steel can be made more damage tolerant and less susceptible to catastrophic failure than joints made using traditional methods, This greater performance (by up to a factor of 2 on creep life) is achieved through application of improved joint geometry and controlled welding procedures.
A well-engineered weld repair has the ability to not only exhibit an acceptable creep life but, perhaps more importantly, to provide an improved life management solution as compared to the traditional weld design and fabrication methods. Because the HAZs for the original and the repair welds are offset, and since defect growth is driven by net section stress considerations, propagation must occur through the base metal. As the Grade 91 base metal has high inherent toughness, the rate of crack growth is slow compared to the propagation in the HAZ. Thus, the risk of fast through-wall, catastrophic failure is significantly reduced.
It is apparent then that future plant will benefit from use of materials which exhibit excellent strength and ductility and from design methods which consider the risks of fracture mechanisms for different operating scenarios. Even so, for most components it is recognized that in service inspection and assessment of damage will need to be undertaken. Ensuring that the design methods and the manufacturing approaches used, provide components with a low risk of fast fracture (and thus a greater expectation of leak before break) is also of major benefit to non destructive testing. It is established that as the size of the ‘indication’ that must be detected decreases, the likelihood of defects being missed and of false calls increases. The need to properly identify and characterize fine damage below 1 mm in linear dimension greatly increases the difficulty and costs associated with inspection.
Footnotes
Acknowledgements
Thanks to the support and contributions from the invited experts, the EPRI facilitated discussions on Creep–Fatigue have been successful in identifying key issues and agree research plans. This work has included involvement of the following Bob Ainsworth, Manchester University; Dave Dean and Mike Spindler, EdF UK; Ashok Saxina, University of Arkansas; Stuart Holdsworth, EMPA; Yukio Takahashi, CRIEPI; Bilal Dogan (formerly of GKSS and EPRI); Peter Skelton, consultant UK (formerly CEGB and Imperial College); Warwick Peyton, ANSTO; Toshimitsu A. Yokobori, Tohoku University: Karl Maile and Andreas Klenk, MPA Stuttgart; Olivier Ancelet, CEA; Hellmuth Klingelhöffer, BAM; Fujimitsu Masuyama, Kyushu Institute of Technology; Ian Perrin, Structural Integrity Associates Inc: Andre Pineau, Ecole des Mines de Paris; Alfred Scholz, T.U. Darmstadt.