You are on page 1of 6

Arturs Kalnins

Lehigh University, Bethlehem, PA 18015-3085 e-mail: ak01@Lehigh.edu

Fatigue Analysis in Pressure Vessel Design by Local Strain Approach: Methods and Software Requirements
The purpose, methods for the analysis, software requirements, and meaning of the results of the local strain approach are discussed for fatigue evaluation of a pressure vessel or its component designed for cyclic service. Three methods that are consistent with the approach are evaluated: the cycle-by-cycle method and two half-cycle methods, twiceyield and Seegers. For the cycle-by-cycle method, the linear kinematic hardening model is identied as the cyclic plasticity model that produces results consistent with the local strain approach. A total equivalent strain range, which is entered on a material strain-life curve to read cycles, is dened for multiaxial stress situations DOI: 10.1115/1.2137770

Introduction

Cofn 1 and Dowling et al. 2 explain the basic idea of the local strain approach. Detailed coverage is given in Chapter 14, Strain-Based Approach to Fatigue of Dowlings book 3 . It has been used in practice for fatigue assessment of pressure vessel components. Its appeal is the applicability to any smooth local geometry that can be dened. The ASME Boiler and Pressure Vessel B&PV Code 4 uses the local strain approach for fatigue evaluation on plastic basis. The paper considers applications to cases in which cyclic action experiences alternating plasticity. The main objectives are to identify the methods and software that are capable of calculating stress and strain ranges within the framework of the local strain approach and to identify a multiaxial total equivalent strain range that is consistent with the approach. This strain range is the counterpart to the uniaxial total strain range that is listed as ordinate of a material strain-life curve.

Basic Assumptions and Consequences

The local strain approach follows from the assumption that a sufciently small crack of the same size is developed at about the same number of cycles on the surfaces of a smooth test specimen and a smooth location of a pressure vessel component, when both are cycled at the same surface strain range and made of the same material. It is assumed that this equality holds true from the very rst crack appearance up to some crack size, which depends on the local geometry and the magnitude of the cycled strain range and is generally unknown. Although the actual crack size up to which the specimencomponent equality can be relied on is of no importance in its design procedure, the local strain approach is justied for design purposes only on the condition that the number of cycles to develop a crack of a given size in the component is not less than that in the smooth fatigue test specimens, at least on a statistical basis. If that is true, then the allowable cycles for a component taken from a material design fatigue curve that is constructed from the smooth specimen data can be expected to have a positive margin with respect to failure however dened, but the same for the specimens and component . It is possible that for some pressure vessel applications the
Contributed by the Pressure Vessels and Piping Division of ASME for publication in the JOURNAL OF PRESSURE VESSEL TECHNOLOGY. Manuscript received August 8, 2005; nal manuscript received October 10, 2005. Review conducted by G. E. Otto Widera.

above condition leads to a margin that may be judged overly generous. If that is unacceptable, a different design procedure has to be formulated and followed. If that is not an option, the generous margin has to be accepted as part of the price for the simplicity and wide applicability any modelable geometry, loading, and material of the local strain approach. As an illustration of a case for which the above condition is met, consider a component in which a crack enters a plastic zone with a decreasing strain range eld that is surrounded by elastically cycled material, which is a common situation in pressure vessels. It is expected that, after a certain crack size is reached, the crack growth in this component will be slower than that in a small-diameter round bar fatigue specimen, in which the crack enters an almost uniform strain range eld. The paper by Kalnins and Dowling 5 supports this scenario. It uses test data cited in Figs. 10.8 and 14.9 of Dowlings book 3 on blunt double-notch plate components and smooth, 6.35 mm dia 0.25 in. dia , round fatigue test specimens, both made from the same heat of AISI 4340 steel. Both are cycled to two predened conditions: the appearance of a 0.5 mm 0.02 in. crack and failure. Figures 1 and 2 show the number of cycles obtained from the tests for each of the two conditions. As seen from Fig. 1, the number of cycles to develop a 0.5 mm crack in the fatigue specimens and plate components is about the same for the lower strain ranges, but for strain ranges above 0.014, the plate takes more cycles to reach a 0.5 mm crack. In Fig. 2, the difference in cycles to failure is far greater. For example, for a strain range of 0.0136, the specimen fails at 2400 cycles, while it takes 6027 cycles for the plate component to reach failure in the test. These results show that the above condition is met. If the strainlife curve is constructed from the specimen data, the positive margin for the plate components is apparent. Of course, for design purposes, factors will be applied to the specimen curve, which will increase the actual design margin.

Irregular Loading

The local strain approach applies to constant-amplitude cycling. If the loading histogram consists of an irregular loading pattern, it has to be resolved into loading ranges that produce individual stress-strain cycles before the local-strain analysis can be begun. This means that if a repeated loading block is dened over a time interval for which all loading components are specied at a number of time points, the local strain approach requires that all stressstrain cycles that are produced by the loading block be identied. This can be achieved by appropriate cycle-counting methods see Transactions of the ASME

2 / Vol. 128, FEBRUARY 2006

Copyright 2006 by ASME

Downloaded 10 Oct 2007 to 203.78.221.48. Redistribution subject to ASME license or copyright, see http://www.asme.org/terms/Terms_Use.cfm

Fig. 1 Cycles to reach 0.5 mm crack in plate and specimen by test

3 . The accumulated usage factor is then calculated over all the individual stress-strain cycles of the loading block, following the Palmgren-Miner rule 3 , Chap. 9 . The same applies to cases in which more than one loading block may be applied in a random sequence, each repeated a specied number of times. In the remainder of the paper, the analyses will be assumed applied to one stress-strain cycle.

Stabilized Cycle

Fig. 3 Monotonic lines only and cyclic markers curves for SA-516 Grade 70 steel. L denotes longitudinal and T denotes transverse orientation of specimens machined from a plate. Reprinted from Fig. 2 of 6, Copyright 1984, with permission from Elsevier.

As part of the design procedure, the local strain approach assumes that a single value of a strain range is used for assessing fatigue damage for the life of a pressure vessel component. Since hardening and softening with cycles accompany the initial phase of cycling, during which the strain and stress ranges may change, the question is: What strain range shall it be? Since for many metals the stress and strain ranges tend to stabilize, so that stabilized hysteresis loops are experienced for the major part of life, the obvious answer is to bypass the hardening and softening with cycles and to accept the strain range of the stabilized cycle as representative of whole life. In some cases, stabilization may be difcult to achieve even until failure. Such cases notwithstanding, the local strain approach assumes that the cyclic action from which the cyclic stress-strain curve of the material is derived has stabilized. That has to be accepted as part of the design procedure.

Cyclic Material Curve

Substitution of a monotonic curve for the cyclic curve may cause problems. For example, Lefebre and Ellyin 6 present curve ttings to test data on specimens made of SA-516 Grade 70 steel, shown in Fig. 3. Stress amplitude is plotted versus strain amplitude for cyclic loading and compared to plots for monotonic loading. The material shows softening with cycles up to strain amplitude of about 0.4% and hardening above that level. Problems may arise within strain amplitudes of 0.10.4%, where the use of the monotonic curve can predict strain ranges in a component that err on the unconservative side. Of course, an accurate curve tting to the cyclic data of the material under consideration is preferable, if one is available. However, for design purposes, approximations could be agreed on for certain classes of materials, which would parallel those of the design fatigue curves now used in the ASME B&PV Code 4 .

Having established that stabilized action is the target, it follows that the cyclic-stress-rangestrain-range or amplitude curve of the material provides the information that is needed for the material model used in the analysis. Such cyclic test data are available for many materials. Typical curve ttings to these data can be obtained in terms of three parameters: cyclic elastic modulus E, a stress parameter, and an exponent.

Applicability

Fig. 2 Cycles to reach failure in plate and specimen by test

The local strain approach is applicable to cases in which all structural features that affect fatigue damage are dened and can be modeled with sufcient accuracy. It is not applicable to cases in which some structural detail is known to affect fatigue damage but cannot be modeled, either because its geometry is unknown e.g., aws at the weld toe of an untreated weld or because its model is unreliable e.g., very sharp notch . Such cases require approaches that incorporate the unmodelable details in the test data, such as, for example, those described by Maddox 7 for weld joint classes, and more recently by Dong et al. 8 . Limitations on loading are not so clear. Proportional loading presents no problems, but cases when the principal stress and strain axes rotate have been shown to pose a problem. Itoh et al. 9 present test data for shear and axial strains that are imposed nonproportionally to the test section of a thin cylindrical shell, forcing the principal axes to rotate. The data show unsatisfactory correlation with predictions using the principal and equivalent strain ranges that are in the current ASME B&PV Code 4 . Kalnins 10 has shown that the hysteresis loops for some of the nonproportional cases e.g., case 10 in 9 exhibit no elastic unloading and the reversal points of different components do not coincide, which prevents the use of the methods of Sec. 7. It may FEBRUARY 2006, Vol. 128 / 3

Journal of Pressure Vessel Technology

Downloaded 10 Oct 2007 to 203.78.221.48. Redistribution subject to ASME license or copyright, see http://www.asme.org/terms/Terms_Use.cfm

Fig. 5 Stabilized hysteresis loop Fig. 4 Cyclic curve and calculated ranges

be that such situations are rare in pressure vessel components. Whether a caveat for these cases is or is not needed in design standards is an open question. It is assumed for the remainder of the paper that stress and strain reversal points of all nontrivial components coincide so that a multiaxial equivalent stress-strain cycle can be dened Sec. 9.2 . This is ensured for proportional loading. The conditions for which it may also be true for nonproportional loading require further investigation.

Methods and Software

As per Sec. 3, the analysis is applied to each individual stressstrain cycle that is produced by the loading histogram. These cycles must be identied, and the two time points at their stress and strain reversals determined. The loading components at the reversal points can then be evaluated and the loading range for the cycle determined. At this point, it has been established that the cyclic curve of the material will be used to model the material and the loading will consist of the loading range. Now the question is: What methods and software will solve the problem in a way that is consistent with the local strain approach? Two basic methods are discussed: cycle-by-cycle and half-cycle methods. For the latter, the twiceyield and Seegers methods are included. The cycle-by-cycle method is discussed only because of its appeal for modeling cyclic action, but, when used for design purposes, it requires far more effort and is less generic to software than the half-cycle methods. The stress and strain ranges obtained by all three methods are the same. 7.1 Cycle-by-Cycle Method. Elastic-plastic nite element analysis FEA is performed over a sufcient number of repetitions of a selected cycle until the stress and strain values at the reversal points stabilize. The cyclic-stress-amplitudestrainamplitude curve of the material is used as input for the monotonic uniaxial material model. The loading can be specied as either between the loading at the reversal points of the cycle or between plus and minus of the loading amplitude. If in the former case the hysteresis loop indicates a mean stress, its effect is neglected as per Sec. 8. Since up- and downloading is performed, a cyclic plasticity model must be specied to model the unloading and reloading phases. Two cyclic plasticity models will be considered for the cycle-by-cycle method. One includes linear and the other nonlinear hardening. For details, see the ABAQUS 11 Standard Users Manual II, 11.2.2-2, or Refs. 12,13 . The question is which cyclic plasticity model is consistent with the local strain approach. To answer that question, a validity check is given next. 7.1.1 Cyclic Plasticity Validity Check. According to Sec. 4, the cyclic stress-rangestrain-range or amplitude curve is used, which means that the hysteresis loops in the specimens stabilize. Since the material of the component is supposed to be the same as that of the test specimens, the loops in the components should also 4 / Vol. 128, FEBRUARY 2006

stabilize in the same way. The software that is used for the cycleby-cycle method must reect this behavior. This means that it must be able to replicate the cyclic curve for a uniaxial stress state in a component. In other words, the calculated stress and plastic strain ranges must lie on the cyclic-stress-rangeplastic-strainrange curve that has been input. This requirement will be used as a validity check in evaluating the cyclic plasticity models of the software. To illustrate the validity check, consider a uniaxial stress state cycled in strain control using a cyclic plasticity model that is to be checked for consistency with the local strain approach. The cyclic curve shown in Fig. 4 rewritten in amplitudes is input for the monotonic material model. The calculated stabilized hysteresis loop is shown Fig. 5. It shows a plastic strain range of 0.033 and a stress range of 1330 MPa. Is the cyclic plasticity model consistent with the local strain approach? This is decided by plotting the coordinates for the calculated ranges, 0.033 and 1330, in Fig. 4. It is seen that this point lies on the cyclic curve. If the cycling were done at different strain ranges, the cyclic curve in Fig. 4 would be duplicated. Therefore, the cyclic plasticity model used in this analysis is consistent with the local strain approach. 7.1.2 Linear Hardening in Cyclic Plasticity. Cyclic plasticity models with linear hardening involve two separate components: isotropic and kinematic. The model with isotropic hardening expands the yield surface until purely elastic action remains, as shown in Fig. 6. This result does not meet the validity check of Sec. 7.1.1 and is not acceptable. The kinematic component is considered next. Cyclic plasticity models with linear kinematic hardening have been developed that assume Masing behavior Sec. 2.5 of 14 , according to which magnifying the cyclic stress-strain amplitude curve by a factor of 2 approximates the two branches of a stabilized hysteresis loop. These models pass the validity check of Sec. 7.1.1. Among the popular nite element programs, ANSYS 15 linear kinematic hardening model KINH supports a multilinear

Fig. 6

Stabilized cycle using linear isotropic hardening model

Transactions of the ASME

Downloaded 10 Oct 2007 to 203.78.221.48. Redistribution subject to ASME license or copyright, see http://www.asme.org/terms/Terms_Use.cfm

Fig. 7 Stabilized cycle using linear kinematic hardening model

curved curve tting to the cyclic data for the input of the monotonic material model, while that of ABAQUS 11 Version 6.3-1 supports only a bilinear curvet to the cyclic data when the parameter hardening= kinematic is invoked. Note that ABAQUS permits the input of a curved curve tting to the cyclic data for the input of the monotonic material model when no cyclic plasticity with kinematic hardening is specied, which is the case for the half-cycle methods. For an illustration, consider a single eight-noded brick element, cycled in uniaxial strain control between a strain of 0.03 and 0.01. Figure 7 shows1 the stress-strain response for which the stress range of 1330 MPa is obtained. When cycled with the same strain range, but fully reversed between 0.02 and 0.02, exactly the same stress range is predicted. In both calculations, the cyclic curve in Fig. 4 rewritten in amplitudes is used. The square marker in Fig. 4 shows the point with the coordinates of the calculated stress range and plastic strain range. The fact that it lies on the cyclic curve indicates that the test of Sec. 7.1.1 has been met. 7.1.3 Nonlinear Cyclic Plasticity Models. Nonlinear cyclic plasticity models e.g., 1113,15 , which contain combined isotropic/kinematic components, are not designed to receive a generic cyclic stress-strain curve of the material as input and calculate the stress and strain ranges that represent a stabilized cycle of the same material. The problem is that the input is written for a specied strain range, which is the end product of the analysis and, therefore, unknown before the analysis. For this reason, the nonlinear hardening models used in Refs. 12,13 do not meet the validity check of Sec. 7.1.1. The following example illustrates the problem. Again a single eight-noded brick element is subjected to fully reversed, straincontrolled cycling in one direction, producing a uniaxial stress state. ABAQUS 11 data-type= stabilized parameter is selected, for which an approximation derived from the cyclic curve shown in Fig. 8 is used as input. No isotropic component is used. The model includes only the nonlinear kinematic NLK component. The response is shown by the curve marked NLK in Fig. 9. It shows clearly that the stabilized cycle of the cyclic curve that was input has not been replicated. The stress range given by the NLK model is 1240 MPa, while the corresponding value on the cyclic curve is 962 MPa, which is also plotted in Fig. 8. It is clear that the model does not replicate the cyclic curve that has been input and fails the validity check of Sec. 7.1.1. 7.2 Half-Cycle Methods. These methods take advantage of the stabilized form of the hysteresis loop of the cycle. There is no need to perform the cycle-by-cycle method over a number of cycles if the two branches of the loop are geometrically similar, as shown in Fig. 7. FEA over just one branch of the loop gives the desired stress and strain ranges. That is the basis of the half-cycle methods. They require only one monotonic FEA of one load step,
1 Jrgen Rudolph of the University of Dortmund, Germany, performed the calculations for this gure using ANSYS KINH linear kinematic hardening model.

Fig. 8 Cyclic curve and calculated ranges using NLK model

with no unloading and reloading, and do not require cyclic plasticity models. The advantage is simplicity no FEA over cycles and that they can be performed with any nite element program that has an incremental plasticity option for static loading. The half-cycle methods give strain and stress ranges that, for practical purposes, are the same as those obtained by the cycle-by-cycle method of Sec. 7.1. The two half-cycle methods are considered next. 7.2.1 Twice-Yield Method. Theoretical support of this method can be found in the work of Mroz 16 . Dowling 17 and Dowling and Wilson 18 applied it to some special cases. More recently, Kalnins 19 proposed it as a general method for design and called it the twice-yield method. It is applicable to cyclic primary and nonprimary e.g., transient thermal loading; that is, its applicability is the same as that of the cycle-by-cycle method. The only limitations are stated in Sec. 6. From an FEA perspective, the twice-yield method is explained by the observation that if in the input the load is specied as the loading range and the cyclic stress-rangestrain-range curve is used for the material model, then in the output the stress components are the stress component ranges and the strain components are the strain component ranges. Thus, in one FEA load step, for which the loading is specied from zero to that of the loading range, the output provides the stress and strain ranges that are needed in the local strain approach. When coupled with the multiaxial total strain range Sec. 9.2 , the twice-yield method is far simpler than the cycle-by-cycle method. After the reversal points of the loading for the cycle and the loading range have been determined, the method is straightforward. The quantities that are taken from the output are the multiaxial equivalent stress range, eq, given by Eq. 4 , and the

Fig. 9 Calculated cycle using NLK model

Journal of Pressure Vessel Technology

FEBRUARY 2006, Vol. 128 / 5

Downloaded 10 Oct 2007 to 203.78.221.48. Redistribution subject to ASME license or copyright, see http://www.asme.org/terms/Terms_Use.cfm

equivalent plastic strain range, peq, given by Eq. 5 . Typical nite element programs calculate them automatically. For example, ABAQUS 11 calls eq MISES, and peq PEMAG. ANSYS 15 uses similar variable names in the output. A generic output le is scanned for the maximum value of peq, and eq is then recorded at the same location. No search of the solution database is required. The total strain range is then obtained from Eq. 6 using a hand calculator. 7.2.2 Seegers Method. Seeger gives the general background in 20 . Rudolph and Weiss 21 describe the procedure and discuss its application to weld seams with postweld treatment. It is applicable to proportional loading; that is, to cases in which all loading components are multiplied by a single function of time, say, L. Seegers method performs only one FEA of the component from L = 0 to the greatest magnitude of L on the histogram and and strain measure at a number of records a selected stress L values that is sufcient to permit a curve tting by an equation e.g, Ramberg-Osgood . The curve tting between L and is and is called the component yield curve and that between called the local - curve. After the two curvets are derived, the unload branch of the hysteresis loop of each cycle is constructed by assuming Masing behavior see 14 . The stress and strain ranges are determined from this branch. For details, see 20,21 . Regarding the comparison between the two half-cycle methods, twice-yield method performs an FEA for each stress-strain cycle separately, but determines the stress and strain ranges with no postprocessing. The advantage of Seegers method is that the results of a single FEA can be used for a number of stress-strain cycles with different loading amplitudes of the same set of loading.

needed here only to identify the consistent stress and strain parameters in the local strain approach. The user of any of the methods discussed in Sec. 7 does not have to construct one for an application. In multiaxial situations, hysteresis loops are commonly constructed for corresponding stress and strain components separately. This does not reveal which strain range is entered on the material strain-life curve. Kalnins et al. 22 developed the concept of a multiaxial equivalent hysteresis loop and showed how to construct one. This is discussed next. 9.1 Uniaxial Stress State. To obtain a template for a multiaxial stress case, a procedure is outlined rst for a uniaxial stress state in a fatigue test specimen. The calculation is performed with the cycle-by-cycle method of Sec. 7.1 using the cyclic plasticity model with linear kinematic hardening. 1. Plot stress versus axial plastic strain over one stabilized cycle and obtain a hysteresis loop, which may look like that in Fig. 5. 2. Note that its height is the uniaxial stress range , and its width is the axial plastic strain range p. 3. Calculate the axial total strain range from
t

where E is the modulus of the elastic portion of the cyclic curve. 4. Enter t as ordinate on the strain-life curve to read cycles. 5. Note that and p lie on the cyclic stress range-plastic strain range curve of the material, just like the square marker in Fig. 4. 9.2 Multiaxial Stress State. The ve steps in Sec. 9.1 are now retraced for the multiaxial stress case. 1. Use again cycle-by-cycle method to calculate all stress ij and plastic strain pij components at a number of output points over a stabilized cycle that would be sufcient to draw a graph. Then the following two quantities are calculated at each of the output points:
1 eq = 2

Mean Stress

What is known about each stabilized stress-strain cycle is only its loading range, which is used to calculate the stress and strain ranges of the stabilized cycle. The loading may begin with asymmetric components at the reversal points, but once the stress-strain cycle has stabilized, no information is available regarding the mean stress of the cycle. Since the only description of cyclic behavior of the material is taken from the cyclic stress-range strain-range curve Sec. 5 , which contains no information on mean stress, the magnitude of mean stress, if one is present, is unknown. The lack of knowledge of the mean stress is not a problem when the design fatigue curves of the ASME B&PV Code 4 are used, in which mean stress is assumed zero when alternating plasticity is present. This is supported by Ellyin 14 , who has shown that mean stress approaches negligible magnitudes when test specimens of SA-516 Grade 70 steel are cycled with various degrees of mean strain. A reasonable assumption is that the effect of any mean stress that may actually occur in an individual cycle with alternating plasticity in a real component can be neglected. This is made part of the design procedure considered in this paper.

2 2

2 3

2 1

+6

2 12

2 23

2 31

2
peq =
2 3

p11 p22 2 + p22 p33 2 + p33 p11 2 +


r ij r pij = pij

3 2

2 2 2 p12 + p23 + p31

3
where ij = pij, i , j = 1 , 2 , 3. The superscript ij, r , r = 1 , 2, refers to the stress and plastic strain components at the reversal points, t1 and t2. The minus signs apply to the right-hand r r downward leg see Fig. 5 of the hysteresis loop, with ij , pij xed at the upper extreme. The plus signs apply to the left-hand r r upward leg of the loop, with ij , pij xed at the lower extreme. The resulting curve of eq versus peq is the multiaxial equivalent hysteresis loop of the cycle, which is the counterpart to the uniaxial hysteresis loop of the fatigue test specimen. It may look like that in Fig. 5. 2. Just as in the uniaxial case, this loop identies the stress range and strain range that describe the size of the loop. Its height is the multiaxial equivalent stress range and its width is the multiaxial equivalent plastic strain range, which are now dened by

Multiaxial Stress and Strain Equivalents

The objective of this section is to obtain the multiaxial stress and strain equivalents that are appropriate for the local strain approach. This will be achieved based on one multiaxial equivalent hysteresis loop that represents the cycle as a whole. This loop is

1 eq = 2

+6

2 12 +

2 23 +

2 31

6 / Vol. 128, FEBRUARY 2006

Transactions of the ASME

Downloaded 10 Oct 2007 to 203.78.221.48. Redistribution subject to ASME license or copyright, see http://www.asme.org/terms/Terms_Use.cfm

2 peq = 3

p11 p22 2 +

p22 p33 2 +

p33 p11 2 +

3 2

p2 + p2 + p2 12 23 31

where pij = pij pij , i , j = 1 , 2 , 3. The superij = ij ij , scripts denote the subscripts at reversal points t1 and t2. 3. In analogy to Eq. 1 , the multiaxial total equivalent strain range2 is dened by
eq = eq

peq

4. The strain range of Eq. 6 is the multiaxial counterpart to the uniaxial total strain range of Eq. 1 and is entered on the strainlife curve to read cycles. 5. According to the theory of plasticity used in typical FEA software, the calculated eq and eq lie on the cyclic stress range-plastic strain range curve that is input for the required monotonic material model, just like the square marker in Fig. 4, thus meeting the validity check of Sec. 7.1.1. 9.3 Discussion. The multiaxial total equivalent strain range of Eq. 6 is superior to the maximum principal total strain range, which is used in the ASME B&PV Code 4 , both in Section 8-Div. 2, 4-136.2 c , and in Section 3, NB-3228.4 c . A simple example of equibiaxial, in-plane cycling of a plate3 refutes its general application in the local strain approach. The maximum principal strain is perpendicular to the plate while the stress component in that direction is zero. This produces a degenerate hysteresis loop of a straight line on the strain axis, which is not consistent with the hysteresis loop observed in the cycling of a fatigue test specimen. The multiaxial total equivalent strain range is also superior to the equivalent strain range that is dened in terms of total strain component ranges, which is used in 4 , Section III-NH, Appendix T, T-1413 for elevated temperature service. Its problem is that it does not reduce to the correct strain range for a uniaxial stress state. The multiaxial total equivalent strain range of Eq. 6 is dened in terms of Mises stress and plastic strain components. It provides a smooth transition to purely elastic action per cycle if the Mises multiaxial elastic stress range is taken as the stress measure for fatigue analysis in the elastic case. This was assumed here because it has been recommended for the new Division 2 of Section 8 of the ASME B&PV Code 4 . However, the current 2004 edition of the ASME B&PV Code 4 uses the Tresca stress components for purely elastic action. In that case, an effective combined strain range can be dened on the basis of the maximum shear strain as shown in a recent paper by Reinhardt 23 .

4. The twice-yield and Seegers methods require software with only incremental plasticity model for monotonic loading. No cyclic plasticity models are used. 5. The multiaxial total equivalent strain range dened in the paper is the multiaxial counterpart to the strain range listed on the design fatigue curve. 6. Twice-yield is the simplest method for calculating the multiaxial total equivalent strain range for a selected stressstrain cycle.

Acknowledgment
This research was supported in part by the Pressure Vessel Research Council through Grant No. 01-DIV2/PNV-23AS. The author also wishes to thank Dr. Wolf Reinhardt of Babcock & Wilcox Industries, Cambridge, Ontario, Canada, for many helpful discussions on the topic of this paper.

References
1 Cofn, L. F., Jr., 1973, Fatigue at High Temperature, Fatigue at Elevated Temperatures, American Soc. for Testing and Materials, ASTM STP 520, pp. 534. 2 Dowling, N. E., Brose, W. R., and Wilson, W. K., 1977, Notched Member Life Predictions by the Local Strain Approach, Fatigue Under Complex LoadingAnalysis and Experiments, Society of Automotive Engineers, Warrendale, PA. 3 Dowling, N. E., 1999, Mechanical Behavior of Materials, 2nd ed., Prentice Hall, Englewood Cliffs, NJ. 4 ASME, 2004, ASME Boiler and Pressure Vessel Code, ASME, New York. 5 Kalnins, A., and Dowling, N. E., 2004, Design Criterion of Fatigue Analysis on Plastic Basis by ASME Boiler and Pressure Vessel Code, ASME J. Pressure Vessel Technol., 126, pp. 461465. 6 Lefebre, D., and Ellyin, F., 1984, Cyclic Response and Inelastic Strain Energy in Low Cycle Fatigue, Int. J. Fatigue, 6, pp. 915. 7 Maddox, S. J., 1998, Fatigue Strength of Welded Structures, 2nd ed., Abington Publishing, Cambridge, England. 8 Dong, P., Hong, J. K., Osage, D. A., and Prager, M., 2003, Master S-N Curve Method for Fatigue Evaluation of Welded Components, Weld. Res. Counc. Bull., 474, pp. 150. 9 Itoh, T., Sakane, M., Ohnami, M., and Socie, D. F., 1995, Nonproportional Low Cycle Fatigue Criterion for Type 304 Stainless Steel, ASME J. Eng. Mater. Technol., 117, pp. 285292. 10 Kalnins, A., 2003, Fatigue Analysis of Pressure Vessels Based on Individual Stress-Strain Cycles, J. L. Zeman, ed., Proc. ICPVT-10, Vienna, July 710, Oesterr. Ges. f. Schweisstechnik, Vienna, pp. 349356. 11 ABAQUS Finite Element Program, Version 6.1, Hibbitt, Karlsson, and Sorensen, Inc., Pawtucket, RI, by Educational License to Lehigh University. 12 Armstrong, P. J., and Frederick, C. O., 1966, A Mathematical Representation of the Multiaxial Bauschinger Effect, Berkeley Nuclear Laboratories, R&D Department Report No. RD/B/N/731. 13 Chaboche, J.-L., Time-independent Constitutive Theories for Cyclic Plasticity, Int. J. Plast., 2 2 , pp. 149188. 14 Ellyin, F., 1997, Fatigue Damage, Crack Growth and Life Prediction, Chapman and Hall, London. 15 ANSYS Finite Element Program, Southpointe, Canonsburg, PA. 16 Mroz, Z., 1973, Boundary-Value Problems in Cyclic Plasticity, Second Int. Conf. on Structural Mechanics in Reactor Technology, Berlin, Vol. 6B, Part L, Paper No. L7/6. 17 Dowling, N. E., 1978, Stress-Strain Analysis of Cyclic Plastic Bending and Torsion, ASME J. Eng. Mater. Technol., 100, pp. 157163. 18 Dowling, N. E., and Wilson, W. K., 1979, Analysis of Notch Strain for Cyclic Loading, 5th Int. Conf. on Structural Mechanics in Reactor Technology, Vol. L, Berlin, Paper No. L13/4, pp. 18. 19 Kalnins, A., 2001, Fatigue Analysis of Pressure Vessels With Twice-Yield Plastic FEA, ASME Bound Vol. 419, pp. 4352. 20 Seeger, T., 1996, Grundlagen fur Betriebsfestigkeits-nachweise, Stahlbau handbuch, Band 1, Teil B, Stahlbau-Verlagsgesellschaft, Koln. 21 Rudolph, J., and Weiss, E., 2003, Service Durability of TIG Dressed and Ground Weld Seams Subjected to Proportional Loading Criteria, Proc. ICPVT-10, J. L. Zeman, ed., Vienna, July 710, Oesterr. Ges. f. Schweisstechnik, Vienna, pp. 465475. 22 Kalnins, A., Updike, D. P., and Park, I., 1999, Evaluation of Fatigue in the Knuckle of a Torispherical Head in the Presence of a Nozzle, Welding Research Council Progress Report, Sept./Oct., pp. 3193. 23 Reinhardt, W., 2005, Strain Measures for Fatigue Assessment Using ElasticPlastic FEA, ASME PVP Conference, July, Denver, PVP200571547.

10

Conclusions

1. The local strain approach gives allowable cycles with a design margin that depends on the local geometry and the magnitude of the cycled strain range. 2. The cycle-by-cycle method must be used with linear kinematic cyclic plasticity model, not isotropic. Nonlinear isotropic/kinematic cyclic plasticity models do not give results consistent with the local strain approach and should not be used. 3. The cycle-by-cycle method is more labor intensive and requires software with a cyclic plasticity model but gives the same strain ranges as the twice-yield or Seegers method.
2 This strain range was introduced by Dowling 3 who called it the effective strain range 3 Professor Masao Sakane of Ritsumeikan University, Shiga, Japan, pointed this out to the author.

Journal of Pressure Vessel Technology

FEBRUARY 2006, Vol. 128 / 7

Downloaded 10 Oct 2007 to 203.78.221.48. Redistribution subject to ASME license or copyright, see http://www.asme.org/terms/Terms_Use.cfm

You might also like