ML20054G480

Crbr Special Matls Considerations, Crbr Engineering Study Rept

ML20054G480
Person / Time
Site:	Clinch River
Issue date:	05/31/1982
From:	Diamond S, Mallett R WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:
Shared Package
ML20054G450	List: ML20054G449 ML20054G456 ML20054G465 ML20054G480 ML20054G488
References
ES-LPD-82-011, ES-LPD-82-11, NUDOCS 8206210595
Download: ML20054G480 (44)

Text

{{#Wiki_filter:ES-LPD-82-011 CRBRP SPECIAL MATERIALS CONSIDERATIONS CRBRP ENGINEERING STUDY REPORT i MAY1982 l WESTINGHOUSE ELECTRIC CORPORATION ADVANCED REACTORS DIVISION P.O. BOX 158 MADISON, PENNSYLVANIA 15663 Compiled by: Approved by: Snowl ( S. Diamond R. H. Mallett, Manager W-ARD CRBRP Piping Design and Mechanical Equipment l OSO'$$00Po!$88$37 l A PDR

ES-LPD-82-011 I CRBRP SPECIAL MATERIALS i CONSIDERATIONS CRBRP ENGINEERING STUDY REPORT I l MAY 1982 l APPLIED TECHNOLOGY Any further distriDution by any holder of this document or of the data therein to third parties representing foreign interests foreign governments, foreign companies and foreign Subsids-aries or foreign divisions of U S companies should De coordinated with the Director. Division of Reactor Aesearch and Technology. Department of Energy WESTINGHOUSE ELECTRIC CORPORATION ADVANCED REACTORS DIVISION P.O. BOX 158 MADISON, PENNSYLVANIA 15663 l Compiled by: Approved by: id l S. Diamond R. H. Mallett, Manager l W-ARD CRBRP Piping Design and ) Mechanical Equipment l l 1 l l

ABSTRACT This presentation provides responses to coecific NRC questions on CRBRP materials considerations. Metallurgical isotch effects are examined, elevated temperature acceptance tests are discussed and the evaluation of creep rupture damage from residual welding stresses is described. It is concluded that metallurgical notches are accommodated by the ductility of the materials used in the plant. It is shown that the current Code rules and acceptance tests are adequate for plant design. It is concluded that the relaxation of residual stresses cause negligible plastic strain. Finally, it is shown that creep rupture and fatigue strengths for notches are greater than minimum data used in determining ASME Code allowable stresses for design, ii ~

TABLE OF CONTENTS Page ABSTRACT 11 1.0 SilMMARY 1

2.0 INTRODUCTION

2 3.0 SPECIAL MATERIALS CONSIDERATIONS 5 M. MANJOINE (W-R&D) iii

1.0

SUMMARY

Discussions are presented on special CRBRP materials considerations with the purpose of examining notch effects, the need for elevated temperature dCCeptance tests, and the evaluation of creep rupture damage from residual forming and welding stresses in response to NRC questions CS 210.1 and CS 250.4. It is shown that creep rupture and fatigue strengths for notches under biaxial and triaxial strasses are greater than minimum data used in deter-mining ASME Code allowable stresses for design. It is concluded that the current Code rules and acceptance tests are adequate for plant design. It is shown that the residual stresses relax in service but the resulting plastic strains are less than 10% of the creep-rupture ductilities of the zones of the weldment. Finally, it is concluded that the strain concentration of a metallurgical notch is accommodated by the high ductilities of the zones of a weldment. 1

l

2.0 INTRODUCTION

A series of questions was sent to the CRBRP Project to address concerns about a intended CRBRP materials, high and low temper'iture regions of the plant, design and analyses approaches, and specific welded joints in the plant, l 1.e., the reactor vessel transition joint and the IHTS transition joints. There were two specific questions concerning special CRBRP materials considera-tions, CS 210.1, Part B and CS 250.4, Parts (1), (ii) and (iv), which are presented below: CS 210.1 In piping systems at elevated temperatures, local deformation i may occur at areas of geometric discontinuity, such as at fittings. Provide methods and procedures for the following: l A. Define elastic follow-up. l i B. Evaluate creep rupture and fatigue damage. C. Justify the use of simplified creep ratcheting bounding techniques used in computer codes. i CS 250.4 Provide the method for structural evaluation of weldments l and associated materials for service at elevated temperature. I The following should be addressed: (i) room temperature and elevated temperature material acceptance criteria: (ii) creep rupture damage resulting from residual forming I and welding stresses: (iii) mass transfer effects: (iv) metallurgical notch effects: (v) fracture toughness criteria: (vi) thermal aging effects; and (vii) irradiation effects. 2 l 1 I

- To address these and other questions a CRBRP/NRC meeting was held at Bethesda, Maryland on April 6-7, 1982 at which there was a topical presentation concerning special materials considerations for CRBRP applications. Figure I was used in introducing this presentation. This report provides the responses to the NRC concerning the CRBRP special materials considerations, i I i l i l l

FIGURE 2.0-1 HTS MATERIALS AND STRUCTURES VI. Special Materials Considerations

• Purpose (CS 250.4.iv)

- To explain notch effects e Conclusion - Notches are accommodated by ductility z. Purpose (CS 250.4.i) - To discuss the need for elevated temperature acceptance tests e Conclusion - Current code rules and acceptance tests are adequate e Purpose (CS 250.4..ii) - To evaluate creep rupture damage from residual stress e Conclusion - Relaxatio$ of residual stress causes negligible plastical strain

• Responds to O Cs 250.4

SPECIAL MATERIALS CONSIDERATIONS By M. Manjoine (W-R&D) 1 55928-444B:2 (53597) 1 5

SPECIAL MATERIALS CONSIDERATIONS Three of the special materials considerations that were identified in the NRC questions are addressed here. The first one is 250.4 and a subpart on the question of notch effects of various sorts in our structures and how they should be evaluated. The conclusion is reached that the ductility of the materials is sufficient to accommodate the effects of those notches. Another subpart of question 250.4 addresses the question of the acceptance tests for materials to be operated at elevated temperatures. This presentation will show that the current Code rule, specifically t he acceptance tests, are adequate for that purpose. And yet a third subpart of section 250.4 addresses the question of the effects of residual stresses as they relate to creep rupture damage. The conclusion which we will present is that the relaxation of those residual stresses which do exist early on causes negligible plastic strain in the equipment. The tests to be described really come out of the technology program and have been going on for about the last seven years (Figure 1). These include fairly large specimens into which geometric discontinuities can be introduced and welds made where the residual stresses are still retained since they are not removed by machining the weldment into a small specimen. What we will discuss is two parts of the program; one where we started out by using,just the base metal materials and studied the effect of notches in the base metal material. Later on, we produced weldments in which we actually introduced notches and tried to make a comparison between the kind of elastic stress concentration factors that were in these notches and in the metallurgical notch that occurred in the weld itself. 5592B-444B:2 (S3597) 2 6 L

It will be shown that when a geometric notch of a given K is introduced, t the specimen will not fail at the notch, it will tail in the heat-affecteo zone of tne weld, where tne metallurgical notch occurs. Another question, CS 210.1, says, " Discuss the local deformation and geometric discontinuities and evaluate the creep rupture and fatigue camage at geometric notches." This is addressed by work in which notches were made in both the base material and the weld. The conclusion is tnat. creep rupture and fatigue strengths for notcnes unoer either biaxial stress or triaxial stresses are greater tnan tnat of the minimum cata used in determining the ASME code allowable stresses for ocsign. Just a short review of the Code design criteria (Figure 2). The lower of S t and S IS S Tne stress to rupture turns out to be the most restrictive m mt. criterion for a given time t. So this, really, is the controlling criterion thich was seen in some of the earlier tests. Rupture is usually more controlling at these elevated temperatures. be have the strain limit and will take a look at the kinds of strains that actually occur at geometric discontinuities to see if the strain limits that the Code uses are conservative. The strain at welds will also be checked since the strain limit is only one-halt of that for base metal. The other criterion is that the weld metal strengths shall be greater tnan tnat of the Dase metal as well. Ana then, of course, the strain limits were reduced by a factor of two. The basic specimen contigurations are shown in Figure 3, but for the base n.etal, a simple tension test is run on the materials. Another plate specimen wnicn is used is 8.5 by 85 oy 76.2 milimeters witn almost a square cross-sectior. The advantage of this specimen is tnat it is large. The specimens are reinforced at the ends, so plane strain conoitions are maintained at the end. Now, the plane strain does drop off toward the midale of the specimen, but a state of biaxial stress exists in this specimen of uniform cross-section. 7 55926-444B:2 ($3597) 3

4 i To simulate a notch, a hole is made right in the midale of the plate, whicn j has an clastic stress concentration factor of 2.7 for the wioth of specimen used. The width to thickness ratio of these plates is ten to one. For a l plate with that ratio, the stresses in the thickness direction are essentially zero, and not too much lateral strain is obtained. However, tne stress aoes build up in that direction. .i I The stress distribution obtained in the plastic case arouna the hole will be shown to describe the relationship between the means of effective stress ana the peak axial stresses that occur in such a case. For triaxial stresses a circumferentially notched bar with K of 4 was used t with two notches in the bar. The purpose of such a sample is tnat it can be i loaded for a given time, say, 30 or 40 percent of life, and then taken out, j cut in two, and the other tests can be continued while the first specimen is l examined for cracking. Some of the results of those tests will be shown. I Also tension and bending can be introauced into the test. There is interest 1 in knowing wnat happens in the case of an eccentric loaa on a test specimen, I and how the stress and the strains vary in this type of specimen. There is a j discontinuity, of course, at the ends, but there will be a nigher stress on the tension-tension siae than on the tension-compression side. The strains in such a specimen will also be discussed. Finally, there is shown the plane strain specimen. We have plane strain at i l both ends and the specimen is quite short in comparison to the others, so a very high degree of plane strain throughout the test is obtained, and a higher transverse stress is developed. 1 j The weld test is made in a rather unique manner. A welo is placeu right in the middle of the plate to duplicate a circumferential weld in a pipe where all the parts of the weld are strained the same overall amount from end to end. So all of the elements of the specimen are given the same strain, because the testing has shown that a strain criterion is a better criterion if j damage is being considered. However, we're going to look at tnat in comparison to the Code rules to see it the Code rules are conservative enough Sb92B-444b:2 (S3597) 4 8 i

l i when their particular method of calculating is used. It may not be the best method, but it may be satisfactory for design. In the case of the plate with the hole, again a welu is put down the middle, and then a hole is drilled to take out half of the width of the weld. Thus, a defect has been introduced.' When we have transverse welds, we just put the weld across the plate in the same place and again drill out half of it. In the case of the notched bor, we put tne welu between the notches and put the root of the notch in the middle of the weld. There are some cases where i tne root of :he noten is placed at the neat-affected zone. i figure 4 shows the stress situation for the flat plate ana the plane strain j conditions. The plate, axial stress, width stress, and thickness stress are shown. The width-to-thickness ratio is ten. The axial stress is the nighest, { the transverse stress and the thickness stress are essentially zero. Also shown are the effective stresses and the means of effective stress. Plastic strains will be considered because fairly large strains, at least ten times the elastic strains, will be encountered. A square grid of lines was placed on those test specimens so that the strains could be measured throughout the whole plate. If the strains throughout the plate are measured in any one location, the strains throughout the plate can be compared. It the ratio of the strains is measured, some idea can be j obtained of the transverse stress and then for the particular loading and I geometry of the test, it can be determined if inelastic analysis produces the some kind of strain. ] By measuring these strains, we can actually get some iaea of the stress and of l the ratio of these stresses. Note that with this plane strain condition, it tnis is the Von Mises effective stress, the axial stress will actually be higher for the plane strain condition, which has a 0.5 stress ratio. 55928-444B:2 (53597) 5 9 i

Tne two grapns on tne right hand side of Figure 4 show inelastic stress ano strain components for different plate widths and plane strain conditions. For example, the stress ratio o,/o, = 0 (lef t nand ordinate) corresponds to a uniaxial plane stress condition; the stress ratio o /ca (right g hand ordinate) corresponds to a plane strain condition. For plane strain, tne widtn strain is zero. The other two strains, the axial ] and transverse strains, are equal in magnitude but opposite in sign. And so ) we will see if we can actually measure these strains and see wnat degree of plane strain we have, although we can't very well measure the stresses. Figure 5 shows the situation for the plate with a hole in it. This is a plastic analysis made by Evan Davis. The highest stress point is at the i hole. We have an elastic stress corcentration factor of 2.7, and, after plastic strain, we take a look at the axial stress distribution away from the hole, going across the specimen. Tne abscissa is the distance away from the hole divided by the half width, which yields a dimensionless number. The axial stress right at the hole is about one and a half times the nominal stress. The ordinate is the stress over the nominal stress. The radial stress, of course, is zero at the origin but increases a short distance away from the hole and then decreases to zero. Thus, we have the i 4 hignest triaxiality conditions at the peak of the (radial stress) curve. i Knowing the two stresses--or the three stresses, actually, since the tnickness stress is zero, we can actually calculate the Von Mises effective stress across the plate. The effective stress will still be highest somewhere rig'ht near the eage of the hole. For the triaxial stress case in the circumferential notch, Figure 6 shows the axial stress for the notch. The way the axial stress redistributes after the plastic flow (and subsequent stress redistribution) is shown. The stress right at the root of the notch is seen to be about 1.1 times the nominal stress. A little bit below the root of the notch it is actually up to 1.6, then it decays as shown. Also shown is the effective stress after creep. It is about 0.85 of the nominal stress, and of course, smaller than the axial l stress. 10 5b92b-444B:2 (53597) 6

In the last two figures we show the analytical predictions at the gross geometric discontinuities. In Figure 7 we see the experimental evidence of plastic and creep flow at the geometric discontinuity, and how the strain, and consequently the stress, redistributes due to this material nonlinearity. Figure 7 shows what happens in stress rupture to a plate with a center hole under biaxial stress. The original specimen with the hole and the square grid of lines on the specimen are shown. The material is 304 stainless steel and the test was performed at 1100*F. After 1,030 hours another picture of the specimen is taken and the distances between all the lines are measured. Thus, the strains can be determined. It was possible to ascertain first when the first initiation of a crack occurred and the strains were measured then right at the crack and in the material first beyond the crack. It was found in these tests that there are two shear bands that occur due to the high stress. They are visible on both the end pictures althaJgh they are small in this particular case. Thus, two shear bands appear at an angle of about 60 degrees. One band forms right at the tip of the crack. The crack right behind the two lines actually is a high tension point and it jumps from that point into the high tension field, relieves the stresses, and renders the area stress-free for a while. However, the creep load is still being applied on the plate and the plate continues to extend. So we followed the crack, we found that it i jumped a little bit, formed a little "V" pattern, then it jumped again. Thus, even thaigh a static test was being performed, there was a cyclic stress going on at the crack tip. The test was actually stopped and the specimen photographed during the test to get numerous datum points. In other experiments a hole was made in the furnace and pictures were taken during the test. The results of one of the tests are shown in Figure 8. The overall deformation was measured arti the conventional creep curve was obtained for this plate with the hole in the middle. The crack opening displacemeat at the i hole is also shown. The hole becomes elliptical and actually moves in a little bit, also, but it gets a lot longer as the material creeps. 5592B-444B:2 (53597) 7 11

r 4 j One of the objectives of this work was to see what the discontinuities of tne curee could mean in terms of the crack growth rate. A steady creep rate, steady open displacement, and steady crack growth were observed in tnese tests. The specimen was cracking, stress redistribution occurred and then the stress had to build up again. So tne process is not one of an accelerating crack growth rate; it is actually a very slow growth rate until a certain j point is reached. At that point, the crack starts to accelerate, for several l reasons, one of which is that the net section stress is increasing because of l the crack propogation. i So, we can measure the crack growth rate in these specimens where the crack has been initiated at a hole. Notice that it initiates at about the halt lif e of the test, then it grows rather slowly, slower than the displacement, ano, finally, it accelerates near the end of life. l Figure 9 snows the results of the measurements of rupture times for 304 stainless steel for different types of specimens. The Code allowables are shown in the same figure. The upper curve shows the results for the notched bar with nominal stress plotted against time to rupture. Notice that the 1 notch strength of this ductile 304 material is higher than tnat of the i unnotched base metal. The effective stress in the notched bar was only 0.8d { percent of the nominal stress, so if the actual effective stress were plotteo, the curve would be moved down somewhat. t i i We also have simple stress test data (points A and B in Figure 9). We nave comparea it to the uniaxial bar cata obtained by ORNL on the reference 304 stainless steel, and we find tnat our tests ouplicate their results. In the cases of a plate under plane stress, the effective stress was higher, as were the axial stresses and they would have to be plotted higher. The results in Figure 9 seem to indicate that there is an apparent notch strengthening (with respect to the uniaxial ORNL data) in the case of a l notched bar specimen; whereas there is apparent notch weakening (with respect to the uniaxial ORhl data) in the case of a plate with a hole. However, we 4 i should point out that the stress plotted on the ordinate corresponds to the i net section stress and not the peak stress at the notch. In tact, according $$92B-444B:2 I2 (S359/) 8 i

i j to the Code, the peak surface has to be used to compute the t/T D (creep-rupture damage) that would accumulate at the geometric discontinuity. Thus, the actual peak stress is the nominal net section stress times the stress intensification factor. To confirm the adequacy of Code calculations, detailed inelastic analyses have been performed for the C and D specimens. We have shown that the stress rupture in both these specimens can be conservatively predicted according to the Code method and the uniaxial creep-rupture data shown in Figure 9. A simplified stress rupture method has also been developed (instead of performing detailed inelastic analysis) to account for the multiaxial stress state at the geometric discontinuity. There are several kinds of tests which were run. There are step-loaded tests chich start out at, say, a stress of 20,000 and drop down to 10, 15, 20 and up to 22, and these variations usually duplicate the final stress conditions. The Code data are also shown in Figure 9. They are multiplied by one and a half because, as discussed earlier in Figure 2 point 3, in the Code a factor of two-thirds is applied to that data. So there are the minimum data used in the Code and I've given them at 10 to the 4th and 10 to the 5th hours which aill be used for comparison purposes later in this presentation. It is seen that all of our test d ata, just based on nominal stress, with the discontinuities, lie above the minimum Code values. Furthermore, as discussed i earlier, detailed as well as simplified Code type analyses have been performed ] to predict time to rupture in specimens with geometric discontinuities. In all cases the predictions are more conservative than the experimental l observations. Since there has been some interest in creep fatigue interaction, we present some results on fatigue in Figure 10. It is possible to get some pretty drastic results on high strain cycle fatigue with long hold times. These results come from Ben Lazan's work of about 25 years ago. He plotted the ratio of the alternating stress to the fatigue strength at two times 10 to the 7th cycle, so the abscissa is non-dimensional. Then he plotted the mean stress as a function of the rupture strength at 100 hours. So, we have a mean stress and an alternating stress, and the Code goes through some kind of a a SS92B-444B:2 (53597) 9 13

m linear translation and, indeed, at room temperature that's a pretty good representation. At elevated temperatures, these materials are rate sensitive and we find that the creep and fatigue are independent. This justifies the linear damage summation rule that the Code requires since the creep and j fatigue are calculated separately in many cases. i j lt can be seen from this that there is actually a benefit in creep insof ar as totigue is concerned and the structure can stand tnis amount of alternating stress and not suf fer a change in rupture life. how, that is a substantial l anmunt, 40 percent of the f atigue strength when the specimen is subjected to I full stress for rupture in 100 hours. j In Figure 11, the tests that were run were a little bit different. Again, j there was concern about creep fatigue interaction since, with creep rupture, atter about half of life, a crack develops and if the specimen is then l fatigued, the crack cen propogate rapidly. Notched bars were tested and it was found that they Cracked at about 50 percent of life. They were run out to I between 50 and 90 percent of life in creep rupture and then they were fatigued. The fatigue life is plotted versus the percentage of rupture life. The results for 57, 65, 85 and 90 percent are shown. The specimens were creep 4 tested out that far and then run under the same stress as used before. These specimens seem to be much stronger. Thus at the base of the notch, stress redistribution took place due to creep which proves to be beneficial. The peak stresses have been moved subsurface and the strength has been increased. 4 In these tests there were double notches, so one of them was fatigued ana the i other one was used to measure the crack length. Inaeed, cracks were not seen until 50 percent of life was reacned. We looked at the size of the cracks I that occurred shortly af ter 50 percent of life. So, we are talking about cracks that are pretty big (16-33 mils), yet all of the samples appeared to be stronger than the virgin material. So, these results gave us a very 9000 ) feeling again about creep fatigue interaction. l 4 i l 55928-4448:2 (53597) 10 14

Figure 12 summarizes our results on the base material. We conclude that creep repture and fatigue strength for (geometrical) notches under biaxial or triaxial loading are greater than the minimum data used to determine ASME code allowable stresses. In Figure 13 weldments are considered. We first want to look at the acceptance criteria for the materials used in weldments. How can using weldments be justified when all that is done is to use a Code value which is based on the base materials? How can we use those values and still apply it 1 to our weldments? What is the creep rupture damage that results in the i residual stresses? To answer these questions, large specimens were made in which the residual stresses were retained to evalute their effects. We will show you how we evaluted the effects of metallurgical notches that occur in the weldments. i 3 The conclusion was that the current Code rules and the acceptance tests that are applied to the materials are adequate to be used for the materials in the weldments. The residual stress relaxes in service, but the resulting plastic strains are less than ten percent of the creep rupture ductilities of the zone of the weldments. That is, the plastic strains resulting from residual stresses are one tenth the plastic strain to which these specimens were subjected in the 2 actual tests that we have performed. I 1 The strain concentrations of a metallurgical notch are accommodated by t he ductilities of the zone of the weldments and we have actually measured the strains around these metallurgical notches. We took a look at the cracks that were formed and measured the strains ahead of the crack to see when the cracks jumped ahead. These strains were then compared with the Code allowable strain c limits. The residual stresses in a weld were calculated by a finite element anal ysis (Figure 14). One zone was layed down in the weld and the stress resulting after solidification of that zone was calculated. This is the circumferential weld in a pipe. The next bead was put down, the next and so forth, until 55928-444B:2 (S3597) 11 15

seven beads were layed. The finite element analysis was run all the way from the melting point on down making reasonable assumptions of the properties. Then the residual stresses that were measured in the specimen were compared with the calculated stresses. Now, it is seen that there are high tension stresses at the center line of the welds, a little bit higher stresses at the fusion line, and then the stress drops off and becomes compressive. The analyses that were made are shown as solid lines and the measurements made of the residual stresses are shown as points. It is felt that there was excellent agreement between the two considering all the assumptions that were made. Figure 15 shows a sample where an axial weldment was layed down the middle of the plate and the sample was allowed to creep. The effective stress is actually highest in the middle of the specimen at the center line of the specimen, or mid-length, and drops off toward the ends. Plane strain conditions are set up in this test. This test was run for a number of hours and then the sample was examined. In the heat-affected zone, a little crack had formed with little V-shaped shear lines at either end. The test was run 6 little bit longer and, right below the first crack, another crack formed. As the sample crept, when a given straio along the length of the specimen was reached, a crack formed, almost at exactly the same strain. The average of all these strains was taken and it was found that it took four percent strain in this specimen at the heat-affected zone before a crack would form. Figure 16 shows cracks at the HAZ's at two times in a test of a plate with a hole in the weld deposit. We know we had the strain concentrations of the hole at those points and, in the case of the base metal, we saw the point of crack initiation. So, we have a geometric discontinuity at the edge of the hole and we evaluated that. If there were still the high strain concentration l at those points, the sample ought to have failed there but it does not. What happens is that the two shear lines form and where those shear lines cross the I heat-affected zone, that is where the specimen fails (shown by arrows in Figure 16). 16 55928-444B:2 (S3597) 12

50, the cnaracteristic points, the areas where the sample failed, were located dnd the strains were measured. we can teli tne strain at wnicn it failed. The beauty of the test was that the strain at which the sample tailea was neasured and the strain concentrations which occurrea were determinea. i The strains were measured throughout the specimen and are shc T in figure 17. The strain lines emanate from the center and go clear out to the fillets et the specimen. The hole is shown with the hign strains around it. From this 1 strain contour plot it is clear that even in the case of the plate with a hole i in the welo deposit, it still took around four percent strain at the heat affected zone, as the shear line was going across, to get the specimen to fail. Figure 18 shows what happens when the weld was made in the transverse direction. This was supposed to be almost a straight-sided weld, but it was slightly V-snaped. Again, the strain lines are shown, as are the f racture locations. The strains at the tips of the cracks were measured again at tour percent. The sample fails again at the metallurgical notch. Thus, I think we proved conclusively that there are metallurgical notches in the welas ano that it takes around tour percent strain before the crack will 1orm. Figure 19 shows that the crack growth rate in these welus was quite slow, l also. We found about the same patterns for the two weld coniigurations. Data l from ORNL tests and our tests are presented in this figure. The results of our axial weldment are higher than the base metal. The weld metal is a little bit stronger. The crack initiation points are almost all on the smooth bar data or the reference data for this material. The tests in the transverse weldments are also shown. It was thought that the crack growth rate would be much faster in the transverse welds because the crack was going right down the weak zone. This was true with the high stresses; but at the low stresses, it crossed over, showing that it is really harder to get tne crack to grow down tnrough the heat affected Zones. So, there it something about the transverse welds that surprised us. I %92B-4446:2 (S3597) 13 17 l

Again, the Code stresses, the minimum Code strength, is lower than that of any of the samples in this work. The notched bars, of course, have much higher values. It was very difficult to tail the notched bars with these notches in the welds, although we did inspect them and found that in each case there were subsurface cracks. 4 Figure 20 presents a comparison of the stresses for 10 and 106 hours for all of these specimens. The comparison is made on two boses-one, just using the net section stress; and one using the gross section stress. Remember that the allowable stresses are two-thiros of the Code values. The base metal was three percent higher than the code materials, so it is just aDout representative of tne minimum for the code materials. That may not souno very good, but it turns out that this particular heat was solution treated twice, anu everytime these materials are solution treated, their strengtn oecreases o little. However, their strength was still higher than the base mdterials f or the Code, tnd remember that. the actual design stress is two-thiras of Code value plotted in Figures 9, 19 and tabulated here in Figure 20. l The axial weldments were 42 percent higher. The transverse weloments were 39 percent higher. The plate with a hole in it was 29 percent higher based on the net section stress, and the transverse one was 38 percent higher than the code. It was 16 percent higher based on the gross stress. If you just took the gross area, then the transverse welo was 25 percent higher. The notched l samples, of course, were considerably higher. They were estimated to be 200 percent or 100 percent higher. Even the notch at the heat-affected zone was around 10 percent higher although we had figured that that would be the weak section. i 5 The same thing is true at 10 hours. it shows that tne strength of the l weldment without stress risers is greater than that of the Coce aata for tne l base metal. The strength of the axial and transverse welaments with the hole within the weld is greater than that for the Code value f or the base metal when based even on the gross stress away from the hole. Lastly, the notch strength of the weldments is greater than that of the Code value for the base metal and the code design stress allowables are only two-thirds of the numbers shown on the chart. 18 5592B-444b:2 ($3597) 14

At this point we should point out that there have been some other studies which show that the weldments are somewhat v:eaker than the base metal. Consequently, at the present time some discussion is going on in the ASMi Code Committee meetings on the weld strength reduction factors. The reduction 1 factors that are currently proposed are significant only at temperatures significantly above the operating temperatures expecteo in the CRBR plant. There are some Code committee members who feel that, when taken as a whole, the Code provides adequate margin of safety against creep rupture in weldments. They point out the allowable membrane strain limit is 1/2%; whereas strains of about 4% have been observed in test results that we have presented. In addition, the stress-to-rupture curves have also been lowered in elevated temperature Code Case 1592 compared to its predecessor Code Case 1331. At this point we can make the following observations: o The tests presented show that the weld metal as well o the base metal is stronger than the minimum (uniaxial) stress-to-rupture curves utilized in the Code. As is discussed elsewhere by Dr. Brinkman,* there is more scatter in o the creep-rupture strength of the weld metal than there is in the metal. The tests shown here indicate that the weld as weil as the weldment is stronger than the base metal. Figure 21 presents our conclusions from this work. The current Code rules and acceptance tests that we use for the materials are adequate for the materials of the welds. Residual stresses do relax in the surface, but they result in only a plastic strain that is very small. The creep rupture ductilities of these zones of the weldments are around four percent for cracking, which is a lot greater than any of the strains from the residual stresses. Finally, the strain concentrations of a metailurgical notch are accommodated by the high ductilities of this material.

• C. Brinkman, " Materials Data Base", W-ARD, ES-LPD-82-008, p.ll.

5592B-444B:2 (S3597) 15 19

/ FIGURE 3.0-1 Michael Manjoine Westinghouse Research Lab Consultant Presentation to NRC April 7, 1982

• +

-e COMPONENT DESIGN FOR ELEVATED TEMPERATURE CS210.1 LOCAL DEF RMATION AT GE0 METRIC DISCONTINUITIES PUfiPOSE: TO EVALUATE CREEP RUPTURE AND FATIGUE DAMAGE AT GE0 METRIC NOTCHES 1 CONCLUSION: CREEP RUPTURE' AND FATIGUE STRENGTHS FOR NOTCHES UNDER EIAXIAL AND TRIAXIAL STRESSES ARE GREATER THAN MINIMUM DATA USED IN DETERMINING ASME CODE ALLOWABLE STRESSES FOR DESIGN f APPLIED TECHNOLOGY Any further distribution by any holder !of this document or of the data.therein .to third parties representing foreign g interests, foreign governments, foreign

companies and foreign subsidiaries or foreign divisions of U.S. companies should be coordinated with the Director, Division of Reactor Research and I

Technology. Di>partment of Energy. 20

.2 FI GU RE 3.0-2 L ~ _: nrc e.

o. r,s s m F

r n '.o ; : y,-. :.:s S T F.. a 5 L1'.4 c v'. AT p ysv T r.v.o. .ys r. s. ey.4 j

v. ;...,.

0.9 5 MIN. AT ELEVATED TE";. y S w. TE,:' L: 1/3 Sv ".'N, AI RDD" TE",P. I ',- s...~~< E.37 5 M:N AT ELEVAiED TEM ~. = i_ v 1 i q :. : "., D/3 S MIN AT El.EV. TEMP. R;:1 j E IN Tl"E t.

.S S MIN. AT ELEV. TEMP.

Tc sz F TREE; IN TIME t. m F i. CREEP 1% CREEP, STRESS MIN. AT ST RE ',:'- ELEV. TEMP. IN 11ME t. STRAIN t1"11TS E Ex:Est: VE DE;DR"ATION 1% ME"2RANE 3 F DUE ;:, l',:REv, ENTAL 2% EENDIN3 (DLLA?SE AND RACHETING 5% LOCAL K /. LOSS 0; F,:T!DN DJE TO E x:E SE ,T DE F DR'%T I ON LJ1DELINE ~ g. BU:KLIN: D'E TO SHORT TER", LOADINGS GUIDELINE 9. C;EED E_:CN3 DUE TO LON3 TER"i LOADINGS GUIDELINE WELDS: A) HELD METAL STRENGTH GREATER THAN THAT OF BASE METAL e) STRilN L.M!TS REDUCED BY FACTOR OF TWO e y m-E ?. t

-~ c... i:n a P, 's P A. Uniaxially Loaded f[ Plane Stress @p Gage Section: 8.5 x 17 x 76.2 mm N B. Uniaxially Loaded Flat g \\ x Plates with End Restraints P 8' N 'P N txwx1 8.5 x 85 x 76.2 = %s e C. Uniaxially loaded Flat Ps Plates with Holes 8.5m= eia. N K ~ 2. 7 p t 8.5 x 85 x 76.2 m P D. Notched Bars e K ~4 5 " 8"di"5 P t 16.2mm & 24.4== Diameters N. E. Tension + Limited Bending N Pp of Flat Plate p ,. 9,1mm 8 8.5 x 85 x 76.2 mm G. Flat Plates Under Plane Strain l f gN 8.5 x 8 5 x 21 mm 1 N P FIGURE 3.0 Basic specimen types 22

Curve 715613-B ,,,,,ii s i n u n n,

1. 2

\\o a n h ( 1.0 o r -u u m U a a 0.8 ~0 ,/ u

0. 6 o

a t i d d O 0.4 o ~ m u <v , 1464644" 0.2 Width, u =10 thickness t -et 0 '\\' C >U >U a u t 0 0.1 0.2 0.3 0.4 0.5 Effective Stress. g fg u'1/2 2 o a 2 ~0 0 +0 m"I U a au 1.0 mEm E +E +E=0 g/ a o t a Effective Strain. 0.8 (_g 3 (E ~ u + E = a u i m - 1/2 2 0.4 + (E -E 1 (-E t a u E 0.2 e. o m m u I 2) E

• a 0

a o* 0 0.1

0. 2 0.30.40.5 0

m g g"_ 2 ) u a g = o m E:- (0 g 0 a u m FIGURE 3.0 Biaxial Str05505 and Stiains for plate models 23

FIGURE 3.0-5 W ESTINGNOUSE EL ECT RIC CORPOR ATION 1.75 axial STRESS \\ 1.50 \\ 'N 1.25 - STRESS DISTRIBUTIONS FOR PLATE WITH l. 1,gg HOLE AFTER PLASTIC FLOW 0.75 O.50 RADIAL STRESS 0.25 - g 0 0.02 0.04 0.06 0.08 0.1 DISTANCE FROM HOLE HALF WIDTH 24

5' E T C

• 5" E E T
• E5TINGHOUS E EL ECTRIC CORPOR ATION i

~i, i l i I '\\ \\ I ,l i l 3 l Elastic bial Stress STkEEE P/A 2 i / / Axial Stress after Creep CENT E? / t,.. 1 4 l, b ~ ~ - %'N L __ _ Effectice Stress after Creep

at c h ROOT

!? K=5 t 0 0.1 0.2 0.3 0.4 1 X/r i S!RESS DISTRIB**TIONS FOR CIRCGTERENTIAL NOTCH IN A BAR FIGURE 3.0-6 25

7 f. '. aw-O ' . "EE '= F#*-_** P.a We S. = Im a. ' ^ f.% :x-W$1b5b?W' h: =--e*^ 2 i%.I.c '.5 ?q -- $ $$$~fdD0-5!EW Q .-..._o ..y }-.n.o-.~d .. k=z - s,:) a. ~r. _. ;4.. m.xxr n- ~?~--.-=,= _ - _ _ = _ - -= :%__;_ -.... - ., f, & W,; '.*....' ] K r '-? &,.?I,., ---.u._ c--. ?- @ W F 5 ~.3^EiF-;5 X -.:.Y 3-au - s g. i?fE.; c. rw- - ~ h ~~ :

S $~II U~ YW

5 a 5+M-3,_:-.I 4 NIri-S-e i8L.5 J"I~ $:Yr.~ ~ EDM.h,g 4i

?

-4-2 %=;E c g;;:g _---(-

z z?_- 2 c [C* ki" F* E;* r -t -* i &MI-b -T 1,5 s W 'qe,.,8 = - i=s=9 4&:- *'l=iF'9 mt, e.-- t. ~ ir*O,5.*7". _".7'"tr"= 7 3 ; '. r, "ge"7* %.9 r, ' 78 E '" O *=..:./ * -s F-- ..z ,-. i L' = Y__= } ^:- Y _Q [' &_. *+2'~== W._+'^.=? F_ vs-. k M'.M.,Na.;y.14j G. -$ 17 ~ gf.,.:.- t : - -=...m .. ra... Em .1 > ;,r

• ?-
• - - -- ~ y =J t ~ i'I=, =.k5^s: --- ~ ::p z~-

'- M . C 3, C - W N, E W.=IE _=e r C g~ b '.3[*'f2b.N. hN,. ,J..M: r7.I w-nr -3 _ g ~ m%. Y ;;.'=.~~~'^f'~ *E ~ -.,51 h yn & & & $.) "% p --Q{(D h F. Ef.. ;.~ :.m. - ---c -~ di ={.h_ '^ E br-I ),m =E '"h D.EpN. b k **' /:d*555' QW'**.,

==

'. + - - . ~ '.! ~=~~!'r: c ~2 ~.=_u ..7.:=: y' '_.'=~W;2ii :% e mr pf.p',3 := sx --& v.?. sic. '. V.. ~+h. &;e m r-

v. #i:-r 4.'." *

• p.

. =

==---m a - s %..r _mJ= a.r.. t ~-

-..-=c:: m 11t 4

x.G ~-r.+ i12 : _1 .- m cs

~z-s:a - n: an s m -

_ =.. _. -~2- =r==_- C v usE*:.u O_=_rp g.e m --r_r-h t p' *e:-j% "* % -iiMsr.@d,7.pU. O.c *rf. I "~+ - m i'v :5 2 - =-;=4.y~t jfn 44 Q: h:- - =.gra-m.7 :jij r .Y r L - i 1.n - m pm

ms.

M 4~ WF'- ,p:r,23-8 _; -.w== g v.=-_ :*_rw f =:;;i . ;-. a; e _ s -- ;M " ",3 i m- -c r :n ia...4T8 8 e-e .i s m _ _ g 7.. n,=-,2 u.,rT' ' l*25 }.a j.Q EM p.~K r--{ -- hi -,rwrpp,if ,g y,ep.".-s-- -- r ' ' y :1: _ ;.. 5 p.., L. wls.re. * {M- __.:'rg i 7, m - -G a

3. pmF2 s d W -:M_.c --+=.M_. ii:

En !' h; ;:

M ~s s -C. : y .g 3-m E i.. :- '.,e w ?="*y..

y

gr..s=r pr m

. i ..p 6+ a ~ y.aus. v - . eq , m v. =.. r

m.. z.

.g Yi IM I d. !)

• =

-:=. u.: .: : q E b b"*'.$N !.M -d(*h-INM er.M_.[*s-j*g-Z -~h tQI ~~ r.14[:i. ~ h, [ =i ~h * J -I* f., C L.,.' ~. i Y- ~ ~ - C E -'~52.i % =3G.)IN[*y 5Eif-

• j... ;

J-N ~ ag*$ f 5r Rg e .*N E.or. r'.:=4 rs% it-=?. w,[E_r..5.i3 : g .w.-w Kin n .r r.:= :n-= 1 e/ ay.-:.mse. m. -n._f (b) 1030 hours (a) Original - TG.w== MUG;s-===Zna455 RiW %,M M M==-n M.ns2@- -.. -. n = = = ~ - = a.=f-:-r e.- w2 ;.-. 4 2. g ;,, -- m.,ramyg = -== s a I 2=.'x =1=_=n -=

• m-2:mw mJu' : ~==- w}Q 8

~" r n,

==. n.=n-n_=ew =:m:.= 'k, . N ~. V N-:: ;44= m*-i ' _-f25= A;A=.9-=.-zw.2==_= 5522i=% t -- .h _:.crm- = =aam s.=~u.g i r _ w=.p-m=.- p =r;s=_ -mx=.m==n=:-. a'a =

.:1 m += m-

~~ ^ , vs-.'rE= -? ~MFM: =_=U+ V2'=' =22==5R .= =.==X=;i q= UW = .= m t= n c= ?-r=_u : f +=:=iE ";a=Y. ~r'# . :.2 = y==gn. .=+ m %m'k-5~ 5=4~_EM*i;iYEC,'# ~ s.=. mmr m'E~ MT.~ -.-@ - E r- = 3@g ;p. '4'Eii. Y:_;=w=i'.i-EJz*'r.~" J'-~-- ~.)g j,., Qg?g :-- 2.f..rt--.f_=&.= =K=., -ems = e.w w m m= xD.=Mi-f' =: E. =~p==r. 1"=E E-l =L-'r=_::: d-- =r

----,- x -- u

---r--- n%f'=1%ggy?- s=C -w Q un? -= - = _= .n

-: a. _.c.,, e

E,4 i~ ~+'-Ei y

=m-c -- z= : _ _ _ - r f 5-8dk

=_5

.2.-

v: c. 3; v .w_= .f ="__-QMkrg 8! 2 s.wm_ s. =._.,%_. 2w. _, _e._. _-w._ wi=r-=,#..:w._..;!s!. : %/=,: :,.1f... _._=_,2_.__) 2 .=9 e..,;;:r g as-f c.;e. ' c - -2 = 4 ;m;;- g -Qm==A+---l2__% i-.j;'tq - '. 5 '..: rru-g=-r' ...; _ y _; y z_ . =s-m...c -n. -

a

- - =

w. r' =w%.4 5-d'~:al'Y'- g;e;se Q ess-s=m-as_ e. + :

~hX U EX - y;. ? - aEJ-r h-* sk 4 =-a

s.Cc e n-2---

i u. w:= ' % ~ 9 :ii-f f 2 h,:s:P ..q. c Lr as ^f' - r =- =5""=.&=-~.: ;. e.:=..~2s-13 ' ^ ~ = " . wi =..T %+e? 2%,.-~%e.: *-G --s'.E====&=4: M-

1. ~p t 45 ::r.,_g: -~ -- fen =-.

- ' ' = 1. = r.s --

1. =

=4=,= g - =r W.W 2R$: k.,-dan-+, m;L 2+ i

==*2 -1 a - 5. mA ;; =v'.% -My' icc.:- ar ,=~m m i -' g -.-- r = --e,, w'- ^ dei +%...f.h=.r- ~Mdc erU, Y% a & u;yr us sg- , y 4r X1a:'"~..m, y= .r

d. Y:W.MJ J

?.~. - ~~~.ft1 g.e_,q. ?r.W. f- ;_ g-&. g. o.E de e,4l'H %=3 - r .Jt"im ~~~ . ye.ns M._i.=d W 4 2:t, @,e %.4 u __._=:_u _,.u++ i=@c 1 t

.,m

.a- . = = = ..,7 -= (d) 4005 hours (c) D3 hours - Photographs of MMel C4 showing distortions and cracking up to 4005h-FIGURE 3.0-7 26

4 -2 mm 10 in. cm' in. 7.06 dm T 304 SS .278 in 4 -12 Plate Model With Central Hole 3-r 18 ksi (124.1 MPa) and 1100*F (593*C) p -5 L 64 x 10 in/h E 0.42 pm/h 5 E a -0. 8 % j 2-- 8 p " 2-g@ 1.0 x 10 'in/h g g9 P 0.254 m/h --0.63 P 4g l-- 0. 4 1--4 t t O P ged g#2 - 0. 2 \\ g t* g9@

2. 26 pm/h

~~ g tte , CtM, Ged 8 9x10 in/h 0 0 1000 2000 3000 4000 5000 6000 Time, hours FIGURE 3.0 Creep deformations and crack growth curves for Model C4 I 27

forv7 688111.A i i i i i i i i i i l Model Code .30 D A - Plane Stress a 2m Notched Bar B - Plate,1/3 x 10/3 x 9/3 in. O - =2.7 0 C - Plate With Hole, K,4 28 7 D - Notched Bar, K = O g i E - Plate Tension + Bending V t i G - Plate 1/3 x 10/3 x 3/3 in, h - 17 5 26 ~ ' D Step loaded Test, S 3 RNL D a 7 TM 3845 ' Rd. Nat 9 T 2796 0' h j24 E y, Reann. 2000F f 1093C) y 22 - 150 $g _ c s a 8 D 20 Plate With Hole i - 125 18 g5 iiii i i e i i i i i i i 4 5% 10 10 10 10 2 Time in Hours - ion

• ~

i. Fig.29-Rupture data for basic specimens of T304SS at 1100*F (593C) s c...:< / N i.s s,,,, oi s n.- 75

1 i I ( UAe.ich.s So.cim. A. = i.e j TQs f 650 F (E99C) i 3 NQ \\ \\ E k \\ og 2 k h N w\\n\\ .500, q 815C, sg !O 75 F \\\\ l i E*e (24C) \\\\N g 1350F 7 p =e 0.4 (732C)l\\ l iff g 7. Ns i 2 0.2 g g e s O O O.2 0.4 0.5 0.8 1.0 n.ii..i........n..... si... 1 10 0 h,.. i t... n...i,. i., a.is. iO' c,.i..) FIGURE 3.0-10, STRESS RANGE CURVES FOR ALLOY 3-816 29 4

i PE: ICE *,T OF R'.TTURE LIFE O 23 43 60 ED 103 i i 93 ' ~~~'s' 5 tc' 10

• 165,

)2A 's o' 's e4B $ s.22 A s \\ 16 100 hrs. ,334 10 i 1 0 4 8 12 16 20 24 25 32 36 O E 45 (D' 9A~ 93 5 10 's 16A* s '/ 125's I2A 45 2 \\

• C Viaiensile Test j10 lig\\

6 g 22A a C Via Fatigue Te51 g w 0 4 E 12 16 20 24 2E 32 36 O E 45 ~ 3 Crack Lencin, C, mils- !d. 9_5 ~ o 5 ( Cl Ib-10 12A E i ~ ~ o22A a 10

• Dim., 6( A or Bi l

," f ^, A E xt., 3 6 ( A + B ) l 0 1 2 3 4 5 6 7 8 9 10 11 12 Notch Extension, 6, mits FIGURE 3.0-11. FATIriUE LIFE Vs PRIOR RUPTURE LIFE AT 649C 30

FIGURE 3.0-12 i COMPONENT DESIGN FOR ELEVATED TEMPERATURE CS210.1 LOCAL DEFORMATION AT GE0 METRIC DISCONTINUIT PURPOSE: TO EVALUATE CREEP RUPTURE AND FATIGUE DAMAGE AT GE0 METRIC NOTCHES CONCLUSION: CREEP RUPTURE AND FATIGUE STRENGTHS FOR NOTCHES UNDER BIAXIAL AND TRIAXIAL STRESSES ARE GREATER THAN MINIMUM. DATA USED IN DETERMINING ASME CODE ALLOWABLE STRESSES FOR DESIGN 31 i n.,n-. 7 ,,.-~.. ---,, n,

1 FIGURE 3.0-13 HTS MATERIALS AND STRUCTURES CS250.4 STRUCTURAL EVALUATION OF WELDMENTS AND ASSOCIATED MATERIALS FOR ELEVATED TEMPERATURE SERVICE PURPOSE: 1. ACCEPTANCE CRITERIA FOR MATERIALS IN WELDMENTS 11. CREEP RUPTURE DAMAGE FROM RESIDUAL STRESSES Iy. EVALUATION OF METALLURGICAL NOTCH EFFECTS 9 CONCLUSIONS: I. CURRENT CODE RULES AND ACCEPTANCE TESTS ARE ADEQUATE II. THE RESIDUAL STRESSES RELAX IN SERVICE BUT THE RESULTING PLASTIC STRAINS ARE LESS THAN 10% OF THE CREEP-RUPTURE DUCTILITIES OF THE ZONES OF THE WELDMENT IV THE STRAIN CONCENTRATION OF A METALLURGICAL NOTCH IS ACCOMMODATED BY THE HIGH DUCTILITIES OF THE ZONES OF A WELDMENT 32

a 3 ge e. (.se ee. e-ve:...e e*. j e 1 ex, tope g F g,,,,,,w} .'s., 1 ...e we u,.. S ee 449 5 ex A 1 f5 e 1 A 2er e g 1 3 s l (.00 I V. #%. e SX

• i e

4 24 O t 0 e h e $. $ 3 g *te f e3* *e f (ta't"a't. " s MC - te ef .o. T t = {u .e'. s ;. I 3 1

f. JC e

1 4 - o 6 } 'I

.a g

a lm 10C ".* i S C 4 s.

• * *3.*.te..f ec.e.m.e f.C o a.*t.e ' *e, * *

.2.....,.........,.. .........7......... ..q. ... q... ;. ..,.,........;.y....,..,....y.. A s >,1 /g-< i s .c yy l H E5 s .J -.... s ...v ~ <9 s -...,.i i ~ - ~. c

lt M i FIGURE 3.0-14 srvr.*. Ass r ~....r.

W. G g '.i ~7A 33

FIGURE 3.0-15 .r.yvr r e ry. 7 ~ ,'.t'..-'" ,. c.-' W 3 3 T a M y. " "*' .e;.D';# 4%i,5 cWk &'g i _:. :. i i :' p 1 - - ..e:n M. s u.... r r-;/ u 6.'{:.t'Tt',

• w: :..tt

.t feT 4 h 2. W ~ y v.~.' 1, gg 5 ':..

p..n. v.J r.ly *la:e. '. _.. ; o...

.N t.5 r:.t e_3 : - cg q ' ; 7 :. N.. A .L...p 5 R2 h E:7. i A. N-:~. ki!-:'E Q %., : ! fai.. '..; : '. '..:_t.2 d ' k b M }_;* 2 r5 e. O 5 VM 2 y.:.a&n~..p T.@ * ' 4. '.! :..l.$.T rbEM.oi ,S g i: M5.: y.-::rtW. - c.': -:-a-J.:..' c.: 6%.sim;:'-%;.f; r t.m> ! W $(i :::.d', *$ '.l. r s.,: w.iN. :

Hi-1.i1;.3.w.:;.Bi& net.-T: is:

i.

9'..

.,9::.p.cE6.:cyn(-h2 gW-#EraC e.dg*

WFG fll '. J. br*'*; .. :. - Q 0 9y:?. 4: @.f": '-' 4,j;2s,,..S .,,. :N-bU.14: e & -?: M ii.

. v. :z. : n. 1 R w. t. f. p W. 9.;

r.!MMay....x:&w.,. va.?yhoss>6 ;.g... M:mm< .bf H 'L.' :::b.-gp.w m n;&y-Ife F. :we.::r sn wi:::%g_ y. -- ye. p,dL.qu:.4 t!Ps?-y4mh s..-x .1.i-t. A7%,De3m;.a6.ad:t asg rts.t

p -;.

K d @ @e.~.-y.".> :' U.d:T M j).; %:.L.e :a ...' %J.; :;m.3 4 e c,w:. n:.sm>%gy.::w:.r. ~h; @9.t r.,@ ..nnme[, :vya Mw.wu ,r!.i ir.!.W.: c.st29wME:$n-nem M.:*- c r-e oQ.21 gb % s. w w.u:erH M e.,n; rm.:,s sc.3.. .,A... w4 t

t. m t::y x

.m<s-u&s%.-.e ; M,..< ~...., ;?<. &.

_v. e.c. ~....v. n C. $rt

. m..o,_ r.;r m n.

y ...m pt M. M.:.0 & N:.; ; A... 1 .T f g ' g ? n :~ g, o r a e m p e n i m e s w w :e n ' P. '.; v - M m a n.wic;esors ,wi., W e r $ Q g4.cryA ?.:7I-D %;;.. C';:.l*..,1-T'i : G."> w.v'MtRtW.t$4c% 3 d.WT'!2:P&Y:: ;' ' M-2Mtt:1tW src:::w.e.n.:;. : ~.g #.j;1= :.D' Y)c~.tii.a.f;M-uw. tw c.cn;:-

,:n., yp :4' p..rw.i -,.. r g e w u n.e s s fn; v

m-w.tr,r w.:d;Sr. .Y ;...-y. :.i.,..f.::.- , J. 'T.,1 e i '.T yily!p@smn:r.wwW

T:'T:.,2 E,' :AC:? -ci 57.iESA*.Sht' M

<!_3 : 1,.t ?;-- - ne:; *i ;.: !;.;-d:p tK diW W W M W ^ :I r :. q

- i c it..i Mc ;gte.sc55rvI 7

. l.m;..

1. E.s.;. O. e.. t :: 'ime ! -

.1 i : ; % 5.>. m d ii4 F a ysh ?s h.r,:W..n.!.,c",'..:.-( F.&T G ^t ? l ~'. ' 'ri.M .. t,. .e ;.e 4 + '" N YJ l '-- 9. uM M:tl Li !. '.-'J."&, .L..'._... f.V r:w;,g.p.: . l 'T-p . v...,...,. Q %g. a Y 5..*- .ach.. W ).. 4 'f y.," ;-:. N I)Y g-- -,8 ' '!.H ' :..- -:.Ai r ': ;** :..... 3 1

• vi 1 H. i-t1

..., -{s...k. MkH &.r ,ri,t.*M.,e.n.ii:;.,.,:3?V.ir m..T. t ' I ~ c ... f,.. _., ....r. w r >..:."J.. r....., g ,L

- : { 2. ','..

t p. :.. &., i;.,. ;;.: 1

l. t.

7: r ;.:j g !-]\\ : j. :. -t w i s.... i.. '-s m ., p .~,t s t? t: c:4.. v u.' "..:e n,- ?...r.=:n. n 'n z- "3. w .w vi r ..:.ov=

i :.i g. -

Fig.15-Photograph of crack pattern near mid-length of Weldment BAW-2 after 4500 hours (P% of life) at 151.7 MPa and 593 C 34

l ,.. m,. .--. P.,3 M.v,V..,'..h..T* E..P V,.-T. urf F n rr - S. . c ,-i... - p 7,.. y .-t.-+ ? ; ;t. ~. _t.:r > + gy..f j. v x.p. g.. w, .,$ : 3, e' ; q s

.4'..;.3. -... s e

.y. '1 1,.* ~~*. &p... u . _....h . k,h '.',V$s:c;f l.~. .' Y .M ' ' ' ' .. - %' a,, . ~... u . 1.., -- u ~..s. 7. -r ~ r'. 4,',.., ..,.h.?p sn.,, ',.E a u, M,h f ',.M.'$.N' 7- .I I ','" "[.ik.17"'.T$ I k [t[ C.'".,,}b % " ). : ',. 3. - - b.',;'7 M M-6 _. il.f f.i.C ? M~~l-N, $E<f.*?i:. ,..- ~., >.E'. ~

a..>.;-

<.. e.' w -^ d. e *-~ ~ ~ - -~ r J, 5 - +.s L u g-- ),.. s. w:..;

i. --

.~ r -r....,.... L.Q s_* h Y'- L s. 1- >~ 4_m, w'**t er, ' [.'_" ,^ ..% _= ".... L: ~ 9 1 ..... i 1 a e a gg my.+.p [*' '. ' i - 'ir. r 7- {jg,m. . h' n :- ,. 4 Mfn--.-5.,@ WG T* X s,, ^,..s, M 5. ;f U f, gd, , i, j-ph, t g,tgd -i .E,,.DT~,T)7-1:? I t-., 1. r -...2 y, # ~ ,e , 4 j' .Ee 4 d. U_ i h m. 5 4 It F m ~ -- f W l' (A) 7319 HOURS LOADING DIRECTION

p.... 4

.., ).u... _, ....,., g,] h..,. p.. s c e.r -. m ..,y. 3[ ...,. :. l..t 7 , 3.. .. 7.-- --. ' l.y *.... ..e ..... m.:.....,. 4 - -r-,,j7o, L, t p g;- i g -t 1~r.. .t a e it. .....s..

-=

4 ,'h y ...... - p' g.,e q.-.,-. ;,c.,.... _ pbg -}- L ' h,L '.;. r > s.7.;,"m<<y ,a..,..d. '.' d!.m,4 %....l0 1,,2 $,.,l Q.. I -~ e.g. I

f. >

... ; 4.. r M.".dg"f. ,\\ --f-t-t Q qh. ,1 --H ' _a..-a -;4..qQ.$. 1.,c *.,,.. +j.' Y T-

g.7:.5 Qv-=g~; %.T'

' '.t ~t,.J.~.. p M gp,t k,. .g.... W. N..y ) i l l'. .. W o - y.. ( .,.,in m y:.tr & y., y-. } =

- &.J.O

. 8 kf 7 - G: 3 , 4._! M,1 -i-

c., %N a..A i

1 i m.1 1 6 A~e -. s ,e 1 . u..(i L (,,g .9,;th'.;.,.y.'b..l,.Q.'.nyj1 _ i -[ [, ' ' ' ' ;., ' "h ' ~.

• '=,-

. ~.T ' ~~~' - $'bj-fl-,O A,, 4 I,, e ,t ]J.l' Y... g], w 73 p .a 3 t 7 g, 4.g. f 1 J h' I %,Q. q ba; 4,-1 .s .2 h (B) 13,103 H FIGURE 3.0-16 Cracks at HAZ's of CAW 1 at 7319 and 13.103 Hours 3351 4 35

i ~ suoraet wt.o--+. LOAD l t .,+- 20C7 witc.' t 's, 4g , " * * ~ ~ ~ ~ ~ ~~ I s O- \\ I e' m-0 -- 10 10 - og s oe o2 e 20 -- 'O l. ( -- 20 45 \\[ to 'g 8 3 l',,b')) (- ' o'o / s -- 40 w f 40 -- E \\ i, '2' 'b ,/ ! S0 - , o [- - SO 2 30 8 SO --so 1,g + go .o ,3 ) l40 j E 8 6 / / oo' g , oo.,---r + g '\\ - 60 g 60 -- to I 70 S S -- 70 -- ,g, t i; s 's e, l 'i 0* l 2o

1. 50 1

80 -- io l i! I ij so 04 Ls' l N s ' 'o.o l y 90 90 -- or [ 60 - 'N 1 4 g 100 - 's 1 I il I --100 o I l \\ [l] wtLo #007 Fact l l s I 's, i ' ca l I I '__________4_L_______i s l I i ALL simainstst amt Positivt ExCEPT wMtat osottD I I I I I I l O 10 20 30 40 SO 60 70 80 90 t00 PERCENT OF w!OTH FIGURE 3.0-17. AxlaL STRAIN CONTOURS FOR CAW-1 AT 14,571 HOURS 4 4 I 36

10 20 30 40 SO M 70 80 90 100 I t.0 A O ', D\\'o f< 0 0- \\ ~~ IO 10 '- \\ 20 ( 20 - ao so y e: - -- [ hb - -- dr - [N a s . 7 ________________;_ 2 q- - ',M M M N Y = 'o o ,'e 2o SUR ACE [# ~ 00 o 60 ~ ~ ~~~~ ~~~ a-to ,_.H o ,/ o to ) _ f ' '_ _ - _ - = 10 10;_______ /_) - 'o e' ~~ ~ ~ 4o po - to 40 33 l' - 90 o 2 is 99 i.S ~- - 10 0 10 0 -- y, /'D oS V os h s f ALL ST4&itel (%) ARE PO$lTIVE 1 I I I I I I i I O 10 20 30 40 SO SO 70 to 90 10 0 PERCENT OF Wl0TH FIGURE 3.0-18. axial STRAIN CONTOURS FOR WELO ROOT FACE OF CTW 37

/ 'Isw ssatfis ~ = 2 R = ~ o 3 ~ I i 4 1 i i e 1 E. / = / gr i / ,~, E 3, i I ,--e m E 8 ,/ ~ 2 S' / U 5 / ~ e op 9 I .f I J- / ~ z 2 JD q = = g ~ = / I J / me e = = i o / ,o -= 8 a m_ = E 5 8= o G 3 ~ =. = u j _. _m m e. g- ~ ~ a = m u E. -E E -E. x -\\ . ?*. s

• =

a I e o = = c= -E 2-a- ag Es a; 3 I.

? - g

=- c = -= = =_ c U S - -.=- -E- = ~ ( 1 i i 1 i i ^ o e o o e o,, e_ N N N ~ N o e = o.9 h_ k_ 0 C O O o E 2 e m 317]s 001 edv. s!!ns FIGURE 3.0-19 38

COMPARISON OF RUPTURE STRENGTH OF WELDl!ENTS WITH ASt1C CODE DATA AT 593*C b - Stress for 10" Hours Stress for 10 Hours Net Section Gross Section Net Section Gross Section Material,, MPa ksi MPa tsi MPa ksi MPa ksi ASME Code Data Curve 100 14.5 69 10.0 83 12.0 Raieretal 103 15.0 An T30455. HT 9T 2796 112 16.2 Asfal Weidment 142 20.6 TJ00 C"E Weld 123 17.8 Transverse Weidment 139 20.0 T308 CRE Weld CAW - Hole Within Axial 129 18.7 116 16.8 95 13.8 85 12.4 u Weld CTW - Hole Within 138 20.1 125 18.1 1 31 19.0 118 17.I Transverse Weld 147 21.4 DTW - tiotch Within 200 29.0 Transverse Weld 85 12.3 DZTW - Notch at HAZ 110 16.0 for base met.it a) the strengths of ti.e weldments witicut stress risers are greater than that of code data that of b), the strength of the axial and transverse weldments with a hole within the weld is great er than the code value for the base metal Ghen based on the gross stress away f rom the hole c) the notch strength of the weldmente are greater than that of the code value for base metal d)the code design strens allowables are only 2/3 that of the code data value FIGURE 3.0-20

l FIGURE 3.0-21 HTS MATERIALS AND STRUCTURES CS250,4 STRUCTURAL EVALUATION OF WELDMENTS AND ASSOCIATED MATERIALS FOR ELEVATED TEMPERATURE SERVICE PURPOSE: 1. ACCEPTANCE CRITERIA FOR MATERIALS IN WELDMENTS 11. CREEP RUPTURE DAMAGE FROM RESIDUAL STRESSES EVALUATION OF METALLURGICAL NOTCH EFFECTS iv. CONCLUSIONS: 1. CURRENT CODE RULES AND ACCEPTANCE TESTS ARE ADEQUATE THE RESIDUAL STRESSES RELAX IN SERVICE BUT THE 11. RESULTING PLASTIC STRAllis ARE LESS THAN 10% OF THE CREEP-RUPTURE DUCTlLITIES OF THE ZONES OF THE WELDMENT iv. THE STRAIN CONCENTRATION OF A METALLURGICAL NOTCH IS ACCOMMODATED BY THE HIGH DUCTILITIES OF THE ZONES OF A WELDMENT 40 _}}