THE UNIVERSITY OF MICHIGAN COLLEGE OF ENGINEERING of Chemical and Metallurgical Department Engineering Progress Report PREDICTION OF RUPTURE STRENGTHS OF GRADE 11, GRADE 22 AND TYPE 304 STEELS BY SIMULATION OF LONG TIME SERVICE David J. Wilson ORA Project 340020 supported by: THE METAL PROPERTIES COUNCIL NEW YORK, NEW YORK administered through: OFFICE OF RESEARCH ADMINISTRATION ANN ARBOR June 1971

TABLE OF CONTENTS Page LIST OF TABLES v LIST OF FIGURES vi INTRODUCTION 1 EXPERIMENTAL BASIS 2 Utilization of Simulated Service Exposures 2 Extrapolation by Parameter Methods 2 EXPERIMENTAL MATERIALS 3 Grade 11 Steel Grade 22 Steels Type 304 Austenitic Steel 4 PROCEDURE 5 RESULTS AND DISCUSSION 7 Pll (1.25Cr-O.5Mo-0.75Si) Pipe 7 As-Removed From Service 7 Renormalized + Tempered 2 Hours at 1200~F 7 Renormalized + Exposed to Simulate Service at 1000~F 8 Parametric Extrapolations 8 Microstructure of Pll Samples 9 Discussion of Results for Pll Steel 9 Grade 22 (2.25Cr-lMo) Steels 11 Pipe Material 11 Rupture Properties at 1000~F 12 Rupture Strength Above 1000~F for Original Material 12 Simulation of 83,000 Hours of Service at 1100~F 15 Grade 22 Tubes 13 Rupture Strengths for Original Materials 14 Simulation of 10,000 Hours at 1100~F 14 Simulation of 100,000 Hours at 1100~F 14 Microstructures of Grade 22 Tubes 14 Discussion of Results for P22 Pipe and Grade 22 Tubes 15 Type 304 (18Cr-8Ni) Stainless Steel 16 Comparison of Rupture Data with AR-2 Tests 16 Parameter Extrapolation 16 iii

TABLE OF CONTENTS (Concluded) Page Simulation of 100,000 Hours of Service at 1200~F 17 Influence of Stressed Exposure 17 Micro structural Studies 17 Discussion of Results for Type 304 Steel 18 General Discussion 19 SUMMARY AND CONCLUSIONS 21 ACKNOWLEDGMENTS 22 REFERENCES 23 iv

LIST OF TABLES Table Page I. Rupture Test Data for Pll (1.25Cr-0.5Mo-0.75Si) Material AsRemoved from Service After 83,000 Hours at 1000~F 24 II. Rupture Test Data for Simulated 83,000 Hours of Service at 1000~F on Pll Steel Pipe 25 III. Long-Time Strengths at 1000~F for Pll Pipe 28 IV. Rupture Test Data for the Original P22 (2.25Cr-lMo) Steel Pipe 29 V. Simulated 83,000 Hours of Service at 10000F on P22 Pipe 30 VI. Simulated 83,000 Hours of Service at 1100~F on P22 Pipe 31 VII. Long-Time Strength for Grade 22 Pipe 32 VIII. Rupture Test Data for the Original Grade 22 Steel Tubes 33 IX. Tensile and Rupture Data at 1100~F for Grade 22 Tubes Exposed to Simulate 10,000 Hours at 1100~F 35 X. Tensile and Rupture Data at 1100~F for Grade 22 Tubes Exposed to Simulate 100,000 Hours at 11000F 36 XI. Long-Time Strength at 1100~F for Grade 22 Tubes 37 XII. Rupture Test Data for Type 304 Stainless Steel in the Original Condition and After Simulated Service Exposure 38 XIII. Long-Time Strength for Type 304 Austentic Steel 39 v

LIST OF FIGURES Figure 1. Stress versus rupture time data for P11 pipe as-removed from service and after reheat treatment. 2. Stress versus Larson-Miller parameter curves for Pll pipe asremoved from service and after reheat treatment. 3. Stress-rupture time curves at 1000~F for Pll pipe as-removed from service and after reheat treatment and exposure to simulate service at 1000~F. 4. Microstructures of Pll pipe. 5. Stress-rupture time curves at 10000F for P22 pipe in the original heat treated condition and after exposure to simulate service at 1000~F. 6. Stress-rupture time curves for P22 pipe in the original heat treated condition and after exposure to simulate service at 1100~F. 7. Stress versus Larson-Miller parameter curves for P22 pipe in the original heat treated condition and after exposure to simulate service at 11000F. 8. Microstructures of P22 pipe. 9. Stress-rupture time curves for Grade 22 tubes in the original heat treated conditions and after exposure to simulate 10,000 and 100,000 hours at 1100~F. 10. Microstructures of Grade 22, tube A. 11. Microstructures of Grade 22, tube B. 12. Microstructures of Grade 22, tube C. 13. Microstructures of Grade 22, tube D. 14. Stress-rupture time curves for Type 304 austenitic steel in the original heat treated condition and after exposure to simulate service at 12000F. vi

LIST OF FIGURES (Concluded) Figure 15. Stress versus Larson-Miller parameter curves for Type 304 material in the original heat treated condition. 16. Microstructures of Type 304 austenitic stainless steel. 17. Transmission electron micrographs of Type 304 steel, cold reduced 40% and exposed 100 hours at 13500F. vii

INTRODUCTION The long time creep-rupture strengths of alloys, particularly the 100,000-hour rupture strengths, are major factors in the establishment of design stresses for high temperature applications. At the present time, it is virtually impossible to obtain acceptance of 100,000-hour rupture strengths without tests out to longer than 10,000 hours. Even under these circumstances, there is often considerable uncertainty to the strengths at prolonged times. In addition to economic advantages, the following benefits would result from the development of a "practical verification" test (short time) that would define rupture strengths at long times. (1) Avoidance of steel being placed in service with unexpectedly poor strength. (2) Development of the controls necessary to produce material with strengths on the high side of the range. Currently, the data cover a wide range of strengths with the code stresses governed by the minimum values. It is contended that the wide range in properties is tolerated only because it is not practical to determine creeprupture strengths by routine acceptance tests. (3) Delineation of the effects of manufacturing and fabrication conditions on creep-rupture properties. An evaluation of possible verification tests has been carried out at The University of Michigan. The research originated from a program sponsored by the Detroit Edison Company. Clarification and extension of the initial results were carried out under the auspices of the Metal Properties Council. Results from the entire investigation are presented in this report. 1

EXPERIMENTAL BASIS Preferably, a verification test should indicate the actual creep-rupture strengths for whatever service time period and temperature may be of interest. The two techniques involved in the study, simulated service exposures and parameters, were considered potentially useful as the basis of such a test. The research was focused on the utilization of short time tests to predict strengths at times from 10,000 to 100,000 hours. Utilization of Simulated Service Exposures Considerable data has been accumulated at The University of Michigan which indicate that thermally induced structural changes that occur during prolonged service results in the elimination of the instabilities in log stress-log rupture time curves (ref. 1). Thus, the rupture curve for material removed from service can be extrapolated with relative ease and reliability. The available information indicates that the extrapolated strength defines the strength of the unused material. It was suggested that the long times of service could be simulated by short time exposures at temperatures higher than the service temperature. The result would be a method of determining the long-time strengths of unused material before service with confidence by relatively few tests. The investigation involved the study of specific treatments to simulate long time service and the evaluation of methods of extrapolating short time tests of these heat treated materials. Extrapolation by Parameter Methods Parameter techniques are widely used to extrapolate stress-rupture time data (ref. 2). There has been considerable question as to their accuracy. 2

However, when correctly applied they can lead to reliable extrapolations. Since they are often used for verification testing, some consideration was given to their merits relative to methods based on simulated service exposures. EXPERIMENTAL MATERIALS The materials selected for study were Grade 11 (1.25Cr-0.5Mo-0.75Si), Grade 22 (2.25Cr-lMo) and Type 304 (18Cr-8Ni) steels. Grade 11 Steel The material (Pll pipe removed from service) was suited to the investigation because the long time strength at 1000~F had been determined with a great deal of confidence (ref. 1). The pipe had operated for 83,403 hours at 1000~F under 1800 psi stream pressure. It was forged and bored by MidvaleHeppenstall to an O.D. of 12 inches and the I.D. of 9 inches. The heat treatment was a normalize plus temper.* -The simulated service exposures were carried out after reheat treatment (renormalized from 1675~F to restore properties typical of unused material) in an attempt to duplicate the rupture properties of the material as removed from service. The specimens, in the longitudinal and transverse directions were 0. 550 inch in diameter except for specimens tempered at 1325~F or annealed which were 0.505 inch in diameter. Grade 22 Steels The P22 pipe studied had not been in service. The wall thickness was 3-3/4 inches. The heat treatment was established to be either a normalize from 16750 - 1725~F plus tempered at 1225~ - 1350~F or an annealing from *As near as could be finally determined, the pipe prior to service had been heated at 1750~F for 1 hour, A.C. + tempered at 1225~F for 1.5 hours. 5

1750~F. The microstructure indicated the normalize plus temper to be the treatment. The specimens, both longitudinal and transverse, were 0.350 inch in diameter except for 0.505-inch diameter-specimens used to study simulated service at 1100~F. In the latter stages, the program was extended to include four unused Grade 22 tubes with O.D., 2-1/4 inches and I.D., 1-1/4 inches. One tube was treated 1 hour at 1400~F and air cooled. A modified isothermal treatment was used for three tubes. Specimens 0.350 inch in diameter were taken in the longitudinal direction. Type 304 Austentic Steel The type 304 material was bar stock from the ASTM "Standard Specimen Bank." Very long time creep-rupture tests (ref. 3) had been conducted on the material by the Applied Research Panel of the ASTM-ASME Joint Committee on the Effect of Temperature on Properties of Metals (Project AR-2). Although the stock used was from a different billet than the AR-2 material there was every reason to expect that there would be no significant difference in properties. Using this material reduced any problems associated with uncertainties in long time strengths since tests had been conducted beyond 50,000 hours at 1100~ and 1200~F. The steel was furnished as 5/16-inch round-corner squares water quenched after 1 hour at 1950~ - 1975~F. Specimens, 0.250 inch in diameter, were machined from lengthwise quartered stock.

PROCEDURE There was no established basis for selecting exposure times and temperatures to simulate service. Larson and Miller (ref. 4) proposed-the following equation to describe the interrelationship between rupture time and temperature: P = T(C+Log t) where T = Temperature, ~R t = Rupture time, hour P = Parameter C = Constant This relationship offered a basis for selecting simulated exposures. Initially the program involved simulation of 83,000 hours of service at 1000~F for Pll and P22 pipes. There was no reason to expect that constants determined to correlate rupture data would also be applicable for selecting the thermal exposures. Arbitrarily, therefore, exposures of 300, 88 and 20 hours at 1200~F were selected. These correspond, respectively, to constants of approximately 15, 20 and 25. For the P22 pipe these exposures had little influence on the rupture properties. Consequently an exposure of 115 hours at 1300~F (C=20) was carried out to simulate 83,000 hours of service at 1100~F. As there was an uncertainty as to the effect of stress during exposure, samples were exposed with and without stress. The ASME Boiler and Pressure Vessel Committee Code stress of 7800 psi for service at 1000~F (adopted 1951) was used for the stressed exposures. These were carried out in creep-testing units using oversize specimens which were remachined after exposure. Exposures without stress were carried out on machined specimens sealed in evacuated quartz tubes. Rupture tests were conducted at the service temperatures simulated (1000~ and 1100~F) on the materials in the heat treated conditions 5

("original" treatments) and after the simulated exposures. Tests were also carried out at higher temperatures to provide parameter extrapolations to 100,000 hours. Four Grade 22 steel tubes exhibiting a range of rupture strengths were incorporated into the program. Tests were carried out to 10,000 hours at 1100~F for the as-heat treated materials. The rupture curves at 1100~F were extended to 100,000 hours by "parameter" testing. The tubes were exposed to simulate 10,000 and 100,000 hours at 1100~F (based on C=20 these were, respectively, 100 hours at 1242~F and 100 hours at 1313~F). Specimens were machined from the materials after exposure and rupture tests carried out at 1100~F. Testing of the Type 304 steel was carried out to determine if the stock supplied had the same properties as the billet used for the AR-2 tests. Rupture tests were conducted at 1200~F and also at higher temperatures at which results of "parameter" tests had been reported (ref. 3). On the basis of the agreement obtained, bar stock was exposed to simulate 100,000 hours of service at 1200~F (C=20). The exposure of 100 hours at 1425~F was carried out prior to the machining specimens for testing. Specimens were also exposed under a stress of 8500 psi (the estimated 100,000-hour strength of the AR-2 material) at 1325~F until the secondary creep stage was attained. (This took about 100 hours and the total creep was about 0. %.) The exposed materials were rupture tested at temperatures from 1100~ to 1425~F. The simulated service exposures for Type 304 steel did not duplicate the structures evident after long time test exposure. Consequently a number of thermal mechanical treatments were studied which were designed to accelerate the microstructural reactions occurring. The structures obtained were studied using optical and transmission electron microscopy. 6

RESULTS AND DISCUSSION The rupture test data are presented in Tables I and II for P11 pipe, Tables IV through VI for P22 pipe, Tables VIII through X for Grade 22 tubes and Table XII for Type 304 stainless steel. Where longitudinal and transverse specimens were used, the results of the investigation appeared to indicate no marked difference in rupture strength. Accordingly, the results of the tests were combined. Pll (1.25Cr-0.5Mo-O.75Si) PIPE AS-REMOVED FROM SERVICE Consideration of the rupture data (Table 1, Figs. 1,2) for the material from service (ref. 1) led to the following: (1) The stress-rupture time curves at 1000~F and 1100~F were parallel while the 1200~F curve was slightly steeper (Fig. la). The similarity of the slopes of the rupture curves, by the family-of-curves concept, lent confidence in the 83,000-hour strength level at 1000~F of 13,200 psi indicated by straight line extrapolation. (2) Extrapolation of the rupture data at 1000~F, using the Larson-Miller parameter with a constant C of 20 indicated an 83,000-hour strength of 12,800 psi (Fig. 2). This value is slightly lower than that obtained by straight line extrapolation of the log stress-log rupture time curve. RENORMALIZED 1675~F + TEMPERED 2 HOURS AT 1200~F Renormalizing and tempering restored properties characteristic of new pipe. The tests at 1000~F were at much higher stress levels than for the material as-removed from service (Fig. lb). The data at 1000~F for the reheat treated material could not be extrapolated as a straight line since changes in slope were expected to occur at prolonged times. Consequently higher 7

temperature tests were conducted to permit parameter extrapolation. The Larson-Miller parameter curve (Fig. 2) exhibited " " type behavior i.e., an increase followed by a decrease in steepness. The stress for rupture in 83,000 hours at 1000~F was indicated to be 13,300 psi essentially the same value as derived by straight line extrapolation of the curve for the material as-removed from service (Table III). RENORMALIZED 1675~F + EXPOSED TO SIMULATE SERVICE AT 1000~F Exposure of 20, 88 and 500 hours at 1200~F were used to simulate 83,000 hours of service at 1000~F. There was evidence that exposure under stress (7800 psi) resulted in slightly lower strengths than for exposure without stress (Fig. 3). This could simply reflect utilization of small amounts of rupture life during the stressed exposures. In no case was there exact agreement between the stress-rupture time curve for exposed specimens and that established for material removed from service (Fig. 3). Although somewhat steeper, the rupture curves for the materials exposed 88 and 300 hours came quite close over the test stress range (Figs. 3b,c). When extrapolated as straight lines, the rupture curves for the materials exposed 20, 88 and 300 hours indicated 100,000-hour strengths that were respectively, higher, similar and lower than for the material asremoved from service (Table III). PARAMETER EXTRAPOLATIONS Parameter testing indicated that the materials as-removed from service and after re-heat treatment had essentially the same 83,000-hour strengths (Fig. 2). These materials had very different short time strengths. To determine the generality of the observation it was decided to check as many treatments as possible using a stress of 12,500 psi and a test time of about 100 hours. The results of the tests run at 1200~F were in close agreement 8

(Table II). This was true even for reannealed material. Only the material exposed 300 hours at 1200~F resulted in a somewhat lower strength. MICROSTRUCTURES OF Pll PIPE SAMPLES The as-removed from service microstructure (Fig. 4a) showed extensive spherioidization of carbides. There was also precipitation within the grains. Simulated service by exposure for 20 hours at 1200~F after normalizing from 1675~F did not appreciably spheroidize the carbides (Fig. 4b). Considerable spheroidization occurred during 88 and 300 hours at 1200~F (Figs. 4c,d), but was not as complete as in the material from service. Also there was less precipitation within the grains. A stress of 7800 psi had no readily apparent effect on the microstructures. The microstructures of specimens tested for rupture at 12000F and 12,500 psi were very similar for all heat treated materials. The structures were highly spheroidized with large amounts of intragrannular precipitate. Thus the structures reflected the similarity in rupture strengths observed. DISCUSSION OF RESULTS FOR Pll STEEL The material removed from service yielded stress-rupture time curves which did not exhibit the changes in slope so frequently found for new steel (and evident for the reheat treated material). There was no evidence that measurable creep had occurred during service. Calculations, using the actual service temperatures and stresses, indicated that the. amount of creep-rupture life used up was negligible. Therefore the difference in results for the unused (reheat treated) material and the steel after service should primarily reflect the result of thermally induced structural changes that occurred during service. The pipe was thermally weakened by service exposure. Structural changes presumably resulted in the " " type rupture curve observed for the reheat treated material. (Microstructural examination indicated that spheroidization of carbides and precipitation of intragrannular carbides occurred 9

during high temperature exposure.) The downward break reflects a weakening reaction while the upward break is due to completion (or a slowing down) of the thermally induced change. Based on this, the structural changes should have been essentially completed during the 83,403-hour service exposure. The linear rupture curve at 1000~F for the material as-removed from service would substantiate this conclusion. As indicated by parameter tests, at very long times the rupture curve must increase in steepness (Fig. 2). This behavior prevents higher strengths occurring for the as-removed from service condition than for the unused steel (as characterized for the reheat treated material). In accordance with the results, at long time periods at 1000~F the rupture curves for these conditions must merge. None of the simulated service exposures resulted in exact duplication of the microstructure or the stress-rupture characteristics of the material as-removed from service. However, exposure of 88 hours (based on C=20) came extremely close. Straight line extrapolation of the short time data predicted reasonably well the strength at 83,000 hours. This would suggest that for the material studied a constant C of approximately 20 can be used to simulate long times of service. It should also be noted that parameter extrapolation (using Larson-Miller with C=20) for the reheat treated steel gave essentially the same 83,000-hour strength as determined from the as-removed from service material. Based on a C of 20, the exposures of 20 hours and 300 hours at 1200~F would simulate respectively 16,600 and 360,000 hours at 1000~F. Straight line extrapolation of the data indicated strength levels that were fairly close to those determined by the parameter tests for reheat treated material (Table III). The steel was thermally weakened and therefore the rupture time for the unused (reheat treated) material must be closely related to (if not controlled by) the final microstructure resulting from the thermally induced structural changes that occur during testing. This results since the life is used up at the greatest rate at the end of the test when the steel is in the weakest

structural condition. Assuming no further structural changes, testing of material after simulated service exposure could be expected to yield a linear rupture curve passing through the actual rupture strength at the time simulated. These concepts are consistent with the observed results as follows: (1) Exposures of 88 and 300 hours at 1200~F (for C=20) simulated times at which thermally induced structural changes were expected to be complete. The rupture curves were essentially linear and extrapolations (to 83,000 and 360,000 hours, respectively) resulted in good strength predictions. (2) The structurally weakening reactions were not completed by the 20 -hour exposure at 1200~F. The rupture curve increased in steepness (due to continued thermal weakening) at long time periods as indicated by the parameter test. However, straight line extrapolation of the shorter time tests did predict reasonably well the prolonged time strength. A striking feature was the similarity of the rupture strengths at the higher temperatures and times (i.e., high parameter values) for all of the heat treatments evaluated. Correspondingly, the microstructures introduced by heat treatment all tended towards an "equilibrium structure" during the high temperature test exposures. These results would suggest that although it is impossible to heat treat the steel to a range of strengths at low parameter values the influence of heat treatment decreases with increasing time and/or temperature. Grade 22 (2.25Cr-lMo) Steels PIPE MATERIAL The P22 pipe had not been in service. Consequently, unlike the Pll material, no data were available after prolonged service which would be used as a basis for accurately predicting long time strengths. Therefore, a parameter technique was used to determine the long-time strengths of the original material for comparison with the results of the experiments to simulated service. 11

Rupture Properties at 1000~F The original program was designed to use simulation of 83,000 hours of service at 1000~F to checkthe long-time strength at 1000~F. The exposures were the same as those used for the Pll tests. The test results (Table V, Fig. 5) showed that the exposures of 20, 88 and 500 hours at 1200~F, with or without a stress of 7800 psi had a tendency to reduce the short time strengths. This lowered the steeper portion of the curve evident for the original material at short times. It should be noted that a decrease in steepness is not an uncommon characteristic of stress-rupture time curves (at 1000~F or 1050~F) for 2.25Cr-lMo steel. Straight line extrapolation of the rupture curves for the exposed materials indicated 83,000-hour strengths similar to that obtained for the unexposed material using the Larson-Miller parameter with C of 20 (Table VII, Figs. 6,7). The test results and the microstructures (Fig. 8) indicated that the P22 was so stable that the exposures were not significantly affecting the material. This was not unreasonable when checking disclosed that the as-produced pipe had probably been tempered at 1325~F for about 4 hours rather than having been annealed from 1675~F and tempered 2 hours at 1200~F. Tempering at 1325~F was probably so effective in introducing structural changes that the additional heating for short times at 1200~F could not be expected to introduce major modifications in structure. For these reasons exposures were also carried out to simulate service at 1100~F. This temperature was expected to be high enough to induce structural changes during testing that could influence the rupture strengths. Rupture Strengths Above 1000~F for Original Material It was apparent from the test data (Table IV, Fig. 6) that the stress rupture curves above 1100~F exhibited drastic increases in slope within the time periods of the tests. This indicated that the rupture curves at lower temperatures particularly at 1100~ and 1050~F should not be extrapolated as straight lines. They were therefore extrapolated using the Larson-Miller parameter with a C of 20 (Figs. 6,7). 12

It should be noted that the difference in strength levels of stress-rupture time curves at 1050~, 1100~ and 1150~F was relatively small until the slopes increased. It is not known why this occurred. Published stress-rupture time curves for 2.25Cr-lMo steel do not normally show such small differences. Simulation of 83,000 Hours of Service at 1100~F The simulated service exposure of 115 hours at 13000F resulted in lower rupture strengths at the shorter time periods than for the unexposed material (Table VI, Figs. 6,7). When extrapolated as a straight line the data at 1100~F indicated the exact same strength at 83,000 hours as determined for the unexposed material using parameter extrapolation. Based on the Larson-Miller parameter (C=20) an exposure of 115 hours at 1300~F (P=38.9) should also simulate 517,000 hours of service at 1050~F. Straight line extrapolation of the data at 1050~F for the exposed material indicated an 517,000 strength of 7,600 psi. Again, this is the same strength as determined for the unexposed material by parameter extrapolation. For the exposed material, parameter tests indicated lower long time strengths at 1100~F than determined by straight line extrapolation (Fig. 7, Table VII). At the highest parameter values the strengths were similar to those established for the unexposed material. These characteristics are similar to those previously discussed for Pll pipe. GRADE 22 TUBES Four unused tubes with a range of rupture strengths were studied. The objective was to compare strength predictions at 11000F using exposures to simulate service against the 10,000-hour strengths from actual test data and 100,000-hour strengths determined by parameter extrapolation. 13

Rupture Strengths for Original Materials For the as-heat treated materials tests were conducted at 1100~F to times longer than 6000 hours (Table VIII, Fig. 9). This permitted accurate determination of the 10,000-hour strengths (Table XI). The strengths at 100,000 hours were obtained by application of the Larson-Miller parameter (C=20) to 1230~ and 1300~F test data. The materials exhibited a wide range of strengths at short times at 1100~F. The rupture curves for the higher strength materials exhibited marked increases in steepness at about 500 hours. The net result was that the strengths became more similar with time. By 10,000 or 100,000 hours at 1100~F the strengths were almost independent of heat treatment (Table XI). Simulation of 10,000 Hours at 1100~F The materials exposed 100 hours at 1242~F to simulate 10,000 hours at 1100~F, had lower short time rupture strengths at 1100~F than the unexposed materials (Table IX, Fig. 9). The decrease in strength due to exposure was greater the higher the short time strength of the original material. When the curves for the exposed materials were extrapolated as straight lines, in all cases the 10,000-hour strengths indicated were essentially the same as determined by actual tests on the unexposed material (Table XI). Simulation of 100,000 Hours at 1100~F An exposure of 100 hours at 1313~F was used to simulate 100,000 hours at 1100~F. Straight line extrapolation of the rupture data (Table X) indicated 100,000-hour strengths similar to those determined for the unexposed material by parameter extrapolation (Fig. 9, Table XI). Microstructures of Grade 22 Tubes Thermal exposure resulted in carbide spheroidization and the precipitation of intragrannular carbide (Figs. 10 through 13). These reactions were ac 14

companied by a reduction in rupture strength. The results showed the following: (1) Thermal exposure had the greatest influence on the materials (tubes A and B) which exhibited the highest short time strengths (Figs. 10,11). The marked microstructural changes were accompanied by considerable losses in short time rupture strength (Figs. 9a,b). Also, the rupture curves at 1000~F for these thermally unstable materials exhibited increases in steepness. (2) For the lower strength materials (tubes C and D) little microstructural change occurred due to the simulated service exposures (Figs. 12,13). Correspondingly the rupture curves did not exhibit marked instabilities and the exposure had little influence on rupture strength (Figs. 9c,d). (3) With increasing exposure (simulated service or test) the microstructures and rupture strengths became similar (Figs. 10 through 13). As discussed for the P1l material, the results indicate that heat treatment variation can be used to produce widely different strengths at low but not at high parameter values. The extent or generality of this observation requires further checking. Discussion of Results for P22 Pipe and Grade 22 Tubes Because of the similarity of the results to those for the P11 material the following comments are considerably abbreviated. Simulated service exposures were used to correctly predict the 10,000-hour strengths of four Grade 22 tubes as determined by actual long time tests. This would tend to suggest that the 100,000-hour strengths determined by simulated service exposures were also close to the actual values. These agreed with those established by parameter tests for the unexposed materials, and therefore, increases confidence in the use of this extrapolation technique. The tensile strengths after exposures to simulate 10,000 and 100,000 hours were similar (Tables IX, X). This could indicate that although the exposures correctly simulated the factors controlling strength at long times other factors influenced the short time strengths. The usual extrapolation practice for 2.25Cr-lMo steel up to 1100~F has been to use straight line log-log curves. Application of the family of curves concept, simulated service exposures and parameter methods, indicate that the 15

rupture curves at 11000F and probably somewhat lower temperatures can exhibit an increase in steepness. Such a slope change could results in lower 100,000 -hour rupture strengths at 1100~ and 10500- than determined by straight line extrapolation. Type 304 (18Cr-8Ni) Stainless Steel Simulated exposures and parametric characteristics were used to establish long-time strengths which could be compared with those established by actual tests in the AR-2 program. In addition, thermal mechanical treatments were evaluated which were designed to accelerate the microstructural changes occurring during prolonged time test exposures. COMPARISON OF RUPTURE DATA WITH AR-2 TESTS Rupture strengths at 1200~F for the stock supplied (Table XII, Fig 14a) were similar to those reported for the AR-2 material (ref. 3). Higher temperature "parameter" tests also demonstrated that the strengths of the two materials were the same. PARAMETER EXTRAPOLATIONS The steepness of the rupture curves at 1100~ and 1200~F for the AR-2 material increase markedly at about 1000 hours (Fig. 14a). Similar changes in slope are not evident at shorter times at higher temperatures. Thus the data are apparently not consistent with the parameter concept of trade off of time and temperature. This was reflected in the 100,000-hour strengths determined by Larson-Miller extrapolation of the AR-2 "short time parameter tests". Using an optimized constant C of 17.9 or the standard value of 20 considerably higher strengths were obtained than determined by the long time tests (Table XIII). This is clearly evident from Figure 15 where the 1100~ and 1200~F test data are presented along with the parameter data. 16

Extrapolation of the short time data using the Manson-Haferd parameter reportedly resulted in 100,000-hour strengths close to those determined by testing (ref. 3). There is considerable question however, as to whether this is a valid prediction or simply an accident of the downward curvature enforced by the mathematics of the parameter. SIMULATION OF 100,000 HOURS OF SERVICE AT 1200~F The simulated service exposure (100 hours at 1425~F) resulted in a slight increase in the short time rupture strength at 1200~F. Extrapolation of the data for the exposed material as a straight line indicated an 100,000-hour strength of 10,200 psi. This strength agrees with that established by LarsonMiller extrapolation for the unexposed material but is higher than indicated by the prolonged time tests (Table XIII). INFLUENCE OF STRESSED EXPOSURE Material was exposed under stress in an attempt to introduce a dislocation substructure representative of that which would occur in a test at 1200~F at the approximate 100,000-hour strength. The test exposure conditions (100 hours at 1325~F under 8500 psi) were selected so that second stage creep would be attained and the rupture life utilized would be minimal. The rupture characteristics after stressed exposure were similar to those established for the material exposed without stress for 100 hours at 14250F (Table XII, Fig. 14b). The results also showed that the ductility (Table.XII) after exposure, with or without stress, is appreciably higher than for the unexposed material. The significance of this is not known. MICROSTRUCTURAL STUDIES The results of a study of the microstructures of AR-2 test specimens were reported recently (ref. 5). Sigma and carbide particles were shown to have precipitated during the long time test exposures. It is possible that 17

the increase in steepness of the stress-rupture time curves is associated with the occurrence of one or both of these reactions. Equivalent precipitation did not occur during the parameter tests or simulated service exposures (Fig. 16). Consequently, thermal mechanical treatments were designed to accelerate the precipitation reactions. These could form the basis of a visual test of whether or not a break down in long time rupture strength might be expected. Such a test would not however, provide an estimate of the long time strengths. The treatments investigated were: (1) Cold reduction 400 plus 100 hours at 1350~F. (2) Cold reduction 40%0 plus 100 hours at 1450~F. (3) Cold reduction 40% plus 20 minutes at 1625~F plus 100 hours at 1350~F. (4) Cold reduction 40%o plus 20 minutes at 1625~F plus 100 hours at 14500F. All of the treatments promoted carbide and sigma precipitation (Fig. 16). The inclusion of the treatment at 1625~F tended to reduce the amount of precipitate formed. A study using transmission electron microscopy showed that the structures (Fig. 17) were very similar to those reported for the long time test specimens. Sigma particles were evident in the grain boundaries while M 2C6 carbide was present as a finely dispersed intragranular precipitate. DISCUSSION OF RESULTS FOR TYPE 304 STEEL The appearance of sigma phase is almost always accompanied by a loss in creep resistance. There is therefore, a possibility that it was responsible for the increase in steepness of the rupture curves. If so then this would explain why the use of parameters, simulated service exposures or any method that trades an increase in temperature for rupture time can indicate erroneous extrapolated strengths. Sigma phase has a "C" type TTT curve so that it does not form above about 1550~F. Presumably 1100~ and 1200~F are near the temper 18

ature of maximum sigma. Therefore, precipitation at these temperatures cannot be simulated by shorter time higher temperature exposures. If carbide formation was the controlling factor, then the response was different to that for the CrMo steels. For the latter materials, there was a parameter relation. The yield strengths at 1100~F and 1200~F do not appear to have been determined. It is, however, entirely possible that the increase in slope of the stress-rupture time curves is related to decreasing amounts of yielding when the stress was applied. Further research is necessary to clarify the above questions. It is considered important to develop the capability of correctly predicting the long time rupture characteristics of Type 304 steels. This could lead to the development of a practical short time test that reliably predicts 100,000 -hour strengths of materials for which "C" type reactions influence the properties. At the present time no such method is available. General Discussion For all of the steels studied there was close agreement between the strengths determined by parameter testing of unexposed materials with those established by the simulated service technique. This was not entirely unexpected. Both methods are highly dependent on the acceleration of thermally induced metallurgical processes by trading time for temperature. The basic relationship used in both cases was the Larson-Miller parameter. The oxidation effects differed. Extrapolations of rupture data for Grade 11 and Grade 22 steels using parameter methods have been questioned due to the extensive surface oxidation that occurs during the high temperature exposures. On the other hand, predictions based on simulated service exposures are not suspect on this basis. Specimens were machined after high temperature exposure and tested at the relatively low service temperature. The similarity of the strengths determined by the two techniques is therefore

evidence that parameter methods can correctly predict long time strengths even under circumstances where extensive oxidation occurs. It is probable that the accuracy of the strength predictions for the unexposed materials could have been improved by utilization of methods other than the Larson-Miller parameter with C of 20. This, however, was beyond the scope of the investigation. A verification test should be as simple as possible. For parameter extrapolation the number of tests required can be limited by using fixed constants. Hence methods such as the Larson-Miller parameter with C of 20 or the Manson Compromise method are most attractive. It is of interest to compare the minimum test points required for these methods and simulated service exposures to determine any long time strength, (1) The material could be exposed to simulate 100,000 hours of service, and subsequently rupture tested. A minimum of two short time rupture tests would be required to permit straight line extrapolation to the 100,000-hour strength. The testing could be reduced if it were possible by a single test to characterize the rupture behavior of the exposed material (i.e., the structure developed). Such a test was not defined by the present study. (2) Parameter testing to determine an 100,000 hour strength requires at least two "high temperature" tests. Thus the results do not show that the use of simulat ted service exposures at the present stage of development have an advantage over parameters as the basis of a verification test. 20

SUMMARY AND CONCLUSIONS A study was made of the use of simulated service exposures as the basis of a verification test. The technique was utilized to predict the long time strengths for Grade 11, Grade 22 and Type 304 steels. The strengths were also determined by parameter extrapolation. Consideration of the results led to the following: (1) Simulated service exposures reduced the uncertainties in the extrapolation of log stress-log rupture time curves for Grade 11 and Grade 22 steels. The rupture curves for the exposed materials did not exhibit the instabilities evident for unused steel. The evidence indicated that straight line extrapolation of short time rupture data for exposed materials gave good predictions of the long time strengths. The strengths determined by simulated service exposures were similar to those established by parameter testing of unexposed material. Assuming these were reasonably accurate determinations of the 100,000-hour strengths then values obtained by straight line extrapolation were in many cases in error by significant amounts. For both steels the results indicated that heat treatment variations resulted in a wide range of rupture strengths at low parameter values. At longer times and/or higher temperatures (high parameter values) differences in heat treatment had little influence on strength. (2) Accurate strength predictions at 1200~F were not obtained for the Type 304 (AR-2) steel by use of the simulated service technique (or parameter testing). This was probably due to the occurrence of a "C" type metallurgical reaction. It is recommended that additional research be conducted, directed at the development of methods for predicting long time strengths under these circumstances. (3) Simulated service exposures were not shown to offer a better basis for a verification test than parameter methods. 21

ACKNOWLEDGMENTS Acknowledgment is expressed to The Detroit Edison Company and The Metal Properties Council for sponsoring the research. Special thanks and appreciation are expressed to the late Professor James W. Freeman under whose guidance the research was conducted. The encouragement of I. M. Rohrig and J. J. Bodzin of The Detroit Edison Company, the M. P. C. Subcommittee I-Task Force on Acceptance Testing, C. E. Spaeder Chairman, has been greatly appreciated. 22

REFERENCES 1. Cullen, T. M., Rohrig, I. A., and Freeman, J. W., "Creep-Rupture Properties of 1.25Cr-0.5Mo Steel After Service at 1000~F," Journal of Basic Engineering, 88, 669 (1966). 2. "Time-Temperature Parameters for Creep-Rupture Analysis," ASM Publication No. D8-100 (1968). 3. Williams, W. Lee, "Parameter and Long Life Creep-Rupture Tests of Type 304 Steel," Time-Temperature Parameters for Creep-Rupture Analysis, ASM Publication No. D8-100, (1968). 4. Larson, F. R. and Miller, J., "A Time Temperature Relationship for Rupture and Creep Stress," Transactions of The American Society of Mechanical Engineers, 74, 765 (1952). 5. Biss, V., "Metallographic Examination of Type 304 Stainless Steel CreepRupture Specimens," RP-32-1969-7, Climax Molybdenum Company of Michigan, March 27, 1970 (unpublished). 23

TABLE I Rupture Test Data for P1 (1. 25Cr-0. 5Mo-0. 75Si) Material As Removed from Service after 83, 000 Hours at 1000~F Specimen Code I. l Test Temp. (~F)! Stress (psi) Rupture Time (hours). Elong. (%) R.A. (%) Longitudinal Orientation A- 1 A -4 A -2 A - 13 A -6 1000 1000 1000 1000 1000 A -5 A -3 A -7 A - 10 1100 1100 1100 1100 38, 300 24, 000 22, 000 22, 000 18,000 18, 000 16,000 15, 000 14, 000 12, 500 11,000 8,000 STTT 72 183 106 1566 44 194 425 1019 63 184 1505 41 51 44 37 27 77 68 78 76 59 53 48 44 31 74 76 67 48 A- 11 A - 12 A-16 1200 1200 1200 37 39 14 71 63 36 Transverse Orientation B-2 B- B-3 1000 1000 1000 24,000 22, 000 20, 000 39 179 434 51 43 46 73 67 56 STTT - Short Time Tensile Test 24

TABLE II.Jupture Test Data for. Simulated 8.3, 000 Hours of Service at 1000~F on P11 (1. 25Cr-0. 5Mo-0. 75Si) Steel Pipe Specimen Code* t,! _- -~~~~~~~~~~~~~~~~~~~~~~~~ Test Temp. (~F) Stress (psi) Rupture Time Elong. (hours) (%) R.A. (%) Base Material - Re-Normalized 1675~F + 2 Hours at 1200~F 1L13 1L14 1L16 1L15 1T7 1T8 1L34 1L29 1L 1000 1000 1000 1000 1000 1000 1050 1100 1100 1150 1150 1200 1200 1200 1200 54,300 40,000 39,000 34,000 37, 000 31, 000 25,000 25, 000 16, 000 20,000 17, 500 13, 000 11,000 15,800 8, 000 STTT 174 313 900 833 1057 252 41 658 24 33 72 269 21 1056 32 35 25 37 20 25 23 44 45 61 40 34 32 21 20 77 74 71 75 60 61 72 81 61 58 79 74 53 66 38 1L30 1T16 1L33 1L31 1 T15 1L Re - Normalized 16750F + 20 Hours at 12000F (C=25) lLl 1L2 1L 1L3 1L4 1T1 1T2 1000 1000 1000 1000 1000 1000 1000 49,200 40,000 35,000 28,000 26,500 34,000 24,000 STTT 8.5 30 438 567 27 694 36 46 47 33 49 37 51 81 77 80 81 80 74 75 * Code L - Longitudinal Specimen T - Transverse Specimen STTT - Short Time Tensil Test 25

TABLE II (Cont. ) Specimen Code 1L35 Test Temp. (~F) 1200 Stress (psi) 12, 500 Rupture Time Elong. (hours) (%) 109 50 Hours at 1200~F (C=20) R.A. (%) 72 Re -Normalized 1675~F + 88 - -- -- 1L5 1L6 1L 1L8 11:7 1L 1T3 1T4 1000 1000 1000 1000 1000 1000 1000 1000 1200 46, 750 35, 000 30,000 25, 000 22, 000 20,000 30,000 20, 000 12, 500 1675~F + 300 STTT 12 33 116 382 826 50 1003 75 39 45 52 23 39 51 49 45 81 79 84 79 84 82 74 78 1L35 60 79 Re -Normalized Hours at 1200~F (C=15) 1L9 1L10 1Lll 1L12 1T5 1T6 1000 1000 1000 1000 1000 1000 1200 45, 000 35, 000 24, 000 18,000 30, 000 20, 000 12, 500 STTT 12 160 1224 24 677 35 39 37 49 62 47 61 67 81 80 83 67 73 79 1L32 84 Re-Normalized 1675~F + Tempered at + 88 Hours at 1200~F 1325~F for 1 Hour 1 L40 1L47 1000 1000 1200 25, 000 21, 000 12, 500 153 761 102 61 47 83 85 1 L48 50 80 Annealed 1675~F, 1 Hour, F.C. at 125~F per Hour to 1300~F and F.C. to Roo,-n Temperature in 16 Hours D-11 1200 12, 500 91 53 78 26

TABLE II(Cont.) Specimen Code Test Temp. (~F) Stress (psi) Rupture Time Elong. (hour s) (%) R.A. (%) Re-Normalized 16750F + 20 Hours at 1200~F, under 7,800 psi - -- -- - - --- - -- --- ---- 1L17 1L19 1L20 1T9 1T10 1000 1000 1000 1000 1000 52,250 40,000 30,000 35,000 25,000 STTT 5.9 131 15 507 37 36 44 41 51 82 76 65 72 77 Re-Normalized 16250F + 88 Hours at 12000F, under 7,800 psi 1L21 1L22 1L23 1L24 1T11 1T12 1000 1000 1000 1000 1000 1000 45,000 32,000 23,000 2p, 000 26,000 21,000 STTT 10 378 980 99.6 565 39 45 62 48 54 61 82 81 82 79 78 80 Re-Normalized 1675~F + 300 Hours at 1200~F, under 7,800 psi 1L25 1L 1 L26 1L27 1L28 1T13 1T14 1000 1000 1000 1000 1000 1000 1000 39,125 30,000 24,000 20,000 17,500 22,000 19,000 STTT 13 110 455 1288 289 630 3~ 36 36 39 52 57 64 82 79 79 78 72 77 77 27

TABLE III Long-Time Strengths at 1000~F for Pll Pipe Extrapolation* Condition Method Rupture Strength (psi) 83,000-Hour Strengths As Removed from Service "Original"-Re-normalized + 2 hrs at 1200~F Simulated Service Exposures Re-normalized + 20 hrs at 1200~F Re-normalized + 80 hrs at 1200~F Re-normalized + 300 hrs at 1200~F SL LM-20 SL LM-20 SL LM-20 SL LM-20 SL LM-20 13,200 12,800 13,300 15,300 13,000 11,900 12,200 10,900 10,900 16,600-Hour Strengths "Original"-Re-normalized + 2 hrs at 1200~F Simulated Service Exposure Re-normalized + 20 hrs at 1200~F 360,000-Hour Strength "Original"-Re-normalized + 2 hrs at 1200~F Simulated Service Exposure Re-normalized + 300 hrs at 1200~F *SL —Straight line rupture curve. LM-20 Larson-Miller parameter with C of 20. LM-20 SL LM-20 SL 17,000 19,200 10,600 9,600 28

TABLE IV Rupture Test Data for the Original P22 (2. 25Cr-lMo) Steel Pipe (As Produced + 2 Hours at 1200~F) Specimen Code 2L16 2T1 2L17 2L18 Test Temp. (~F) 1000 1000 1000 1000 2L21 2L19 2L46 1050 1050 1050 2L47 2L29 2T23 2T3 1100 1100 1100 1100 Stress (psi) 50, 000 35,000 31,000 27, 000 25, 000 19,000 18, 500 19,000 18, 000 17, 000 15, 000 19, 000 17, 000 13, 000 13, 000 10, 000 10, 000 10, 000 8,000 7, 000 4,000 Rupture Time (hours) STTT 31 72 239 49 1210 2190 233 360 1752 3465 53 308 1035 225 754 141 55 119 168 1096 Elong. (%o) R.A. (%) 28 42 69 50 40 41 51 47 42 25 27 40 41 43 43 34 43 58 77 33 35 79 80 85 86 87 86 86 88 85 81 77 88 87 86 87 88 74 94 95 93 91 2T4 2T24 2L25*.* 2T2 2L26"** 2L27** 2L28 * * 2T8 "*, 2L20 2L33 1150 1150 1150 1200 1200 1250 1300 1300 1300 1300 Code L - Longitudinal Specimen T - Transverse Specimen STTT - Short Time Tensile Test - Exposed for 20 hours at 1200~F - used on the basis that this would not influence the properties 29

TABLE V Simulated 83, 000 Hours of Service at 10000F on P22 (2. 25Cr-lMo) Steel Pipe Specimen Code Test Temp. (~F) Stress (psi) Rupture Time (hours) Elong. (%) R.A. (%) Exposed 20 Hours at 1200~F (C = 25) L,..1 * 1 2L22 2T5 2L23 2L24 2T6 2T7 1000 1000 1000 1000 1000 1000 50,500 35,000 31, 000 25, 000 23,000 22, 000 STTT 26 66 520 1801 29 48 46 43 44 79 81 83 85 86 Exposed 88 Hours at 1200~F (C = 20) 2L30 2T11 2L31 2L32 2T12 1000 1000 1000 1000 1000 50,500 49,600 31, 000 27, 000 22, 000 STTT STTT 34 166 1767 32 27 50 77 30 79 68 83 84 81 Exposed 300 Hours at 1200~F (C = 15) 2L38 2T17 2L39 1000 1000 1000 47, 000 31, 000 28,000 STTT 109 105 37 38 46 82 82 84 Exposed 300 Hours at 12000F, under 7,800 psi 2L42 2T20 2T22 1000 1000 1000 44, 700 35,000 27, 000 STTT 7.4 124 27 22 29 78 75 82 *Code L - Longitudinal Specimen T - Transverse Specimen STTT - Short Time Tensile Test 30

TABLE VI Simulated 83, 000 Hours of Service at 11000F on P22 (2. 25Cr-lMo) Steel Pipe Spe cimen Code* Test Temp. (0F) Stress (psi) Rupture Time (hours). ~ Elong. (%)_ R.A. (%) Exposed 115 Hours at 1300~F (C=20) 2L13 2L14 2L10 2L9 2Lll 1050 1050 1100 1100 1100 1250 1300 1300 1300 20,000 17, 500 16, 000 14, 000 12, 500 9,000 8, 000 7, 000 4, 000 182 597 155 451 1297 108 38 140 1272 63 62 64 44 53 100 86 47 31 87 85 89 89 90 91 95 96 93 2L15 2L12 2T 19 2L40 *Code L - Longitudinal Specimen T - Transverse Specimen 31

TABLE VII Long-Time Strengths for Grade 22 Pipe Extrapolation* Method Rupture Strength (psi) Condition 83,000-Hour Strength at 1000 F "Original" —As Produced + 2 hours at 1200~F SL LMSimulated Service Exposures As Produced + 20 hrs at 1200~F SL As Produced + 88 hrs at 1200~F SL As Produced + 300 hrs at 1200~F SL 83,000-Hour Strength at 1100~F "Original"-As Produced + 2 hrs at 1200~F SL LMSimulated Service Exposure As Produced + 115 hrs at 1300~F SL LM-: *SL-Straight line rupture curve LM-20 Larson-Miller parameter with C of 20 20 13,000 16,000 15,800 15,900 15,000 20 20 14,200 7,600 7,600 6,700 32

TABLE VIII Rupture Test Data for the Original Grade 22 Steel Tubes Tube Stress (psi) Rupture Time (hours) Elong. (%) Temp. R.A. (%) Heat Treated 1 hour at 14500F, Air Cooled A 1100 16,500 14,500 13, 500 12, 000 10, 181 10, 000 6, 000 6, 000 864 1575 3225 7014 155 174 1672 216 43 43 36 28 32 23 63 67 56 51 85 88 79 1230 1300 28 90 Modified Iso-Thermal Treatment B 1100 B 1230 1300 21, 000 19, 000 19,000 16, 000 12, 500 10, 000 12, 500 9,000 5, 000 9,000 6, 000 4,000 20,000 17,500 15, 000 12,500 10, 000 10, 000 8,000 5, 000 150 613 464 1279 2382 8475 43 291 2335 16 210 651 55 139 496 1942 10, 327 141 500 3788 12 54 20 50 38 31 30 75 54 56 36 52 86 87 57 89 72 69 25 C 1100 45 43 37 75 69 57 42 36 1230 53 33 14 72 60 52 33

TABLE VIII (Concluded) Rupture Test Data for the Original Grade 22 Steel Tubes Stress Temp. (psi) Rupture Time (hours) Elong. ( ) R.A. Tube Modified Iso-Thermal Treatment C 1300 1100 D 8,000 6,000 6,000 4,000 20,000 15,000 12,500 12,000 10,500 9,000 8,500 6,000 7,000 4,475 19 554 256 777 12 106 466 861 1578 5871 145 1551 95 837 55 66 47 34 47 34 22 20 75 67 81 51 70 51 48 42 45 36 1230 1300 52 52 61 34 78 49 34

TABLE IX Tensile and Rupture Data at 1100~F for Grade 22 Steel Tubes Exposed to Simulate 10,000 Hours at 1100~F Stress (psi) Rupture Time (hours) Elong. JIL) R.A. (~) Tube Heat Treated 1 hour at 14500~F, Air Cooled A 29,500 17,500 15,000 12,000 STTT 40 189 1572 53 51 61 39 86 86 86 83 Modified Iso-Thermal Treatment B C D 30,600 17,500 17,500 15,000 12,000 29,700 17,500 15,000 12,000 28,000 15,000 13,500 13,500 12,000 10,000 STTT 20 68 247 1915 STTT '51 246 1910 STTT 29 209 274 450 1927 56 59 47 37 43 56 35 55 36 54 50 51 38 53 30 91 87 89 89 76 89 76 65 54 83 70 66 59 67 49 STTT-Short Time Tensile Test 35

TABLE X Tensile and Rupture Data at 1100~F for Grade 22 Steel Tubes Exposed to Simulate 100,000 Hours at 1100~F Stress (psi) Rupture Time (hours) Elong. (i), R.A. (~) Tube Heat Treated 1 hour at 14500F, Air Cooled A 30,100 17,500 15,000 12,000 STTT 53 156 741 53 60 51 41 88 87 87 87 Modified Iso-Thermal Treatment B C D 32,400 17,500 15,000 12,000 10,000 31,300 17,500 15,000 12,000 10,000 28,900 15,000 13,500 12,000 10,000 STTT 41 156 574 2165 STTT 32 163 568 3184 STTT 9 189 512 1450 52 54 65 74 48 52 57 62 33 57 65 59 52 35 90 89 88 92 86 87 81 79 74 86 64 65 63 53 STTT-Short Time Tensile Test 36

TABLE XI Long-Time Strengths at 1100~F for Grade 22 Tubes Extrapolation* Method Rupture Strength (psi) 10,000 Hour 100,000 Hour Tube Condition.... A Original Simulated Exposures B Original Simulated Exposures C Original Simulated Exposures D Original Simulated Exposures SL LM-20 SL SL LM-20 SL SL LM-20 SL SL LM-20 SL 11,100 10,900 10,100 9,700 9,700 9,800 10,000 10,000 10,000 8,400 8,400 8,200 7,800 6,500 6,100 5,900 5,900 5,800 7,500 7,000 6,700 6,300 6,300 6,200 *SL Straight line rupture curve LM-20 Larson-Miller parameters with C of 20 37

TABLE XII Rupture Original Test Data for Type 304 Stainless Steel in the Condition and After Simulated Service Exposure Test Temp. (~F) Stress ( psi) Rupture Time (hours) Elong., (I) R.A..(%k Annealed 1950 - 1975~F, 1 hour W.Q. 1200 1200 1300 1300 1335 1350 1425 1450 1500 24,000 18,500 16,500 13,500 13,500 28,225 7,500 8,000 7,500 92 790 98 516 160 0.2 1173 311 141 29 25 37 33 39 55 21 43 29 22 35 27 31 47 24 30 40 Annealed + 100 hours at 1425~F 1200 1200 1200 1200 1200 1300 1300 1350 1350 1425 30,000 24,000 21,000 18,500 18,000 16,500 13,500 13,500 13,500 7,500 15 149 415 1295 1030 115 421 117 123 838 49 57 75 64 92 79 76 44 39 48 55 56 63 57 61 62 36 Annealed + 100 hours at 1525~F, under 8500 psi 1100 1200 1200 1200 1200 1350 4o,000 30,000 24,000 21,000 18,000 13,500 20 14 155 374 1033 148 59 57 56 53 48 55 47 52 49 41 38

TABLE XIII Long-Time Strengths for Type 304 Austentic Steel Temperature (~F) Extrapolation Method 100,000 Hour Strength (psi) 1100 1100 1100 1100 1200 1200 1200 1200 Straight Line Straight Line Larson-Miller Larson-Miller Straight Line Straight Line Larson-Miller Larson-Miller As Produced Material - Tests to 1,000 hrs. - Tests to 50,000 hrs. Parameter (C=20) Optimized Parameter (C=17.95) - Tests to 1,000 hrs. - Tests to 50,000 hrs. Parameter (C=20) Optimized Parameter (C=17.95) 19,000 13,500 16,200 14,900 10,200 8,000 11,000 9,800 Simulated Service at 1200~F by Heating 100 hours at 1425~F Straight Line Log - Log 10,200 Parameter 10, 200 1200 1200 39

(a) As Removed from Service 40 i/n un U. U' I — u, 20 10 8 6 1000 0 100, AA" IZVV Longnitudinal 0 Transverse 0 40 In UR I&' 20 p- P% - T0 I.T 05 1-. - -- - - (b) Re-Normalized 1675~F plus 2hrs 1200 ~F "'^_ _- " Extrapolated using Larson-Miller -,_ Parameter -Fig. 2 I I I I 11111 1 I I 1 I 1 10 8 6 I I I I I I ll I I I I I I I 1 I I I I I I I I I I I I I 10 100 1,000 10 RUPTURE TIME, HOURS Figure 1. Stress versus Rupture Time Data for P11 Pipe As-Removed from Service and after Reheat Treatment.,000 100,000

40 Re-Normalized 1675~F plus 2hrs 1200 OF 20 0 0. 0 ~8|~ | As Removed from Service, 1 1 10 83,000 Hours 8 - at 1,000F ~ 6 32 34 36 38 P=(T + 460) (20 + LOG t) X 10 Figure2. Stress versus Larson-Miller Parameter Curves for P11 Pipe As Removed from Service and after Reheat Treatment.

40 (a e4t --- - Norm. 1675~F + 20hrs 1200 ~F 0 30 v 20 As Removed from Service I. 10 40 (b) 30675F 88hrs 1200 20 m, 7800 psi No Stress amt | Longnitudinal 0 w) L Transverse I 0 10 40 (c) ~ 30 75 F + 300 hrs 1200 OF 20.I-, C1 10 II 1111... I I I I I 1i l 1 11 1 1111 I I 10 100 1,000 RUPTURE TIME, HOURS Figure 3. Stress-Rupture Time Curves at 10000F for P11 Pipe As-Removed from Service and after Reheat Treatment and Exposure to Simulate Service at 1000~F.

mIt 0.X 00 at 0* E; n 0 z 0.t.9 0 c a t 0 0 -4 __ 0 0.t N 0 0 4.4 0 IN:@ 3 odr _1S f _g n 5 l* 0 0 z 0.e 0 4S0 etj~a

' 60 0 0 0 _- An I As Produced Plus 2hrs. at 1200~F 0 -~~- ~20 hrs. " D 88 hrs t" 0 300 hrs " A ~7 -- - -- ^^Z^ ^[>^o~*-""*^^ ^^300 hrs ff at7800 psi. V iL fv r u I4m 20 I 111 I I I I I I 11 I I I I I I I I I I I 10 100 1000 RUPTURE TIME, Hours Figure 5: Stress-Rupture Time Curves at 1000~F for P22 Pipe in the Original Heat Treated Condition and after Exposure to Simulate Service at 1000~F.

40 V 20 a 10 us 8 ac I#A 6 4 k As Produced Plus 0 2hrs at 1200~F -- ^ -.. ----^* ~ 115hrs at 1300~F --- -- ~ ^ B~r'-~ Extrapolated using Larson-Miller Parameter -Fig 7 1250......., 1300 i I I I I 1I I I I I I I I I II I I I I I I I i I I I I I I I 10 100 1,000 RUPTURE TIME, 10,000 100,000 Hours Figure 6: Stress-Rupture Time Curves for P22 Pipe in the Original Heat Treated Condition and after Exposure to Simulate Service at 1100~F.

40 83,000 hrs at 1100~F 20 1-._ $ o &e 10 As Produced Plus 2hrs at 1200~F O 115hrs at 13000 F * ut IL IV) 8 6 4 I I I I I I I I I m 32 34 36 38 P = ( T+460)(20 + Logt) X 10O3 Figure 7: Stress versus Larson-Miller Parameter Curves for P22 Pipe in the Original Heat Treated Condition to Simulate Service at 11000F. 40 and after Exposure

(a) As Produced + 2 Hrs. at 1200~F (b) As Produced-+ 8:r': a t 1200 ub) As& crostructure s of P22 P0pe X 1000

(a) TUBE A Extrapolations: 20 -_ Original -Larson Miller, C=20 ""-~~'""J^ i,,i o^"^^ -- = Exposed - Straight Line 10 v' 1300 10 = ~ 6 4 ~~~~~~~~~~~~~~~~~~(b) TUBE B ~Heat Treated Condition 20_ _ T 0 Original 20'p::::^ "t ~Sulat10, Simulate 10,000 hrs(1100~F 0S. E5-. L9-^ --- ^^- -= Simulate 100,000 hrs Data) ~~~~~~~~~~~10 1 6 - - 1 300 4 L, 1111111 I Illill,11111 I I I 111111 III 1 11111I 10 100 1,000 10,000 100,000 RUPTURE TIME, Hours Figure 9: Stress-Rupture Time Curves for Grade 22 Tubes in the Original Heat Treated Conditionsand after Exposure to Simulate 10,000 and 100,000 Hours at 1100~F.

a. Ia. 0 LAJ 1v 20 10 8 6 4 20 10 8 6 4 (c) TUBE C 1100~F = ~~~~~-=~~~~ ~230 ==_..,.-._ CX ~~ ~ ~ 300 (d) TUBE D 1230...1100. F 1230 fmt~ -a 300 I I I I I I 111 I I I I I I 111 I I I I I I 111 I I I I I il 10 100 1,000 RUPTURE TIME, Hours 10,000 100,000 Figure 9: (Continued)

' ~@;~.;~,'.:.t.x,.,~........:::,: ',,..,':,,"ii:!:I:.........:# +. ':" ":...............sX.,'% *............. ~...::':.::, ~;::::;:: '.%::'::::.::::::::::::::: ':~: ' - ~:,:;.::::::::::::::?::::::.::::!::: ~...... l% 4:. '.7 ':::.:-:t-,,,',,,: s 8 ta t * ^- z S IE ~ ~~~~~~,,.- jl~,.~.,.., 0 * 40w 2 tt j Ad0XXf2i0XAt*-'Oe t@ o s' * > XX X ' + Y ' '.oA..0 0 %:'.........:'::::~ ~i:jrj_::: i:~i~iiiiiii ~ B 60 ~*:::::::::-:: -::::::~::::i::,:I:ii:-~l:.l:`-::: I~~l~~:::::.~:Ot tiiiil~ii x a:i~:,-r::::i rr~~ili: 4~;li::~:::_ il-ii-i:::i~i.:l::i::_:tl:::-:::::::-:: I iii~iiil:;ii-:-i *14P::a,~::::-::~i ~;;cl~~j~~: ~~: ~l::: -::i —iii'::illii-~~iiiil~iiiii:-il:i:-:~i~il ~:::::::i~ii:.::_:::-:::::::T

M i: r.!0W0,' 0 Sf 0x: 0ldii04 ':'................ '..........:. f;i.:; t!}i:: E s ' *'4.:::'T:': AX.......~ ' _'f 'i:.:.0 4.i:i;::.''::0: 000 'i;/ivo.? S' f,/ Z:i..` _ 6, 7: i k. - +*a 'fe -v><>* f+-, *~ 2 0 * I * 0; '0+0..t;4 C: - O ~ Os~ '"YS m,%ws-.S.4 ~'~~~~~~~~~~~~~~~~~~'. 4,,, C. r:................. ' I,"....s,:. '; i: ':.::.,' -le - r;_S ~' s}0 o:': ':':2..:\".':':'.'4+':':.:: _A'... -:..::u, >:;E,> + *:,. E::*:. < t *<:,..::.. "~:} '~' "..4t;":.,*0~~~~~~~A 0" "...... "~' - " '.~... 0,'~~~~~~~~~~~~~~~~~~~~~~~~ ',' >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~" '-4 '.,,'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:.:...::, *," ":'".... Hil........:..,... ~S:..:................., C ~~~~~- ~ ~ A O! ~.......'....,..... x~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.. *~~~~~~~ ~...... '.4~~. ~:.'::.:..:..:':' ~: ':...:::":' _7,. i,.b. ".*9$0-:g,$.:0 4":-Ma "':-4*~4 4.:.?.'1>%:L. 'i.,::0 o ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ wow~~~~~~~~~~o x.'4 *4',~~~~~ S..~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~X

0 0 o 0 0 0 0 0 0 rm C. E Vm 0 0 0 X U. o:: o 0 0 W NO' 0 0 0 0 4 -

(a) As Produced (b) Simulated 10,000 Hrs. at 1100~F ^ ^ ^ ^ ^r^.^- i+ 4'.4 -t 4~7-* F **Wx,9. 'a/ ha. ki.Q. -..-t - (c) Simulated 100,000 Wrs. at 11000F (d s871 Hour Test at 1100~F Figure 13* Microstructures of Grade 22, Tube D. X 1,000

I( t0 l (a) As Produced 60 0 AR-2 ^ 41 -0 -- _r a 1350~F u=e U. M ~0 3 20 10 m- 8 6- - (F ~~~~~~~~~~~~~~~~~~~~~b) ~~~(b) Simulated Service Exposures ro 2 -00 hrs at 1425 ~F c 1 20 F Q- O 100hrs at 1325~F under 8500psi 6 I I I I I I I I I I I II I I - - I I I I 1 1I 1 10 100 1000 10,000 100,000 RUPTURE TIME, Hours Figure 14: Stress-Rupture Time Curves for Type 304 Austenitic Stainless Steel in the Original Heat Treated Condition and after Exposure to Simulate Service at 1200~F.

40 I-I Parameter Data o 1,000 ~F Data o 1,200 ~ F Data 20 ^ 10 co*D m - 8 33 35 37 39 41 43 45 47 P=(T+460) (20+LOG t) X 10-3 Figure15: Stress versus Larson-Miller Parameter Curves (C=20) for Type 304 (AR-2) Material in the Original Heat Treated Condition (ref.3)

(a) As Produced (b) Exposed 100 Hours at 1425~. (c) Cold Reduced 40 Percent plus 20 Miin. atof 16250: plus 100 HFours atof 1425F Figure 16: Microstructures of Type 304 Ausftenitic Stanless Stfeel. X 1,000

:0 0 3; 0 o: 0.O o _e 0 0:O C 0 ~ 0 0 0. 0.. 0.