The American Society of Mechanical Engineers 54-A-73 29 W. 39th St., New York 18, N. Y. lh'5s | ~ ~ i~CARBON-MOLYBDENUM STEEL ST TEAM PIPE %^ AFTER 100,000 HOURS OF SERVICE V I R. Jo Sinnott, Member, ASME Engineering Department c9'~~~j~~ ~The Detroit Edison Company Detroit, Michigan I. A. Rohrig J, W. Freeman and A I. Rush Engineering Research Institute, )m^ w -~ University of Michigan Detroit, lfMichigan Advance Copy Released for general publication upon presentation Contributed by the Joint ASTM-ASME Committee on Effect of Temperature on Properties of Metals for presentation at the Annual Meting, | New York, NoYo - November 28 - December 3, 1954 of The American Society of Mechanical Engineerso (Manuscript received at ASME _;(l0a(^ headquarters August 2p 1954o) ^\ I Written discussion on this paper will be accepted up to January 10, 1955. Decision on publication of this paper by ASME had not been taken when ^A ~ this preprint was preparedo Discussion is published only if the paper is published.. - w Pried i n 5 (Preprints will be available ntil October 1, 1955) j~~xatea znr Ugibolo

ABSTRACT Rarely does the engineer or metallurgist have an opportunity to evaluate design considerations and laboratory data in terms of creep in services Carbonmolybdenum steam pipe, carefully measured for service creep during 100,000 hr of operation at 900 F, was subjected to laboratory examination after removal from service, The purpose was to check calculated service creep rates, assess creep damage, and to compare longtime performance prediction based on short-time laboratory data. Remarkable correlation was observed between calculated service creep rates and those established by subsequent laboratory creep testing. Full agreement with average values used by the Subgroup on Allowable Stresses for Ferrous Materials of the ASME Boiler Code Committee in setting allowable stresses for this material was established for both creep and stress-rupture propertieso

2653 154 CARBON-MOLYBDENUM STEEL STEAM PIPE AFTER 100,000 HOURS OF SERVICE By Ro J. Sinnott, Io Ao Rohrig, J, Wo Freeman, and A. I.o Rush INTRODUCTION When Unit No0 14 was being erected at the Delray Power Plant of The Detroit Edison Company in 1938, the lack of reliable data on service-life properties of metals subjected to high-temperatures was felt acutely. The practice of basing design on stresses obtained by 1000-hr high-temperature laboratory tests extrapolated to 100,000 or more hours has repeatedly been questioned0 Designers accepted such data only because no better method of property prediction of hightemperature characteristics has been available, It is the purpose of'.this paper to present (a) The results of creep measurements made on carbon.molybdenum pipe subjected to 100,000 'hr of actual power-plant service, (b) results of after-service laboratory testing, and (c) a discussion of service results as compared to properties predicted by laboratory testing prior to serviceo Service History With the installation of Unit No0 14, a service creep-measurement program was initiated~ Stainless steel-measuring points were arc-welded on the external surfaces of two 10-ino nominal diamo carbon-molybdenum steel pipes connecting the turbine emergency stop valve to the upper and lower steam chests of the turbine. The two Schedule 80 steam leads were designated as "North" and "South"t indicating their position with respect to the turbine. The stainless-steel buttons were located to provide both diametral and axial measuring stations and were ground to give accurate measuring surfaces. Figo 1 illustrates the measuring stations. Weighted average pressures and temperatures during the operating period are 835 psig and 900 Fo Maximum temperature fluctuations of + 20 deg F represent normal operating conditionso Measurements were made at prevailing temperatures during five outage periods and were taken after the unit had cooled dorn and the thermal insulation removed from the locations undergoing testo Dimensions were measured with an outside micrometer caliper and a special micrometer trammel0 Readings taken were corrected to a base temperature of 68 Fo The results of these measurements are shown in Table Io Plotted service creep calculations based on diametral measurements indicated a typical low stress creep versus time curve over the first 67,000 hr of operation; viz, a relatively rapid initial creep rate followed by a reduction in rate of creep, which is characteristic of material operating under stress conditions producing creep at a low rate0 Total diametral elongation at this time was observed to be in the order of Ool per cent, a relatively low value0 The total axial elongation during the 67,000-hr period as determined by axial measurements did not exceed 0003 per cent and in most cases was considerably lesso

-2 -After 75,500 hr of operation, weld samples removed from the valve-to-pipe and pipe-to-pipe joints of the South connection indicated the presence of graphite ranging from dispersed nodular in the valve joint to chain type in the pipe-topipe jointso At shut-down intervals during the next 5000 hr of operation, all. the welded joints in the turbine leads were first normalized at 1725 F, and shortly thereafter all welds were gouged out, rewelded, and normalizedo After 100,135 hr of operation, the carbon-molybdenum turbine leads were removed and replaced with chromium-molybdenum pipe, which is highly resistant to graphitization. Final service creep measurements were made just prior to the removal of the carbon-molybdenum pipe from services Calculations based on diametral measurements indicated a drastic increase in creep rate over the rate established at the end of the fourth periodo Concern over the carbon-molybdenum pipe remaining in service in other portions of the system indicated the advisability of laboratory tests as a check on calculated service creep rateso Description of Material The pipe was ASTM A158-36 grade P1, schedule 80, made from National Tube Company's Heat 10043o The reported chemical composition wast Per cent Carbono 0 0 0 0 0 0 0 0 0 o 13 Manganese. 0 0 0 0 0 0 0 0 0 o0045 Silicon o o o o o o 0Ool31-0ol35 Molybdenum 0 0 0 0 060-0.62 Molybdenumo o o o o o o o o o o 60-062 The pipe was 10o75 ino OD by 0o593 ino wallo After bending and upsetting the ends, it had been given a full anneal at 1900 F for 2 hro Physical properties were reported as followss Tensile strength, psio o o o o 60,880 Yield strength, psio o o o o o 399730 Elongation in 2 in,, per cento 45 EXPERIMENTAL PROCEDURE Creep Testing tNorth Connectiono Creep tests were run on both tangential and longitudinal specimens at the operating temperature of 900 F and under the estimated operating stress of 7500 psio In addition, tests were made under the present 12,500-psi allowable stress of the ASME Boiler Codeo

-3 -"South Connection"o One tangential specimen and one longitudinal specimen were taken from the pipe, Creep tests of these two specimens were run at 900 F under a stress of 7500 psio Rupture Testing "North Connection". A stress-rupture curve was established at 900 F for time periods up to nearly 1iOO hro Longitudinal specimens were used for this testO In addition, a check test was run on a tangential specimen which was tested at 37,000 psi, a stress intended to cause rupture in about 1400 hro "South Connection"o Tests were run on longitudinal and tangential specimens at a stress of 37,000 psi Impact Tests "North Connection"o Charpy V-notch tests were made on both longitudinal and tangential specimens0 Similar tests were made on the pipe material after it had been reheated 2 hr at 1900 F and furnace-cooled~ All of the Charpy V-notch specimens were tested at 80 Fo No impact tests were made on material from the South Connection Tensile Tests North Connection"o Room-temperature tensile tests were run on longitudinal specimenso No tensile tests were made on material from the South Connection. Hardness Tests "North Connection" Brin/L hardness tests were run on eight sections representing the entire crcumference of the pipe sectiono "SouthCconnectiono, The Brinell hardness was determined on one section representing the South Connectiono Metallographic Examination "North Connectiono A metillographic examination was made on sections through the center line of the measuring buttons representing the four quadrants of the North Connection, and, in addition, samples were taken adjacent to all of the test specimens. "South Connection" Samples were taken at one measuring button and also adjacent to the creep and rupture specimens.

-4 -RESULTS Dimensional Measurements on the North and South Connections The measurements shown in Table I were made across the diameters of the pipe at the reference buttonso It should be noted that at the time the reference buttons were attached, no attempt was made to keep the height of the measuring buttons uniform. Accurate inside diameter measurement of the pipes was not readily obtainable with the equipment available at.,the time of installationo Therefore, any initial ellipticity of the pipes which may have existed was not detected, and was not indicated by the diametral measurements made during serviceo No changes in diameter due to sectioning were observedo The inside diameters were measured at both ends of the pipe ring containing the reference buttons and are recorded also in Table I. Measurements were made on the pipe diameters containing the reference buttons and at diameters between each buttono Further, the circumferential length covered by the buttons was determined as a matter of record with the results shown in Table Io From these data, it was observed that the minimum diameter of the North connection was at the (4-4) axis and at the (1-1) axis of the South connection. These diameters correspond to those showing the greatest increase from creep as judged by the diametral measurements taken during serviceo Consequently, it must be concluded that the pipes were originally elliptical in shape and were becoming round during services The wall-thickness measurements indicated that both sections varied by a maximum of approximately 0,025 to o0030-ino, a variation which was well within the specified allowances for the size of pipeo However, it should be emphasized that these measurements were made after 100,000 hr of service and that original wall thicknesses were unknowno Creep Rates During Service The average rates of creep during service were calculated from the measurements taken on the measuring points during service. The results obtained from such measurements were used to calculate the creep rates by two separate and different approaches; namely, changes in diameter of the pipe, and, changes in the perimeter of an ellipse having the dimensions established by the changes in diameter. The latter method is preferred; therefore, the results given for creep during service are based upon changes in the perimeter of an ellipse having the dimensions established by the measurements of the pipe given in Table Io The results of these calculations are given in Table IIo Fig. 2 shows graphically the changes in circumference of the pipes as a function of time and serviceo The average circumferential creep rates during service show the followings 1 During the first period, the creep rate of both sections was somewhat higher than 0.003 per cent per 1000 hro

-5 -2 During the next three periods, the rates were less than 00022 per cent per 1000 hro However, the minimum creep rates were observed for both sections during the third periodo 3 During the fifth period, the average rate for the South connection apparently increased to Oo0047 per cent per 1000 hr, whereas the creep rate for the North section was essentially the same as for the second and fourth periodso It should be emphasized that the measurements made during service reflected very small changes in the diameters as shown by Table Io Consequently, it would not appear that too much significance should be placed on the minor differences in creep rates observed for the different periods? ioeo the low creep rates calculated for the third period of operation The somewhat higher average creep rates during the first period might have been expected due to the combined effects of primary creep and possible relief of stress concentrations0 The apparent increased creep rate of the South connection during the fifth period is clouded by the effects of the removal of the old welds and the reweldingo It does appear highly significant, however, that in general the creep rates and total deformations, are of the order expected from the extrapolation of laboratory datao Tensile and Hardness Properties Tensile tests conducted at room temperature gave the data shown in Table IIIo Material from all four quadrants had very uniform tensile strengths and ductility valueso Yield strengths and proportional-limit values varied somewhato There did not, however, appear to be a consistent relationship between specimen location and variation in these valueso Lack of specific comparative data raises the question of whether the low tensile strength was the result of service or if it is characteristic of the somewhat high temperature and slow rate of cooling during the original heat-treatment. This heat-treatment might be expected to give somewhat lower tensile strength than those usually used for reported data for Oo5 Mo steelo For similar reasons, the yield strengths might have been expected to be lowo It, therefore, appears that service may have raised the yield points somewhat as would be expected. Brinell hardness determinations were as follows: Brinell hardness Connection Location Number North Top, side and bottom of pipe 103-107 South Side of pipe 101 There did not appear to be a significant variation in hardness around the pipe circumferenceo The hardness values, like the tensile strength, appear to be somewhat low for 0O50 Mo steel0 Apparently the hardness was characteristic of the original heat-treatment, because reheat-treatment at 1900 F for impact testing resulted in a Brinell hardness of 101 to 106o There is, therefore, no real evidence of softening during serviceo

-6 -Impact Properties Charpy V-notch tests were made at 80 F on specimens taken from the pipes as removed from service and after sections of the pipe had been given the original heat tre&tmento The data obtained, Table IV, indicate the followings 1 The impact strength ranged from 8 to 11 ft-lb for the pipe material asremoved from service0 2 Reheat-treatment of the pipe material with the same nominal heat-treatment as the original pipe gave an impact strength of 18 to 22,5 ft-lbo 3 The range in impact strengths was too small for any real significance to be attached to the difference in specimens taken longitudinally and tangentially or from the locations of estimated highest and lowest stress concentrations~ 4 The influence of prolonged service at 900 F on the impact strength of 0.50 Mo steel is not well establishedo Such information as is available would indicate that sme -deterioration would be expected, At least, that is the usual experience in creep-testingo Consequently, it would seem the observed values of 8 to 11 ft-lb do not appear unusual for material with the original heat-treatment used for the pipe in question0 However, if the pipe had been normalized from 1650 F in place of 1900 F, it is probable that the values would have been higher than 8 to 11 ft-lbo Laboratory Creep-Test Properties The creep-rate data obtained from the individual tests are shown in Table V, and the log stress-log creep rate curves derived from the creep-test data are shown in Fig. 3o The curves of Figo 3 indicate that the stress for a creep rate of OoO1 -per cent per 1000 hr was 12,500 psi. Tests were run at two stress levels: (a) 7500 psi - the stress corresponding to the operating stress during service calculated by The Detroit Edison Company. The measured creep rates ranged from less than OoOOl to 0,002 per cent per 1000 hro These rates are not considered as precisely established because the sensitivity of the extensometer system, particularly for the 1-ino gage lengths of the tangential specimens, was too low to give exact valueso There is no doubt, however, that the creep rates were of the order of 0,001 per cent per 1000 hro (b) 12,500 psi - the stress corresponding to the present allowable stress under the ASME Boiler Code, Both a longitudinal and a tangential specimen from the North connection gave final secondary creep rates of 0.01 per cent per 1000 hro Tangential, as well as longitudinal, specimens were tested because the creep during service was largely circumferential and it was felt that creep characteristics should be established for material with the same orientation as the service creepo The absence of any substantial difference between the two types of specimens appears somewhat unusual, because specimens taken across the direction of

-7 -metal flow during working usually have slightly higher creep resistance. The absence of specific data on this point for pipe produced and heat-treated in the same way as the pipe tested, however, make this uncertain for the particular caseo The tangential specimens tested at 7500 psi showed a decrease in length during the first few hours of testingo The reason for this is uncertain0 It could have been due to relief of some complex internal stress system or to testing technique variables Careful review of the testing procedure showed no reason to believe that it was testing technique0 The 7500-psi tests on specimens from the North and South connections did not show a significent difference in creep rateo The creep rates were so low, however, that it is not certain that some difference in creep resistance did not existo Rupture-Test Properties The data obtained from the rupture tests are given in Table VI and are plotted as the usual log stress-log rupture time curve in Figo 4o From this curve, the following rupture strengths at 900 F have been estimated: Stress for Rupture in Indicated Time Periods at 900 F (psi) l00-hr l000hr 10,000-hr 10000 0-hr 37500 37000 33000 2900Q The extrapolation to 10,000 and 100,000 hr appears somewhat uncertain in that there seems to be a break in the slope of the stress-rupture curve at about 1000 hr. The tests at longer time periods are insufficient to define the slope of the curve for longer time periods with certaintyo The curve has been extended using the greatest slope (lowest strengths) indicated by the available datao Elongation and reduction-of-area values decreased with rupture timeo The short-time elongation was approximately 40 per cento Tests between 1000 and 2000 hr in duration showed elongations between 30 and 20 per cento There was no significant difference between longitudinal and tangential specimens or between the two pipe sectionso Metallographic Examination Test specimens were taken adjacent to or between measuring buttons, The microstructures typical of the test specimens were as follows2 1 The microstructures were ferrite and slightly spheroidized pearliteo In addition, there were apparently fairly massive carbides in the grain boundaries and a light general precipitation throughout the matrixo Scattered graphite nodules were presento 2 The grain size was mainly 5 to 80 No areas of larger grains were observed in the North connection, Occasional nodules of graphite could be found, and it was

-8 -noted that the massive carbides were absent in the presence of graphite0 3 No significant variations in structure around the circumference of the North connection was observedo The examination of the South connection was less extensive, All samples were, however, similar to the North connection, except for the occurrence of small patches with grain sizes up to grain size Noo 39 ASTM Designation E 19o 4 A rather general graphitization was observed on both the inside and outside wall surfaces of both connectionso The graphite was quite fine and appeared to be concentrated in the grain boundaries0 The microstructures beneath the stainless-steel buttons had the following characteristics ~ lo Segregated graphite approaching "chain"' graphite was present under all buttons. In most instances, the chain graphite was either parallel to the surface of the pipe and approximately half-way between the inside wall and the heat affected zone of weld or progressed diagonally in the wall near one end of the button0 2 Only a few isolated nodules of graphite were observed in the heat-affected zone of the welded stainless-steel buttonso The microstructure of the material reheat-treated for impact tests revealed the following. 1 The structure of the reheat-treated material was similar to that of the pipe as removed from service except that there was somewhat less spheroidization of the carbides 2 The massive carbides observed in the grain boundaries of the pipes also were present in the reheat-treated material0 It was concluded from the metallographic examination that where the metal was not influenced by the stainless-steel measuring buttons, little graphitization had occurredo There also seems to be little question, but that the original carbide lamellae in the pearlite had spheroidized slightly. The presence of what appears to be massive carbides in the grain boundaries was somewhat surprising in view of the relatively slight spheroidization of the pearlite and raises some doubt as to whether they are actually carbideso The slight general precipitation in the matrix was characteristic of o050 Mo steel after prolonged exposure to stress and temperatureo Other than graphitization, the slight spheroidization, massive carbides, and characteristic general precipitation appear to have been the major structural changes during service0 The resistance of the "massive carbides" to removal by heat treatment was interesting and indicates an unusual composition if they are actually carbideso DISCUSSION The creep curves, Figo 2, obtained from the measurements during service are a very unusual set of datao The engineer or metallurgist rarely has an opportunity to evaluate design considerations and laboratory data in terms of actual creep in service.

-9 -The exact creep and rupture properties of the particular pipe material at the time it was placed in service were not establishedo There are, in fact, very few laboratory tests which establish stresses for creep rates as low as those measured in service (less than 0o002 per cent per 1000 hr, Table II)o Creep strengths of C Mo steel are rather variable, depending on heat-treatment and melting practice0 The exact creep rate for the service stress of 7500 psi at 900 F is therefore not well establishedo Extrapolation of the available laboratory creep data at higher stresses back to 7500 psi, however, suggests that on an average, rates of the order of 0oOOl per cent per 1000 hr might be expected (see Figo 5), The observed creep in service, therefore, is considered to be in good agreement with the predictions of laboratory dataO The creep rates measured in service and those measured in the laboratory for tests at 7500 psi on coupons cut from the pipes after they were removed from service are in good agreemento This confirms the service creep measurementso It would be expected that, after a brief period of adjustment, the two creep rates ought to agree. The stress for a creep rate of Oo01 per cent per 1000 hr established in the laboratory, 12,500 psi, and the stress for rupture hours, 29,000 psi, turned out to be in exact agreement with average values for new steelo The observed creep strength is the average value found by Miller and Heger (3) and used by the Subgroup on Allowable Stresses for Ferrous Materials of the ASME Boiler Code Committee in setting the present allowable stress of 12,500 psio The same group also developed an average value for rupture strength of 29,000 psio The average values for strength of the pipe material after prolonged service may appear to be somewhat surprisingo Actually, however, this is to be expected if it is assumed that the pipe metal had nearly average properties initiallyo It is generally considered that exposure to stress and temperature will permanently use up the available life of metals by creepo Consideration of the service conditions, however, indicates that this should have been very smallo If the stress-rupture time curve of Fig. 4 was extrapolated to 7500 psi, the indicated rupture time would be many millions of hourso On this basis, 100,000 hr of service was negligible in comparison to the total life availableo Certainly the percentage of life used on this basis was too small to alter significantly the rupture-test results. The laboratory creep data show no evidence of weakening from the 100,000 hr of serviceo As previously discussed, the creep rates were the same as those existing during serviceo Creep rates from laboratory tests, however, would be expected to show evidence of structural deterioration only if so-called third-stage creep had started during serviceo There is no precedent of which the authors are aware for estimating when third-stage creep should become evident in such prolonged time periods and at such low creep rates as existed in service. General laboratory experience indicates that, in the absence of substantial loss of strength by structural alteration due to temperature, third-stage creep would not be expected before the total creep deformation reached 1 per cento The creep deformation during service, Figo 2, was of the order of 0o2 per cent and the probability is that the onset of third-stage creep was remoteo The test results support this viewo The rates observed were average for the steel and there was no indication of increasing creep rates during the tests or in service at either 7500 or 12,500 psio The possibility exists that structural alteration under the influence of temperature and stress during the prolonged service could have altered substantially the creep properties from those characteristic of the new pipeo Unfortunately creep data

-10 -for the particular pipe samples before service are not available~ The data obtained from the tests on the pipe after service are compared in Figo 6 with those presented by Weaver (4) in which he attempted to-estimate the effect of spheroidization during service. At OoOl per cent per 1000 hr, the creep strength was in between that for the annealed condition and for a condition of spheroidization equivalent to 23 years of service at 900 Fo Actually the strength indicated is almost exactly that which would be estimated from Weaver's datao The 100,000 hr of service represents about one-half the 23 years estimated by Weaver to attain the degree of spheroidization of his test bars and the creep strength of the pipe material is also about half-way between his two conditionso The mierostructure also showed about the degree of spheroidization consistent with the creep strength in comparison to Weavergs datao The creep strength of the pipe at 7500 psi in both service and laboratory tests did not agree as well with Weavergs datao Both showed a stress for a creep rate of Oo001 per cent per 1000 hr of about 7500 psi or a little lesso This corresponded to Weaver's material spheroidized for an equivalent of 23 years of service, rather than to his annealed stocks which showed a stress of 11,600 psi for 0oOO1 per cent per 1000 hro Because the pipe showed creep rates fran early in its service life of the order of OoOOl per cent per 1000 hr, the indications are that the pipe material had different stress - creep rate relations than Weaver's test stock, Weaver annealed his stock from a lower temperature than was used for the Detroit Edison pipe. White and Crocker (5) reported creep data at 925 F for C Mo pipe given the same heattreatment but from different heats as the pipe used for the service creep tests Extrapolation of their data to 900 F indicated strengths of 11,000 and 12,500 psi for Oo01 per cent per 1000 hro In view of such variation, there seems sufficient precedent to assume that the service creep data reflect initial properties more than changes during service, The similarity of creep and rupture properties of tangential and longitudinal specimens indicated that there was no great difference from this source in the properties of the pipe after service, Because creep in service was circumferential, it could be expected that any damage effect might be most evident in the tangential specimens. The absence of any difference between the longitudinal and circumferential specimens then could be further support to the indication that the 100,000 hr of creep in service had used. up only a small fraction of the available service life of the pipe. Without supporting data on the probable relative strengths of longitudinal and tangential specimens for new pipe or for pipe subjected to creep, there is uncertainty in these conclusionso The elongation and reduction-of-area values for the rupture-test specimens were not as low as have been observed for C Mo Steelo The change from 40 to 20 per cent elongation as the time for fracture increased to 2000 hr is not at all unusual, Certainly there is nothing in these values to suggest undue deterioration during the 100,000 hr of serviceo The impact values after service were lowo However, impact values of 10 ft-lb are not at all unusual for C Mo steel, particularly when heat-treated at a relatively high temperature, as was the pipeo The difference in relation to the samples reheat-treated after service does not represent an unexpected change, The absence of any difference between longitudinal and tangential specimens again supports the view that creep damage during service was very slight, The tense tensile properties after service possibly show same evidence of structural change during service, There is same uncertainty in this observation, because the

-11 -comparative data for new material are inadequate. In either event, the properties were satisfactory and the changes were no more than might be expected for 100,000 hr of exposure to 900 F under stresso SUMMARY AND CONCLUSIONS A unique set of creep curves is presented for the creep of two C Mo steel steam pipes during 100,000 hr of service at 900 Fo Analysis of the data shows that the observed creep rates of 0~002 per cent per 1000 hr or less under an operating stress of 7500 psi were in accordance with the predictions of laboratory creep data. Laboratory creep and rupture tests were carried out on coupons cut from the pipe after the 100,000 hr of serviceo The creep rates under 7500 psi agreed with the rates observed in serviceo The creep strength of 12,500 psi for a rate of Oo01 per cent per 1000 hr and the stress for rupture in 100,000 hr of 29,000 psi are the same as the average values for new C Mo steelo Analysis of the service conditions indicated that 100,000 hr of service at 900 F under 7500 psi would have been expected to use up only a negligible amount of the available creep- rupture life of the steelo All laboratory test results support this conclusiono Deterioration of creep strength due to structural changes, such as spheroidization, was certainly no greater than might have been expectedo The general conclusion is that the pipes performed in service to a remarkable degree in accordance with the predictions of laboratory datao

2653 '54 }2 - BIBLIOGRAPHY 1 "Properties of Carbon and Seamless Alloy Steel-Tubing for High Temperature High-Pressure Service," The Babcock and Wilcox Company, Technical Bulletin, No. 6E, 1948. 2 "Digest of Steels for High Temperature Service," Timken Roller Bearing Company, Steel and Tube Division, 1946o 3 "Report on the Strength of Wrought Steels at Elevated Temperatures," by Ro Fo Miller and Jo Jo Heger, American Society for Testing Materials, Special Technical Bulletin No, 100, 19500 4 "The Effect of Carbide Spheroidization upon the Creep Strength of CarbonMolybdenum Steel," by So Ho Weaver, Proco American Society for Testing Materials, volo 41, 1941, po 6080 5 "Effect of Grain Size and Structure on Carbon-Molybdenum Steel Pipe," by Ao Bo White and Sabin Crockery Trans, ASME, volo 63, 1941, ppo 749-764o Captions for Illustrations Fig. 1 Sketch showing location of stainless-steel reference buttons used for axial length and diameter measurements during service of C-MO pipe Fig, 2 Change in circumference of 10ino C-MO pipe during 100,000 hr of service at 900 F under an operating stress of 7500 psi Figo 3 Stress-creep rate data for Oo0 MO steel steam pipe after 100,000 hr service at 900 F Figo 4 Stress rupture-time curve for Oo5 steel steam pipe after 100,000 hr service at 900 F Fig. 5 Creep data at 900 F from C=MO pipe of Delray Station, Detroit Edison Company, after 100,000 hr of service compared with data for C-MO steel reported by The Timken Roller Bearing Company (2) and the Babcock-Wilcox Tube Company (1) Figo 6 Creep data at 900 F for C-MO pipe after 100,000 hr of service compared with data for spheroidized C-MO steel, as reported by Weaver (4)

TABLE II TABLE I DIIAMETRAL MEASURFIMETS CF 1lO-INO SCHEDULE 80 CARBON MOLDIADEtL PIPE DAUJRIMNT TR SCAHULERS F SERVICE CALCUIATED CREEP RATES OF SCHEDULE 30 CAR1ON-DLYBDENUi August March July June August PIPE DURING 100,000 HOURS SERVICE AT 900 F UNDER A Novembor (12,789 (26,937 (43,832 (66,987 (100,135 Cold relaxed position Looa- hr.) hr.) hr) hr). hrs) Tot. Change after removal/lin. NOMIL STRESS OF 7,500 PSI tion 1938 1940 1942 1944 1947 1951 5 Periods (100,135 hra) DIAMETRAL OVER STAINLESS STEEL REFERENCE BUTTONS (INCHES)* Circumferential Creep iRates Duration Accumulated Time %/1l00 hours North Position Period (hours) (hours) North Connection South Connection 1-1 11.2876 11.2921 11.2933 11.2948 11.2984 11.3046 +0.0170 11.3052 2-2 11.2756 11.2791 11.2808 11.2818 11.2844 11.2887 +0.0131 11.2912 First 12,789 0 to 12,789 0.0031 0.00 3-3 11.3161 11.3201 11.3213 11.3218 11.3244 11.3277 +0.0116 11.3292 4-4 11.3536 11.3581 11.3608 11.3608 11.36A4 11.3734 +0.0198 11.3740 becond 14,148 12,789 to 26,937 0,00165 0.00220 Third 16,895 26,937 to 43,832 O.00020 -0.0015 South Position Fourth 23,155 43,332 to 66,987 0.00140 0.00130 1-1 11.2304 11.2455 11.2512 11.2548 11.2593 11.2945 +0.0641 11.2732 Fifth 33,148 66,987 to 100,135 0.0165 0 7 ( 2-2 11.6044 11.3103 11.3137 11.3153 11.3193 11.3305 +O.02G 11.3332 Ffth 33 66,987 to 100,170 3-3 11.4924 11.4873 11.4852 11.4798 11.4783 11.4785 -0.0139 11.4812 4-4 11.3794 11.3813 11.3812 11.3808 11.3823 11.3865 +0.0071 11.3892 INSIDE DIAMFTRS AFTER SECTIONING (INCHES) (a) Cutting out welds and reweldinc may have influenced measurement and creep Between Between Between Between Iuri:lg the fifth period. 1-1 1-2 2-2 2-3 3-3 3-4 4-4 4-1 North Connection 9.665 9.666 9.665 9.659 9.666 9.646 9.646 9.657 9.666 9.654 9.665 9.658 9.661 9.645 9.633 9.653 TABLE III South Connection 9.564 9.575 9.598 9.651 9.683 9.678 9.648 9.612 --- 9.525 9.539 9.611 9.673 9.697 9.676 9.641 9.583 TENSILE PROPERTIES AT ROC( TEMPERATURE FROM THE NORTH CONNECTION C-Mo PIPE AFTER 100,000 CIRCUMFERENTIAL LENGTH COVERED BY REFERENCE POINTS (INCHES) HOURS CF SERVICE AT 900 F North Connection -- 5.090 South Connection - 5.075 Tensile Yield* Proportional Elongation Reduction of Specimen Strength Point Limit in 1.5 in. Area Location (psi) (psi) (psi) (%) (%) * All measurements corrected to 68- 2BT 53,000 35,300 27,500 43,3 71.5 2BB 53,400 32,600 - - 42.0 71.8 4BT 53,400 32,100 27,000 41.3 71.8 4BB 53,900 38,000 32,500 42.0 71.0 * Specimens showed a very sharp yield point so that this value also is the same as the offset-yield strength values.

TABLE IV TABL CHARPY V-NOTCH IMPACT PROPERTIES AT 80 F FOR CREEP TEST DATA AT 90C 'F FCOR 0.50 io STEEL PIPE AFTER PIPE METAL FRaC THE NORTH CONNECTION 100O00O HOURS OF SERVICE Stress Creep Rate - %/1000 hours Section Direction at indicated time (psi) 1000-hours 2500-hours Impact Strength -_ - As Removed From Service (ft - lb)orth Longtudi 7,500 0.002,~~Tangential 7, 8~ SNoruth Longitudinal 7,500 0.002 Sorth Tangential 7,500 (1) -- South Longi tldinal 7,500 0.002 -- T'ngentisl 7, 8 South Tangential 7, 500 (1) Longitudinal 11, 9.5 North Longitudinal 12,500 0.012 0.010 North Tangential 12,500 0.013 0.010 Longitudinal 8, 9 Heated 2 Hours at 1900 F, Furnace Cooled Tangential 19, 22.5 (1) These specimens showed a decrease in length during the first few Tangential 20, 23 hours of testing and no measurable creep thereafter out to 1000 hours. Longitudinal 18, 20.5 Longitudinal 20, 22 TABLE VI RTTPIPRE-TEST DATA AT 900 F FOR 0.50 MO STEEL PIPE AFTER 100,000 H0URS OF SERVICE Note (1): Specimens were held 16 hours at 200 F after machining. Before testing the specimens were equalized to a temperature of 80 F by holding in a water bath at 80 F. Rupture Elongation Reduction Specimen Location Stress Time of Area _psi) (hours) (% in 1.5 in.) (%) North Conneotion 2 CB - longitudinal 42,700 S.T.T.T. 38.7 76.8 2 CT - longitudinal 40,000 3.5 44.0 76.2 2 CB - longitudinal 38,000 775 30.0 39.6 2 CT - longitudinal 36, 000 1803 21,3 26.2 NA2 B3 - tangential 37,000 1482 25.0* 31.4* South Connection SA3 L2 - longitudinal 37,000 1558 20.0 26.1 SA R1 - tangential 37,000 1349 23.0* 29.8e Tangential specimens were 0.250-inch in diameter with a 1-inch gage length.

REFERENCE BUTTONS. NoR —H NOR1H- LONGiTUDINAL FOR AXIAL MEASUREMENT __ - ~ - NORTH - TANGENTIAL 0 - SOUTH - LONGITUDINAL U - SOUTH - TANGENTIAL _ ( > 1 30,000 20,000 ~. 3 __- - REFERENCE BUTTONS FOR.,- '_______ __ DIAMETER MEASUREMENTS o. > > \0 I I I I I I I 900 F TESTS 03 6,000 ~~~~3 X 3 5,000 4,000 0.0001 0001 0.01 CREEP RATE - % / 1000 HOURS Fig.3 Fig.1 ' 0.003 __ -_____ -_____- -- NORTH LEAD -LONGITUDINAL - - =z 0.003' - - - - I I i~ 111 1 11 __*__$-^-NORTHLEAD —TANGENTIAL L__^.~LNEAD - TANGENTIAL I I i 1111111I — SOUTH LEAD-LONGITUDINAL! I) UII_!_ 111111_ I _I -SOUTH LEAD-TANGENTIAL * NORTH CONNECTION Z o SOUTH CONNECTION / 60I0I0i F 23%l20 w /0.002 -_____._!_ _ ______I-JI!! l!!E50,00110 o 0.002 J. -140,000 900 F ZZ~~~~~~~~~ ~ ~ ~~ - 30,000 3 44%ELONG1. 3 w // D // 0 0.001 4,^10 100 1000 10,000 00,000 W{^~~~~~~~~~~~~~~~~~~~~~~~~ ^~~~~~~~~RUPTURE TIME, HOURS -- I I I 11 f I Fig.4 0 20,000 40,000 60,000 80,000 100,000 TIME- HOURS Fig.2

X NORTH - LONGITUDINAL * NORTH - TANGENTIAL O SOUTH- LONGITUDINAL __ - SOUTH-TANGENTIAL - 30,000 - - 900 F TESTS - - - ~20)0O _1 _ 1 _. _1 _1 __' S,tLG __ 20,ooo00- - MILLER-HEGER RANGE - O L _ _ i l - rl EI 5 < I 1 1 T 1 T O -0,000 --- 6,000 5,ooo0 - - - 4,000 - _ _ 0.001 0.01 0.1 CREEP RATE -% / 1000 HOURS Fig.5 X NORTH SECTION - LONGITUDINAL SPECIMEN * NORTH SECTION - TANGENTIAL SPECIMEN O SOUTH SECTION - LONGITUDINAL SPECIMEN - SOUTH SECTION - TANGENTIAL SPECIMEN 30,000 - 900 F TESTS 20,. ( —000 1 ---- I- I I I I I T | ---- -- -- - - | | | *WEAVER DATA 0MILLER-HEGER RANGE 1F HR. -F.. ~15,OO --------- " "~ -,1750 F -8 HR.-F. (C.+) 15 0001 | l l i i { | | | (2) 1500 HR. AT 1292 F EOUIVALENT TO 200 YR. AT 900 F 4,000 - - 1 1 ----- II 1) 100 HR. AT 1292 F EQUIVALENT TO 20 YR. AT 900 0.001 0.01 0.1 CREEP RATE- / 1000 HOURS Fig.6