Creep rupture test were conducted at 538°C (1000°F), 565°C (1050°F) and 593°C (1100°F) on the 1Cr-11/4 Mo-1/4V steel which is used as the high- and medium-pressure turbine rotor shaft up to 1100mm diameter. Specimens were taken out of inner and outer part of an actual rotor, having radial, tengential and longitudinal direction, respectively. Test results are summarized as follows: (1) The longitudinal specimen has about 10% higher creep rupture strength than radial specimen which has higher strength than tangential specimen also by about 10% in terms of 1000 to 10000 hour creep rupture stress at each temperature. (2) Specimens from outer part of rotor show lower rupture strength than specimens from inner part. Though both inner and outer part have the bainite microscopic structure, the inner part has the upper bainite structure and the outer part has the lower bainite structure, since the cooling rate at quenching is slower in the inner part. The present result which shows that the upper bainite has better creep rupture strength than the lower bainite is in conformity with the previous results on small specimens. (3) The creep rupture stress of the notched specimen is higher than that of the smooth specimen in a short time range of testing and lower in a long time range. This intersecting of the rupture stress vs. time-curve for the notched and the curve for the smooth occurs at a shorter time of testing and when the testing temperature is made higher. (4) Comparing with the results of the tests by Werner et al.2), Conrad et al.4) and Rankin et al.5) on the 1 Cr-11/4 Mo-1/4V steel turbine rotor shaft, the present result shows nearly similar creep rupture strength to their results, and speaking at some length, the present data show a little lower strength in short time range and a little higher strength in long time range than theirs. (5) Elongation at creep rupture in the present test is similar to or a little higher than Rankin's. The specimen from inner part of a rotor gives longer rupture time but shorter elongation at rupture than the specimen from outer part.
We have carried out creep rupture tests and high temperature fatigue tests for notched bars, to examine the general effect of notches on high temperature strength over prolonged periods of time. Test materials are ten kinds of heat resisting steels, ranging from low alloy steels to super alloys, and we have given them regular heat treatment in these tests. Notched specimens have U-type and V-type circumferential groove by machine cutting, which are respectively about 1.8 and 3 in stress concentration factor. The creep rupture tests of multiple tension type were carried out aiming at 1000 hours, and fatigue tests of rotary-bending type was practised at a revolving speed of 3000rpm up to N=107 cycles. The typical values of the abave tests are tabulated in this report, with notch rupture strength ratio and strength reduction factor (β). Here, we obtained several conclusions as follows; (1) The notch rupture strength ratio is generally larger than unity, as in the short time tensile test. But, it correlates closely with ductility variation of materials, depending on progressive precipitation and aging, with the increase of applied temperature and time. (2) Within the range of the tests, all the test points of ten materials have about the same correlation for the notch rupture strength ratio to the value of plain bar contraction, at corresponding rupture time, and it falls under unity when the plain bar is contracted below approximately 30% in V-notch, and 20% in U-notch. It is convenient to assume notch effect from plain bar creep rupture tests. (3) This ratio has a tendency to decrease after it has once increased, with the increase of notch sharpness. When it has large ductility the maximum value is high, but with small ductility, it gradually falls. (4) In general, the ratio for austenitic alloys is lower than for ferritic alloys. Particularly, 25-20 stainless steels showed the highest notch sensitivity of all the test materials. (5) Concerning the high temperature fatigue, there was a tendency in the strength reduction factor (β) for ferritic alloys to decrease, and for austenitic alloys to increase at high temperature.
Cr-Ni austenitic stainless steels are widely used for parts of power boiler and equipments of high temperature and high pressure chemical industry. These materials are employed in tubular form or sheet and are jointed together by welding in the construction of equipments. Many investigators, for the need of high temperature strength data about deposited steels, have performed creep rupture tests. However, their strength data are widely scattered even the in same kind of steel. The present work has dealt with the high temperature strength of some Cr-Ni austenitic stainless deposited steels about the effect of microstructure, composition and solution treatment on the creep rupture properties. In this investigation, the following deposited metal compositions are included. (1) Ten kinds of Cr-Ni-Mo (16∼21 Cr, 8∼13 Ni, 1∼4 Mo) (2) 20 Cr-9 Ni-Cb After welding, some of these materials were given solution treatment for 1 hour at 1050°C and 1100°C. The specimen was prepared in such a manner as the direction of applied stress would be parallel to the direction of well-bead deposition. Tension test was carried out at room temperature and in the temperature region from 550°C to 700°C. Creep rupture test was performed at 650°C and 700°C, and rupture strength for 1000 hours was obtained by interpolation and for 10000 hours by extrapolation with log-log plotting. After creep rupture test, their microstructure were examined. The Results obtained are as follows: In general, δ ferrite in Cr-Ni deposited steel is decreased by solution-treatment and their rupture strength, rupture elongation and ductility are increased. δ ferrite in deposited metal change to σ phase during creep rupture testing and creep rupture strength decreases with increasing of δ ferrite. Otherwise, some of low Cr low Ni deposited metal, which locate the unstable austenite region near martensite in A. Schaeffler's diagram, produce martensitic ferrite after long time creep rupture test.
We have examined notch creep rupture properties of several heat-resisting alloys with varying degrees of precipitation hardening. This report is concerned with the effects of heat treatment and of notch sharpness on the creep rupture strength in about 1000 hours of S-816 alloy. Heat treatment of specimens was carried out as follows; (1) solution treatment, (2) solution treatment and normal aging, (3) solution treatment and aging at a higher temperature. In all cases, the heat treatment was given prior to machining to a diameter of 6.35mm (1/4 inch) and to a gauge length of 25mm (about 1 inch) for unnotched specimens. For notched specimens, the diameter at the base of notch is same as that of unnotched specimen. The notch was finished to radius of either 0.06mm or 1.25mm. Notch strengthening was always observed for specimens with the larger radius notch, but it was not always so for specimens with the sharper notch. The case of solution treated notched specimen was extreme, in that it was notch-weakened in shorter time ranges, though the elongation was more than 30 pct.. The reason of this notch-weakening was sought in the aging effect at the base of notch. It was found that the notch-strengthening is to be related more to the reduction of area rather than to the elongation, both taken at the rupture of unnotched specimens test under identical conditions. It was shown that a material is notch-strengthening if its reduction of area value exceeds about 35 pct. as taken of unnotched specimen.
A new technique of photograting method to measure the plastic strain produced at high temperature was proposed by the authors previously. The method is applied here in order to clarify the creep strain state produced in aluminum notched specimen at 200°C. Frequently the creep strength of the notched specimen is higher than the smooth specimen. In fact, the creep rupture time of the notched specimen used here is twice as long as the smooth specimen at the same stress level. In order to precisely and exactly determine the cause of this phenomenon, it is very important and valuable to clarify the state of creep deformation of the notched specimen. The new technique has given us clear interference fringe patterns to measure the strain distributions produced in the notched specimens. From direct observation of the strain distributions, it has become clear that, if the elongation of the specimen is decided within the gauge length, the strain distribution is the same, regardless of the applied stress and the elapsed time. Approximate figures of stress distributions were obtained, using the experimental values of strain. It can be concluded that the average strain in the minimum cross section of a notched specimen produced at a definite elapsed time is less than the strain produced in a smooth specimen when the average applied stress for the notched specimen is the same as the applied stress for the smooth specimen.
Structural members which are subject to load at high temperature are, in general, not in a steady condition of applied stress and temperature. In this paper, as one of fundamental problems, the creep behaior under cyclic stress removal is discussed. If a specimen of metals is creeping under constant load, at high temperature and on the way, the load is removed, elastic strain and creep strain recovery which seems like elastic-after-effect occur. After the period of stress removal, when reloaded, creep recovery which is based on the recovery of work-hardning in each crystal, strain aging and precipitation in commercial materials, take place. When the cyceic stress removal is given creep curves, show the differences from steady load creep curve. As the a result of this experiment, creep curves of stress removal are shown upper or lower than steady load creep curves, but the difference between lower creep curve of stress removal and steady load curve is very small. So in this experiment, by considering creep recovery only, creep curve of stress removal can be estimated well. But as for cases when stress cycle or period of stress removal is changed, this result has some questions. As test temperature is rised, creep recovery is much remarkable. If stress amplitude is synchronized with temperature cycle, the influence of creep recovery is much striking. Experimental results may be summarized up as follows: (1) Under a high stress and high temperature, creep recovery is much remarkable. (2) In the eary stage of transient creep, creep recovery is much larger and therafter gradually decreases. In the stage of steady creep, creep recovery reaches almost constant value. (3) Strain increment in cyclic stress removal test is affected not only by creep recovery but also by precipitation, aging and phase change. etc. (4) In this experiment, this strain increment is well expressed by the following equation: Δε=a log (1+Ctr) Δε: strain increment, tr: rest time a, c: constant in constant strain, temperature and stress.
Numerous articles have been published on the multiaxial creep problem since 1935 when bailey presented his extensive theoretical work together with some experimental results on this subject. However, these articles are concerned mainly with theories, and experimental works, so far made, are not sufficient to verify the theories. Especially, the experimental works on the nonsteady stress problem are two few, there being only those conducted by the Johnson and his colleagues. This report deals with the two dimensional dynamic creep of thin-walled cylinder under axial static tension and alternating torsion, as a special case of the nonsteady-state multiaxial creep. The tests were conducted with a low carbon steel at the temperature of 450°C. The static tensile load was applied to the specimen by dead weight through a lever mechanism, while the alternating torsion was given by vibrating the end of a lever, attached to the specimen grip perpendicularly to the axis of the specimen, with the aid of an eccentric driven by an induction motor. The vibrating frequency was 1500 cycles per minute. These loading assemblies were realized by modifying the elevated temperature fatigue testing machine for combined tension and bending, which was used by the authors previously in making the combined stress dynamic creep test under static tension and alternating bending. Heating of the specimen was carried out by a conventional heating system consisting of an electric furnace and an automatic temperature controller. The axial elongation of the specimen was measured by dial gauges. The results of the tests revealed that the material employed shows similar creep curves under combined static tension and alternating torsion as those in the case of simple tension creep tests. The magnitude of creep under a combined stress, however, was considerably greater than under the static tensile stress that is equal in magnitude to the static tensile component of the combined stress. These experimental results were discussed from the standpoint of the multiaxial creep theory, and a fairly good agreement was obtained between theory and experiment. In this discussion, the multiaxial creep theories based on von Mises and Tresca criterion for the effective stress were first taken up by using the strain hardening criterion for the stress-strain relation, but they failed to agree with the experimental results. As the second step, the authors used an effective stress, derived from the information of multiaxial static creep tests made with the same material under the combined tension and torsion, and obtained a fairly good agreement between theory and experiment for dynamic creep. Finally, the effect of alternating torsional creep strain on strain hardening was also discussed, and it was shown that a better agreement between theory and experiment can be obtained by taking this effect into account. However, this effect was rather small and may be neglected for practical purpose.
The authors have investigated two fundamental experiments on the deformation behaviours of the aluminium sheath which might be restricted by the uranium fuel core, for exsample, as the combination of the uranium rod and the aluminium sheath used for the fuel element of JRR-3 type reactor. The first experiment was studied for the stress behaviours and the micro structure changes etc. of extruded 2S aluminium which was subjected to the same length changes as that of β-quenched uranium under the thermal cycling between 110° and 550°C. In this case, the experiment was done under the condition separated the uranium rod and the aluminium sheath from each other. The second one was studied to clarify the systematic deformation behaviour of the aluminium caused by the thermal growth of the uranium for reference to the first results, and then the experiments carried out for the repeated strain cyclic tests with the growth of constant rate of five kinds in the tensile direction only, that is, 0, 0.020, 0.037, 0.053 and 0.074 per cent per cycle, respectively. In these experiments, the strain cyclic amplitude, the test temperature of 2S aluminium and time per cycle was about 0.59 per cent strain, 75°C and 27.5min., respectively. The testing machine used was“the universal thermal fatigue testing machine”which was designed by the authors. The outline of the experimental results was as follows: The effect of the growth was clarified to accelerate the local deformation of 2S aluminium from these results, and consequently the fatigue life decreases with increase of the growth rate per cycle, and the relation between the growth rate per cycle and the logarithm of cycles to failure was approximately presented by a straight line. The degree of softening and hardening to be caused by strain cycles with the growth and the time for softening and hardening depended upon large or small of the growth rate per cycle. Furthermore, it was defined from the point of view of the microstructure behaviour that the changes of microstructure due to progress of the deformation follow the same process in spite of large or small of the growth rate per cycle, but the progress rate of the microstructre change becomes late with decrease of the growth rate per cycle. That is, it dependens apparently upon the number of various deformation bands and their forms. The experiments showed that the growth and the strain cycle have the reversal effect on the changes in strength each other, and this opposite phenomenon was explained from the point of view of the microstructure changes between the repeated strain cycle and the simple tension.
Leading edge cracks on blading of aircraft jet-engines have been found out frequently and systematic investigations are continued in US Air Force and N.G.T.E. in England. Thermal fatigue due to repeated themal stress induced in transient engine condition is the dominant cause of such crackings. To investigate thermal fatigue behaviour, we estimated local heat transfer around blade profile and calculated thermal stress distributions in transient condition. Experiments were carried out under the similar condition to engine start and stop, using model blades, and investigated origination and propagation of such crackings. It is essentially important to investigate heat transfer around blade profile to estimate temperature distribution in blade section. We applied the approximate method shown by E.R.G. Eckert in N.A.C.A. technical report in 1955. Transient temperature distributions were estimated by using Capacitance-Resistance net work simulator and stress distributions were calculated assuming young's modulus and linear expansion coefficient as temperature dependant by using an electronic digital computer. We concluded that cyclic plastic strain must have been occurred in start and stop of gas turbine. 4 differnt materials N-155, S-816, 25 Cr-20 Ni and 18 Cr-8 Ni were investigated in the thermal fatigue test. Model blades were heated rapidly by hot gas stream and then cooled by cold air, and origination and propagation of crackings were investigated. The conclusions of our study for improvement of blade behavior toward thermal fatigue are as follows: (1) Leading edge radius should be as large as possible within the allowable limit of aerodynamics. (2) Air film cooling injecting air from leading edge will be effective. (3) Experiments showed slits to absorb thermal strain at leading edge are effective. (4) Material of higher thermal fatigue resistibility is preferable.
The service life of a hot working roll is limited by fire cracking caused on its surface by repeated thermal stress. A testing apparatus was developed for the purpose of the estimation of fire cracking characteristics of iron and steel. In this study, fire cracking characteristics of an actual slabbing mill roll is investigated and the relation between the crack length and its number and also the effect of maximum heating temperature are studied. The results may be summarized as follows: (1) Fire cracking characteristics of the portion from where a premature breakage of a slabbing mill roll was happened was inferior. The lamellar pearlitic structure was indicated at this portion by microstructual examination. The microstructure of the other part of the roll was globular pearlitic. This difference of microstructure was supposed to be caused by a mis-control during the heat treatment. The unfavorable lamellar pearlitic structure is a cause of inferior fire cracking characteristics. (2) Distribution of fire cracking characteristics was investigated from the surface to interior of a slabbing mill roll. Fire cracking characteristics of the internal portion is inferior because of the increase of lamellar pearlite and segregation of carbon. By a re-heattreatment, the microstructure was improved, but the improvement of the inferiority caused by the segregation was difficult. (3) Effect of a crack on the stress of its neighbourhood diminishes at the portion where is about twenty times of crack depth off from the crack. Consequently the deeper the crack grows, the fewer the cracking. (4) Effect of maximum heating temperature was studied on a Cr-Mo steel between 300 and 700°C. The maximum crack depth was observed at 600°C. At the temperature above 600°C, the constraint upon the specimen surface given by the regidity of specimen itself becomes week and the thermal stress would be lowered.
The strength characteristics of metallic materials which are to be subjected to thermal stress cycling superimposed on mechanical stress is of interest from the practical point of view in design purpose. The authors have pointed out in the previous papers an importance of the consideration of the problem in relation with the strength characteristics of materials in cyclic mechanical strain fatigue and creep rupture at elevated temperature, and have presented some analytical and experimental works in this line. In the present paper, the thermal fatigue combined with steady mechanical stress is treated. The elongation of material is not prohibited and the mechanical mean stress does not decrease with number of cycles of thermal stress. There is another mode of combination of mechanical stress and thermal stress cycling, that is, both of mechanical and thermal stress are cycling. This category of problem has been discussed in another paper of the authors. The interest of the present study is placed on the former problem, and the fracture life and deformation of austenitic stainless steel subjected to cyclic thermal and steady mean stress is discussed in connection with both strength characteristics of fully reversed thermal stress cycling and simple creep rupture. Test condition of combination of cyclic thermal stress amplitude Δσ and steady mechanical stress σm is characterized by the stress ratio Δσ/σm. Thermal stress in hollow cylindrical specimens arose by prohibiting free longitudinal expansion and contraction of the specimen subjected to temperature cycling. The mechanical mean stress in tension is applied by pull head mechanism. In the test of small stress ratio, specimen elongates and it causes relaxation of the mean stress. In order to realize the test condition of thermal fatigue under mean steady stress, the relaxation of mean stress must be compensated. This function is realized by a relay mechanism, in which a pulling head worked by the signal sent from the electric circuit of wire strain gage mounted on a weighing bar that is installed in series with specimen. The lower stress level in the cycling of combined stresses is taken as the basis of the automatic control of combined stresses, and the control unit is set to maintain the level of the minimum stress constant. From the present study the followings are concluded: (1) Fracture life of materials subjected to thermal fatigue combined with tensile mean stress can be predicted approximately from both test data of completely reversed thermal stress cycling and simple creep rupture, by employing the criterions of fatigue damage and life consumption in creep rupture for the material subjected to simultaneous cyclic variation of stress and temperature. The analytical results are in good agreement with experimental results on 18-8 Cb stainless steel where Δσ/σm is chosen as 0.5, 1 and 2. (2) Tensile deformation of materials subjected to combined thermal cycling and tensile mean stress, where tensile mean stress is larger compared with thermal stress amplitude, is estimated from the test data of simple tensile creep, by using the principle of strain hardening for the material subjected to simultaneous cyclic variation of stress and temperature. The analytical results are approximately in agreement with experimental results where Δσ/σm is smaller than unity.
In recent years, the problem of thermal shock of metallic materials has been the subject of many studies, and several researchers have made fundamental works in this line. Surveying the previous works, it is found most of the studies on thermal shock are limited within the experiments on non-metallic materials, and so-called “thermal shock parameter” is defined as a criterion for thermal shock resistance. Thermal shock parameter is analysed from the concept based on maximum principal stress theory and elastic thermal stress therory. In the past study on nonmetallic materials such as ceramic, the thermal shock parameter is adopted as the measure of thermal shock resistance. However, this is true within the limited case of non-metallic materials. For the cases of most metallic materials, there are several questions as to criterion with which thermal shock resistance is evaluated. Furthermore, there is an open question as to the study on transition phenomenon from thermal shock fracture to thermal fatigue fracture. In this connection, it is necessary to establish the relation between thermal shock strength chracteristics and thermal fatigue strength characteristics. The authors, as well, have been interested in the problem concerned with thermal shock, and have done some works on the relation between tensile strength characteristics and thermal fatigue strength characteristics by employing heat-resisting metallic materials. In the present paper, the results of thermal shock tests for different heat-transfer conditions are presented, and the results are compared with those of thermal fatigue tests. Two types of tests of thermal shock and thermal fatigue are performed by solid cylindrical specimen of cast iron (3.79% C). Two sorts of heat-transfer conditions are realized by using air cooling and air mist cooling. Both tests are made by employing a conventional thermal fatigue test apparatus, and both ends of specimen are constrained at upper temperature level in thermal cycling. Maximum temperature level is chosen as 600°C. The results of the present study are summarized as follows: (1) Cast iron tested behaves transition phenomenon from thermal shock fracture to thermal fatigue fracture. Under air cooling condition (heat-transfer condition β=0.035), the material fractures at the begining of cooling path in the first cycling of temperature, when the upper temperature level is 500°C. If the upper level of temperature cycling is lower than 500°C cycle-dependent fracture occurs. On the other hand, under air-mist cooling condition (β=0.15), thermal shock fracture occurs for the case of the upper temperature level of 400°C. The parameter β is a non-dimensional heat-transfer coefficient and is denoted as β=r0h/k where r0 is outer radius of the specimen, h heat-transfer coefficient and k heat conductivity. The influence of heat-transfer condition on the temperature level under which such a transitional phenomenon occurs is related with the effect of increase in axial thermal stress at the outer surface of specimen due to multiaxial thermal stresses. In this case, the influence of strain rate on breaking strength of the material is not predominant. (2) From the results of thermal fatigue tests of which were carried out by taking the upper temperature level, the thermal strain amplitude and the thermal stress amplitude as variables, it is found that number of cycles to fracture is shorter in the case of air mist cooling condition than the case of air cooling condition. This trend is interpreted by taking account of plastic deformation produced in the material during a cycle of thermal strain cycling. (3) For the transition phenomenon of cast iron tested from thermal shock fracture to thermal fatigue fracture, simple maximum principal stress theory is applicable.