材料
Online ISSN : 1880-7488
Print ISSN : 0514-5163
ISSN-L : 0514-5163
17 巻 , 173 号
選択された号の論文の17件中1~17を表示しています
  • 中沢 一
    1968 年 17 巻 173 号 p. 84-88
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    A report is given at the outset of a modified rotating-bending fatigue testing machine by means of which the specimens can be submitted to programmed alternating stress cycles. The test material is mild steel with 0.47% C which was annealed in vacuum after machining. The results of the tests are reported, of the plain specimens and of four sorts of notched specimens, and the accumulation of the fatigue damage is discussed.
    It is shown that with the plain specimens and with the blunt notched specimens Miner's hypothesis concerning “linear damage law” that the stress cycles are negligible so far as their amplitude is less than the fatigue limit of the specimen in question could agree with the experimental results. But with the specimens with notches the value of whose stress concentrations factor is larger than 1.8 their fatigue life is found in estimation remarkably critical under “the linear damage law”.
  • 中村 宏, 田中 真一, 初野 耕三
    1968 年 17 巻 173 号 p. 89-96
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    With the increase in the speed of railway trains higher equipment is demanded for the safety of the axles of the rolling stock against fatigue strength. The crack initiation of fracture on the axles usually occur in press fit parts, and the ratio of the crack initiation in these parts have formed over 90% of the total. It is necessary then to investigate the fatigue strength on the press fit part, i.e. the wheel-seat, the dust guard seat etc. in order to study fatigue life of the axles. But in estimating the fatigue life of the axles the usual fatigue tests on specimens are inadequate for data since the press fit parts are subject to the effect of stress concentration and fretting, and there are not sufficient data concerning fatigue strength on the press fit parts.
    In this monograph, the current researches in practice at the present time on the strength of the axles are reported, and the problems concerning the estimation of fatigue life of the axles, the initiation and propagation of crack and the application of cumulative damage law are discussed. To solve the problems concerning the fatigue strength of the axles, researches are required on the following items.
    (1) The estimation of fatigue damage due to service load.
    (2) The effect of the load variation in the initiation and propagation of crack in the press fit parts.
    (3) The fatigue strength under superposed mini sinusodial stress wave and shock stress wave.
    (4) The fatigue strength under the fretting phenomena.
    (5) The fatigue tests repeated 108∼109 times.
  • 河本 実, 伊吹 幸彦, 石川 浩
    1968 年 17 巻 173 号 p. 97-102
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    The manifold program loading tests repeated in five stress levels including those in plastic range were carrid out, of S35C steel by rotating bending at the frequency 20rpm at the large stress range and 1790rpm in the ordinary fatigue tests range.
    In this fatigue test Kawamoto's rotating beam bending fatigue testing machine was used, which was of uniform moment type, and the specimen could be bent in any direction. During the fatigue tests the change of deflexion from vertical to horizontal and vice versa at the center of the specimen was observed.
    The main conclusion is as follows.
    (1) The general theory that fatigue fracture occurs at Σn/N=1 is not applicable to the experimental results. When the minimum step stress is relatively small the value of Σn/N is very small.
    (2) The ratio of the horizontal and the vertical deflexion of the specimen δnv is nearly 1 under large stress loading.
    (3) The loading under stress will make damage on the fatigue life under the program loading.
  • 河本 実, 柴田 俊忍, 石川 浩
    1968 年 17 巻 173 号 p. 103-107
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    The annealed S35C carbon steel was tested under the four-step-in-stress bending program loading. The effect of the stress sequence on the fatigue lives was investigated with respect to the effects by the initial stress level and the magnitude of stress levels as a whole. The number of cycles in one program block was n0=3600 cycles. The results show:
    (1) The effects of the initial stress level and the stress sequence on the fatigue lives can not be seen, because the value of n0 in this experiment is very small compared with the total number of cycles to failure.
    (2) When the maximum stress σ4 is large and all the stresses are overstresses, the fatigue lives nearly coincide with Nth's which are calculated by the use of the linear damage law. When σ4 becomes small, the fatigue life Nf also becomes smaller than Nth, and Nf reaches the smallest value at certain stress level. When the all stresses except σ4 are understresses, the fatigue life Nf becomes equal to or larger than Nth.
  • 関 護雄, 藤谷 景三, 田中 道七, 石井 啓道
    1968 年 17 巻 173 号 p. 108-115
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    An electro-dynamic type fatigue testing equipment has been developed for the study of the fatigue strength under random and program loads. The equipment consists of four basic parts; the signal generator, the power amplifier, the power transducer, and the measuring instruments. The available range of the test frequency is from 2cps to 2kc, the maximum power being 360VA, and the maximum oscillating force which naturally depends on the test frequency and the flexibility of the specimen, being approximately 45kg at 60cps.
    The program load fatigue tests were made on notched specimens of 7075-T6 aluminum alloy, in which the amplitude of the alternating stress was controlled along the curve that was similar to the gust load on transport aircrafts. It was observed that the linear damage rule held good when the root-mean-square value of peak stress was higher than the conventional endurance limit, but that the fatigue life became shorter than that of the linear estimate when the root-mean-square value was near or lower than the endurance limit. This fact seems to indicate that the alternating stress whose magnitude is less than the endurance limit contributes to the progress of the fatigue damage.
  • 山田 敏郎, 北川 茂
    1968 年 17 巻 173 号 p. 116-122
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    Bending fatigue tests of metals were carried out under two kinds of complex stress patterns; the one was the two superimposed sinusoidal cyclic loadings, and the other the rectangular cyclic loadings superimposed with sinusoidal ones. The former loadings were applied with a Schenk type fatigue test machine, and the latter with a electro-magnetic moving coil actuater. The materials used were carbon steel (S35C) and alminum alloy. Those test results showed that if the secondary stress amplitudes (fast ones) were comparatively large and n0 (total number of secondary ones in a period of primary one) was smaller, the fatigue life of experimental results agrees with the calculated values which are estimated with the authors' formula and the peak pair counting method. It was also shown that if the secondary stress amplitudes were small and n0 was large, the test life came out longer than the calculated values, and approached the values which were determined only by the maximum amplitude. Only for alminum alloy with two superimposed sinusoidal cyclic loading, the test life proved shorter than that which was determined only by the maximum amplitude, as far as n0 was small (about 4) and the secondary stress amplitude smaller.
  • 川田 雄一, 三沢 啓志
    1968 年 17 巻 173 号 p. 123-127
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    In the design of machines the fatigue life of a member under service load is usually calculated on the assumption that Miner's rule can be applied. However according to the results of many program fatigue tests, the linear calculative cycle ratio is not always equal to 1, and the Miner's rule does not take into account the stress cycles under fatigue limit. But the results of fatigue tests, including the stress amplitudes below fatigue limit, obviously show that they influence the cumulative life. This fact is particulary important in the ordinary mechanical structure, because the load spectrum mostly consists of stress amplitude less than the fatigue limit, and the stress amplitude over the fatigue limit is smaller in number. Therefore simple application of Miner's rule is somewhat questionable.
    The authors have performed the program fatigue tests of carbon steel S35C under the load spectra which have the same pattern of distribution as the frequency diagram measured on the axles of railways. The load spectra are with or without stress amplitude less than fatigue limit to investigate the influence of the stress below the fatigue limit to the cumulative life. From the experimental results a new equation is proposed to estimate the cumulative fatigue life with the effect of the stress amplitude less than the fatigue limit. The total number of cycles of stresses to failure for fluctuating stress can be shown as follows:
    Ng=N1/{σmi=1ii1)c]+σrp=m+1pp1)b×c]}
    in which n1, n2, ……, nk are the numbers of cycles at stress amplitude
    σ1, σ2, ……, σk in each program cycle.
    If n1+n2+……+nk=ng, then α1=n1/ng, α2=n2/ng, ……, αk=nk/ng
    the value c is decided by the equation c=K, K and b is a constant.
  • 砂本 大造, 遠藤 忠良
    1968 年 17 巻 173 号 p. 128-134
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
  • 菊川 真, 大路 清嗣, 城野 政弘
    1968 年 17 巻 173 号 p. 135-144
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    A push-pull random load fatigue testing apparatus of electro-dynamic type was newly designed and constructed. In order to conduct random load fatigue tests of both narrow and fairly wide bandwidth, the apparatus was required to keep its frequency response flat from 20cps to 175cps by means of a negative feedback of velocity of the moving coil. The stress on the test specimen was measured by using an extensometer of a semi-conductor strain gage type or of a capacity type. The plastic strain of the test specimen was determined by subtracting the elastic strain from the total strain. The stress and the plastic strain were observed by an oscilloscope, and were counted on line by level crossing counters throughout the random load fatigue test. The drift of average stress amplitude for a long period was controlled by comparing an average voltage of a rectified signal of the load cell output with a pre-setted reference voltage.
    The fatigue tests were conducted with unnotched specimens of 0.2 percent carbon steel under the condition of random loading of a narrow band, in addition to the constant stress amplitude test. It was found that a linear cumulative damage assumption expressed in terms of the plastic strain range, gave a good approximation for life prediction.
  • 中川 隆夫, 新田 重光
    1968 年 17 巻 173 号 p. 145-153
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    In the calculation of the fatigue life of metallic materials under the varying stress amplitude, it is common to apply the linear damage criterion, Σ(n/N)=1, based on the stress amplitude (i.e. S-N curve). In the present, the author proposes the fatigue fracture criterion derived from the damage function on the basis of the plastic strain in stead of the customary Σ(n/N)=1. The fatigue fracture criterion is represented as
    1/C0N0εpaadn=1
    for the varying plastic strain amplitude.
    The medium carbon steel S35C (normalized) is tested under the condition of the rotating bending and the repeated tension-compression with constant- and varying-stress amplitude to prove the criterion described above.
    As a result, the experimental values of the cumulative damage based on this criterion
    Df=1/C0mN0εpaadn
    are scattered closely about unity, although Σ(n/N)=1 is also established in this case of low-cycle fatigue: the scatter of Df-value being almost the same as in the value of Σ(n/N). In all the tests performed here, neither effect of the frequency distribution of stress (or strain) nor of their sequence of application on the fatigue life under the varying stress amplitude has been recognized.
  • 小林 英男, 中沢 一
    1968 年 17 巻 173 号 p. 154-159
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    Cumulative fatigue damage of circumferentially cracked 0.35% carbon steel specimens under two step alternating stresses was investigated in rotating bending. Virgin cracked specimens of 0.16mm depth were made of sharp notched specimens by machining away the notch and annealing in vacuum after the fatigue crack was initiated at the notch root. The fatigue limit of the virgin cracked specimens σ1w, based on the net area, was 17.8kg/mm2 (the fatigue strength reduction factor β=1.44). The fatigue crack propagation in the virgin cracked specimens at three first stresses σ1=23.5, 21.0 and 18.5kg/mm2 that were higher than σ1w was examined, and the virgin cracked specimens were subjected to n1 cycles at σ1, a process known as overstressing, until the crack propagated from 0.16 to 0.30mm depth. Then, the numbers of cycles of life remaining n2 and the fatigue limits σ2w of the cracked specimens of 0.30mm depth at various second stresses σ2 were examined. The results obtained are as follows:
    (1) The cumulative cycle ratio Σn/N has values of 1.32∼0.71 when σ21w. Though Σn/N appears to be greater than 1 when σ2>σ1 and less than 1 when σ21, we could say that the linear damage law is nearly substantiated. (2) The fatigue limit σ2w is reduced by the first stressing (overstressing effect) because the crack propagates at the first stress σ1. Though it appears that the higher is σ1, the smaller is the reduction of σ2w, i.e. σ2w=15.6 when σ1=23.5, σ2w=14.6 when σ1=21.0 and σ2w=14.3kg/mm2 when σ1=18.5kg/mm2 for the same 0.30mm crack depth, σ2w is evaluated by means of the crack depth if σ1 is not so high.
  • 遠藤 吉郎, 駒井 謙治郎, 大西 一男
    1968 年 17 巻 173 号 p. 160-168
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    The effects of stress history and corrosive environment on fatigue crack propagation of mild steel are discussed under the constant and the variable stress amplitude, and the following facts are made clear: The crack initiation of corrosion fatigue is very early but the crack rate in saline and in water becomes smaller than that in air from a certain value of crack depth. This may be due to the wedge action of metal oxide in cracks. In hydrochloric acid the decrease of the crack rate is not observed because of the dissolution of metal oxide. Under high-low two-step stress history, both in air and in saline the crack propagation is delayed for a considerable period of time after the stress change and the subsequent crack rate is found to be nearly equal to the expection for the lower stress. On the other hand, under low-high two-step stress history, the rate in air is that which would be expected after the transition to the higher stress, while the rate in saline is a little greater than the expection for a while after the stress change, then gradually decreases to the expection. In saline Σ(n/N) is smaller than unity under low-high stresses, while Σ(n/N) is greater than unity under high-low stresses. Consequently, Σ(n/N)≈1 is concluded in saline under two-stress repeated block history because of the mutual offset of these effects. Since the life of smooth specimens in corrosive environment is wholly occupied by crack propagation, the same argument for the notched specimens mentioned above will be applicable to the corrosion fatigue of smooth specimens under variable stress history.
  • 1968 年 17 巻 173 号 p. 169-177
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
  • 1968 年 17 巻 173 号 p. 178-181
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
  • 遠藤 吉郎, 駒井 謙治郎
    1968 年 17 巻 173 号 p. 182-187
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    The eddy current method was applied to the detection of fatigue crack propagation of notched specimens under rotating bending. The practical device to apply the eddy current method and the test results are described. The depth of fatigue cracks can be measured continuously by the eddy current method, and the accurate value of crack depth is obtained without breaking the test pieces. The method can be applied to ferromagnetic substance as well as to paramagnetic substance, though the detecting probe must be varied so as to suit the materials. A linear relation is obtained between the crack depth and the response of the eddy current circuit irrespective of the environment in which fatigue tests are carried out. The circumferential crack propagation can also be measured by rotating the test piece once or by reciprocating the detecting probe over the notch at several circumferential positions, and the latter method shows better agreement with the crack depth measured by temper coloring. The sensibility of the crack detection is 0.1∼0.05mm. The maximum detectable crack depth is more than 2.5mm in the case of carbon steel (ferromagnetic substance) and is more than 4.0mm in the case of Al alloy (paramagnetic substance).
  • 田中 道七, 出納 真平
    1968 年 17 巻 173 号 p. 188-194
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
    This is the report of the study that was made of the behavior of fatigue strength of aluminum alloy 7075-T6 under a composite stress consisting of rotating bending stress and superimposed stress of higher frequency. The rotating bending stress is denoted by primary stress σ1 and the superimposed stress, by secondary stress σ2. The frequency ratio of the secondary stress to the primary stress is selected to about 12.
    Main conclusions are as follows. It was observed that the fatigue strength only depended on the frequency ratio of the primary and secondary stresses, but independent of the values of respective frequencies. The effect of the magnitude of secondary stress on the fatigue life was more or less depending on the magnitude of maximum stress σmax (=σ12). Let the σmax be higher than 1.5σww; endurance limit), the fatigue life was observed to fall when σ2max which is the ratio of the secondary stress to the maximum stress was more than 0.4, but not to fall when the σ2max was less than 0.4. On the other hand, let the σmax be lower than 1.5σw, the fall rate of the fatigue life was slight in proportion to the decrease of maximum stress. In conclusion a convenient method of estimating the fatigue strength under composite stress is proposed.
  • 田中 道七, 柴田 俊忍
    1968 年 17 巻 173 号 p. 195-201
    発行日: 1968/02/15
    公開日: 2009/06/03
    ジャーナル フリー
feedback
Top