This paper describes the development of a portable detector-technique X-ray stress analyzer capable of measuring the residual stress of large-size structures, the half value width of diffraction lines, etc., and reports the results of actual measurements performed with this analyzer. The results are also useful as the data of materials strength. The analyzer consists of four components, i.e. an X-ray generator, a goniometer, a heat exchanger and an electronic circuit panel. Each unit weighs approximately 18kg on an average and is designed to be carried with ease. This analyzer adopts the parallel beam optical system, Sin2ψ method and X-ray incident angle oscillation method, so that the accuracy of the measurement is equivalent to that of those equipments installed in laboratories for research work. Emphasis is placed on the improvement of operation through simplification of its function and construction and employment of automatic controls as much as possible, to obtain smooth operation even under severe conditions. Moreover, the results of its practical applications on large-size structures, such as the detection of the effect of stress relief on welded parts performed in the field, showed that the change in residual stress before and after the annealing can clearly be obtained with this analyzer. Such measurement has not been conducted so far because reinstallation of existing analytical equipments is impractical.
The strength of large structures such as vessels, slips or bridges depends mostly not only on the residual stress caused by welding but also on constrained stress, and therefore a non destructive stress measurement on a real object is demanded. Based on the experience that one of the authors previously made X-ray stress measurement of ship construction, authors newly developed a compact type of X-ray diffraction stress analyzer which adopts a modified side inclination method and enables the stress measurement at the corner of structures. This analyzer has also such features as an incident X-ray oscillation mechanism for the stress measurement of coarse grained materials, a direct charge system for making counting circuits small and 5×5mm2 large focal spot X-ray tube which is useful in side inclination method. The relation between the measured stress σx' and the true stressσx is theoretically calculated as follow. σx'=0.96σx+0.006σy-0.08Sin2φ(σ1-σ2) The error of σx' is negligibly small, and is found to be comparable to that of the data by using a conventional type of apparatus, based on the experimental results on S15C and S40C heat-treated specimens under uniaxial stress. Goniometer setting error under ±2mm also does not increase the error in X-ray stress measurement.
In the previous paper, the constant ψ method with ten angles incidence by using an X-ray diffractometer has been recommended for the X-ray stress measurement of austenitic stainless steel. This method, however, requires a very long time for the measurement, and furthermore cannot be applied to actual measurement of large machine parts. In the present study, therefore, the constant ψ0 method with oscillation technique by using an X-ray diffraction stress analyzer was applied to the residual stress measurement in the ground surface layer of 18Cr-8Ni austenitic stainless steel in order to clarify the characteristics of ground surface layer and examine the accuracy of stress measurement. The results obtained are as follows: (1) The distribution of micro-strain formed in the ground surface layer corresponds to the changes of hardness and half-value breadth of diffracted X-ray profile in each layer. The depth of the strained layer increases with the increase of the downfeed of grinding, and that in dry grinding is larger than wet grinding. (2) During the grinding of this testing material, the strain-induced transformation γ→α hardly takes place, because the surface temperature of workpiece rises. (3) A high tensile residual stress reaching from 60 to 70kg/mm2 appears in the ground surface which has been ground severer than the condition of downfeed of 0.005mm by wet grinding in this experiment. And in the grinding of 18-8 stainless steel, thermal stress acts effectively to characterize the form of residual stress distribution. (4) Using the constant ψ0 method with oscillation technique by an X-ray diffraction stress analyzer, the residual stress in the ground surface layer of 18-8 stainless steel can be measured accurately. However, the surface residual stress value obtained has to be corrected somewhat owing to the K-value used for stress calculation, the steep stress gradient and the roughness of ground surface.
A non-destructive measuring method of physical quantities such as the X-ray diffraction peak intensity, X-ray broadening and stress distributions inside a sample is proposed. This method is based on the concept of effective X-ray penetration depth. In the case of the poly-chromatic X-ray diffraction technique, many diffraction peaks of different energies can be obtained at the same time. The X-ray linear absorption coefficient is, however, quite different depending upon the X-ray energy. Thus, the physical quantities at various depths of the sample can be easily obtained by the poly-chromatic X-ray technique. The usefulness of this technique is demonstrated here by measuring the peak intensity distribution from the fatigue fracture surface and determining the plastic zone size, the real surface X-ray peak intensity, and the peak intensity gradient. It is also confirmed that the X-ray broadening obtained by the poly-chromatic X-ray method is distinctively different for different samples, for example, the one having fatigue fracture surface and the other having brittle fracture surface.
An investigation has been made on the relation between half value breadth (H. V. B.) and hardness of quenched-tempered steels. The following results have been obtained. (1) There are some good correlations between H. V. B. and hardness of quenched-tempered steels. These correlations are classified by the carbon content of steels and are well worth taking advantage for engineering use. (2) The change in H. V. B. is not influenced by precipitation hardening though it is remarkably affected by solid solution hardening and dislocation density. However, the change in hardness is not so remarkable as that in H. V. B. with dislocation density. (3) The result of tensile test revealed that H. V. B. having a value over 4.0 degree in 2θ decreases with plastic strain until the strain reaches at a certain value before it began to increase with strain, though H. V. B. under 4.0 degree increases monotonously with strain from the initial state.
The strain state of a single crystal of austenitic stainless steel induced by hydrogen charge was measured by the divergent X-ray method. The stress-strain analysis of a face centered cubic crystal has been established already and it is important to measure the six lattice strains in their precise calculation. In this paper, the change of the lattice parameter was measured from a perfect diffraction ellipsoid, reflected from six planes. From the measured six strains, the stress and strain states of single crystals of austenitic stainless steel were analysed and their principal strain caused by solute hydrogen were found to be in    directions. Thus, it is considered that hydrogen atoms occupy the octahedral sites in a face centered cubic lattice.
It is well known that delayed fracture (sometimes called delayed failure, or static fatigue) is a serious problem when high tension steels having the yield strength of more than 120kg/mm2 are used as pre-stressed concrete steel bars, hexagon bolts, members of aircraft etc. Hydrogen embrittlement is considered the main cause of delayed fracture. In this investigation, the notched specimens of Cr-Mo steels SCM2 and AISI 4150 dipped in 5% H2SO4 solution to charge hydrogen were used to examine the change in their mechanical properties and the effect of hydrogen by means of static bending tests, residual stress measurements by X-ray diffraction technique and delayed fracture tests. The experimental results show that the compressive residual stress near the surface caused by heat treatment seems to be released by dissolution of hydrogen and the intrinsic strength level has a close relation to the sensitivity of hydrogen embrittlement. The variation of maximum stress with hydrogen charged time in static bending tests seems to be similar to the curve of delayed fracture diagram. Therefore, it is considered that the static tests or residual stress measurements by X-ray is a useful tool to study the delayed fracture phenomenon due to hydrogen embrittlement.
The effects of temper embrittlement on the crack initiation properties of 3.5 Ni-Cr-Mo-V steel for large rotor forgings have been studied by using the burst tests with application of internal pressure on thin tubular specimens. The results obtained are as follows. (1) Temper embrittled steels with severe segregation show premature fracture before the point of plastic instability, and the fracture appearance has a very brittled grain boundary fracture mode. (2) The strength against crack initiation, which can be measured by the critical circumferential strain at bursting, decreases remarkably in the materials with fracture appearance transition temperature, FATT, above 100°C. (3) Temper embrittled steels with light segregation, however, fracture at the point of plastic instability and show a more ductile shear fracture mode. (4) The materials having higher yield strength show a noticeable decrease in crack initiation strength. (5) The metal resistance against crack initiation depends upon the nature of defects and properties of surrounding matrix. Temper embrittlement enhances the notch-sensitivity of the matrix.
The statistical fatigue property was investigated for 3 steels corresponding to AISI types 4135, 4340 and 403, heat-treated to 7 different strength levels. The fatigue tests were conducted in rotating bending with both smooth and notched specimens, so that 14 cases were studied in total, each using about 120 specimens. The distribution characteristics in fatigue strength were discussed by examining those in the stress deviations of test data from the mean S-N curve determined by the Probit analysis method, and then the P-S-N curves were obtained for all test conditions. The S-N curves for steels 4135 and 4340 revealed, with smooth specimens, no definite fatigue limit within the test range of cycles (>2×107), especially for materials of higher strength levels, while with notched specimens they exibited ordinal fatigue limits. The variation in fatigue strength could be fitted by the normal distribution in most cases of smooth specimens, but it became asymmetric with larger scatter in the weaker side in the case of notched specimens. The magnitude of fatigue variations was analysed in relation to the Vickers hardness values and a possibility of predicting the fatigue variation from that of the hardness value was shown.
The fatigue tests were carried out under completely reversed plane bending stress using the mild steel specimens with or without chemical polishing after grinding. The fatigue limit was 17.2kg/mm2 for the specimens with work-hardened layers and was 11.5kg/mm2 for the chemically polished specimens having no hardened layers. The residual stress in the work-hardened layer was measured by X-ray diffraction technique and was found to be about 15kg/mm2 in tension. Although the tensile residual stress is generally considered to be detrimental to the fatigue limit, the present study seemed to show the opposite effect. The reason for this was followed by using two kinds of experiments, namely the fatigue tests under the mean stress and the evaluation of the strength of thin work-hardened layers by X-ray stress measurements. The results showed that the mean stress has only a slight effect on the fatigue limit. With the assumption that the effect of the mean stress is the same as that of the residual stress, the tensile residual stress in this case is estimated to decrease the fatigue limit by about 1kg/mm2. The strength of grinding surface was found to be increased nearly 50% and this was considered to be the main cause of the high fatigue limit of ground specimens.
To improve fatigue and wear resistances of machine parts, a low temperature gas carbonitriding process has recently come into practical use, in which steel parts are treated in ammonia and endothermic gas phase at about 570°C. The process provides a surface layer similar to that from tufftriding process. In the present study, the effect of low temperature carbonitriding on fatigue strength was investigated by rotating bending tests on Cr-Mo structural steel specimens tempered at 600°C. The reason for fatigue resistance improvement was discussed, based on the hardness and residual stress distribution in the carbonitrided specimen. The results obtained are as follows: (1) The thickness of compound layer and nitrogen diffused zone in the low temperature carbonitrided specimen was approximately 18μm and 0.5mm, respectively. The hardness and compressive residual stress in the nitrogen diffused zone attained the maximum values near the surface and they were HV 575 and -30kg/mm2 at 0.025mm depth, respectively. The core hardness of the low temperature carbonitrided specimen was HV 313, which was equal to that of the quenched-and-tempered specimen. (2) The improvement of endurance limit of plain and notched specimens by low temperature carbonitriding was about 30 and 90%, respectively, in terms of stress amplitude. (3) The S-N relation of the plain carbonitrided specimen was evaluated from the relation between distribution of fatigue strength in the cross section and applied stress in rotating bending. The fatigue strength distribution in the specimen was estimated from the hardness distribution. The evaluated fatigue strength was 10∼20% lower than the experimental one. The difference can probably be ascribed to the fact that the fatigue crack initiated at about 0.5∼0.6mm depth from the surface and propagated to the surface in compressive stress field. (4) In the case of the notched carbonitrided specimen, fracture initiated on the surface, and the S-N relation gave a knee point at about 105 cycles. In the stress level above the knee point, the fatigue strength is considered to be affected by the hardness distribution in the cross section. However, the endurance limit agreed well with the value estimated from the notch factor and the hardness and residual stress at 0.025mm depth from the surface.
The effect of the surface non-martensitic microstructure on the fatigue of gas carburized steel was investigated experimentally. The following results were obtained. (1) The fully reversed fatigue strength of carburized steel was lowered by the surface non-martensitic layer. (2) It was confirmed by X-ray stress measurements that the surface layer deformed plastically during the fatigue test and the residual stress developed in the direction to release the applied mean stress partially. (3) Scanning electron micrographs revealed that internal oxides act as the defects to initiate fatigue cracks. (4) The fatigue fracture mode of carburized steel was characterized by the transgranular fracture with a linear pattern along the direction of crack propagation.
In order to grasp the concept of crack propagation in surface hardened materials, the observation of crack propagation and the measurements of the residual stress and the plastic zone size ahead of the crack tip were performed on a tufftrided steel. The specimens used were made of 0.12% carbon steel, machined to the shape of single notched plate type and tufftrided after annealing. The results obtained are summarized as follows: (1) The relation between the crack propagation rate and the stress intensity factor could not be represented by a simple power law. (2) The compressive residual stress was observed in the vicinity of the notch root before the fatigue test. When the pulsating bending stresses were applied to such a specimen, the decrease of residual stress was observed at the crack tip while the crack length was short. (3) The size of the plastic zone ahead of the crack tip was in proportion to the 4th power of stress intensity factor. However, if the effective stress intensity factor was corrected based on the residual stress, this relation was in agreement with the theoretical relation, the square of stress intensity factor being proportional to the plastic zone size.
The thin-walled cylindrical specimens subjected to internal pressure were fatigued in tension-compression, in order to study the fatigue behavior under the biaxial stress condition. The results are summarized as follows: (1) The circumferential stress increases the fatigue strength, while the tensile mean stress decreases it. But the fatigue limit is nearly constant independently of the stress condition. (2) The number of cycles to form cracks detectable at the external surface decreases with increasing tensile stress on the maximum shear stress plane. Although the specimen is approximately under the biaxial stress condition, the radial stress should be taken into consideration for the crack initiation. (3) The fatigue crack propagation rate is expressed by a power function of the stress intensity range. But the exponent and constant vary with the stress condition. (4) The plastic zone size ahead of the crack tip, ξ, observed with an optical microscope has a close relation with the crack propagation rate, but is independent of the stress condition and the crack length: da/dN=1.65×10-8ξ1.54 (5) The relation between the plastic zone size ahead of the crack tip, ξ*, obtained by the modified Dugdale's model and the crack propagation rate is given by the following function. da/dN=1.24×10-10ξ*1.88
The effect of test temperature on fatigue damage was studied on pure aluminium in fully reversed plane bending by using the X-ray microbeam technique. As a result, the fatigue strength was found to decrease with increasing test temperature. As the test temperature was elevated, the total misorientation β, the excess dislocation density (Db)max and (Db)min increased, and larger and more polygonized subgrains were formed. This resulted in the reduction of fatigue strength. The propagation rate of fatigue crack dl/dN was dependent on test temperature and not uniquely expressed by a power function of the stress intensity factor. However, the excess dislocation density (Db)max and the subgrain size t in the grain at crack tip were closely related to the crack propagation rate dl/dN by the following equations independently of test temperatures: dl/dN=3.26×10-24 (Db)max1.79 and dl/dN=3.08×10-3t-3.69
The rate of crack propagation was measured in notched specimens of annealed 0.04%C steel in high and low cycle fatigue regimes. Fractographic observation indicated that the single micromechanism is responsible for crack growth in the observed range of fatigue crack growth rates of 10-6∼10-2mm/cycle. The mechanical parameter which controlled the growth rate was investigated. The growth rate of fatigue crack under large scale and general yielding conditions was found to be higher than that predicted from the power equation between the rate and the stress intensity factor obtained under the condition of small scale yielding. The plastic part of the crack tip opening displacement Δφplastic and the cyclic J-integral ΔJ were found to be such parameters which gave the single value of the rate da/dN by the following equations: da/dN=2.66(Δφplastic)2.15, da/dN=1.22×10-3(ΔJ)1.35.