Arc sensor, which utilizes the welding arc as a sensor, gives important groove information during welding. Automatic seam tracking control system with arc sensor developed by NKK has a fesature, that a turning point of transverse weaving with constant arc length control (AVC) is decided by the condition, when torch height reaches to constant set level.
Under applying above control system, variation of groove and deposition area can be detected from the trace of weaving. The automatic real-time bead height control system has been developed utilizing above detected information. The area and width of weaving trace has close corelation with the area of groove and deposition. Experiments conducted with varied groove width proved satisfactory control ability for all position TIG welding.

An algorithm and program for determination of welding parameters under given leg length, welding speed and length between tip and base metal by CO₂ arc welding are proposed. The welding parameters were set to obtain the globular transfer.
Calculated parameters were confirmed to given effective leg length and reinforcement by welding experiment.

Welding power supplies for pulsed GMAW have been developed, which have two circuits such as a single-phase SCR (Silicon Controled Rectifier) bridge circuit for pulse current and a three phase SCR bridge circuit for base current. In this power supply, however, the characteristic of the SCR control system is affected with the wave form deformation of power source voltage. caused by electric unbalanced load when plural machines are operated at the same time.
In this case, unstable arc phenomena occurs and the weldability becomes poor.
As the countermeasure for the above phenomena, upper limit of primary circuit's. impedance, which is a cause of the wave form deformation power source voltage, and the relation between the impedance and unstable arc phenomena were confirmed by the theoretical calculation and experiments, and further more, pulse arc voltage feedback control system was devised so that stable welding is possible, regardless of the impedance value.

High temperature strength of joint and susceptivity of stress relief anneal cracking for 2 1/4Cr-1Mo steel EB welded have been investigated.
Results are summarized as follows;
(1) Strength of EB welded joint at room and high temperature satisfy with base metal code.
(2) Susceptivity of SR cracking for weld metal by EB welding is similar to that for HAZ. It's critical stress for SR cracking is superior to that for simulated HAZ by SMAW.
(3) Residual stress of EB welded joint decrease to about 49MPa after 690°C×8Hr SR.

Dissimilar joints of nonferrous metals like as contacts of micro switchgears are required high quality and non-deformation at bonds.
For this purpose, vapor-shielded pressure bonding process and in-process control system applicable to the process are developed.
For the control of material deformation, heating characteristics and temperature distribution while bonding are examined by computer simulation as the first step for the development of the above process.
Then, the optimum electrode materials and their constitution are estimated for the joint couple of Ag and Cu.
Results obtained are summarized as follows.
(1) Electrodes Mo-Mo (φ4 mm) can rapidly heat the joint to the bonding temperature under the condition of comparatively low current density. But, these electrodes cause extreme deformation of Ag and unstable heating characteristics.
(2) Comparing with small diameter (φ4 mm) of electrodes, large one (φ10 mm) can improve the Ag deformation and the stability of heating. However, neither grade of deformation nor bonding stability are allowable for practical use.
(3) Over-excess of deformation at high current region contracts the allowance of practical bonding condition.
(4) Computer simulation can be applied to estimate the heating characteristics by considering contact resistance from the experimental results.
(a) Electrodes Mo-Mo give relatively flat temperature distribution in the joint materials.
(b) Electrodes CuCr-CuCr, which can heat the joint to the bonding temperature by high current condition, can give the peak temperatre at the bonding interface. So, bonding can be achieved under lower average temperature in the whole joint than with electrodes Mo-Mo.
(c) Arrangement of each electrode such as CuCr to Ag side and Mo to Cu side can achieve bonding under relatively low current condition and low average temperature in Ag material.

A copper-tungsten sintered alloy (Cu-W) containing 70 mass% W (tungsten particle about 3μm in diameter) has been friction-welded to a tough pitch copper (Cu). The friction speed for the welding was 40 s^-1, the friction time 0.5 to 2.0s, the friction pressure 50 MPa, and the forge pressure 50 to 340 MPa. The tensile strength of joint increased with an increase in forge pressure P₂, approaching a saturated value of 280 MPa at P₂ higher than 290 MPa. At P₂ below 250 MPa, unbonded area, which decreased with increase in P₂, were observed on fractured surfaces. At P₂ above 290 MPa, though no unbonded area could be observed, fracture occurred mostly in Cu within 140μm from the weld interface. A microhardness test showed that a weld heat-affected zone formed in Cu had the lowest harndess number at a distance more than 0.3 mm from the weld interface. The location of the lowest hardness number did not coincide with that of the fracture, suggesting that the decrease in the hardness number did not cause the fracture in Cu immediately next to the weld interface. A number of tungsten particles, which were probably picked up by mechanical mixture during the friction process, were observed in Cu within 50μm from the weld interface. On fractured surfaces of Cu side, many tungsten particles were observed at the bottom of dimples. From these results, it can be concluded that the tungsten particle picked up in Cu acts as the nucleation site of fracture and so is responsible for the fracture in Cu immediately next to the weld interface.

Behavior of the oxide film on the intimate contact zone of diffusion-welded interface has been investigated with Auger electron spectroscopy, SEM and EPMA. Results obtained are summarized as follows.
1. Behavior of the oxide film on the intimate contact zone of iron, SUS304 stainless steel, copper and titanium is similar to that on the void surface of diffusion-welded interface.
2. On the diffusion-welded interface of aluminum, the oxide film is disrupted to form metallic bonds by the deformation of contacting asperities. As the oxide film of aluminum does not disappear during welding, tensile strength of welded joint is low.
3. On the intimate contact zone of diffusion-welded joints in copper, iron and SUS304 stainless steel, inclusions are formed owing to spheroidizing of the oxide film. Inclusions consist of various oxides and sulfides. The amount of inclusions decreases with a decrease in the thickness of the oxide film of welded surface and with a increase of welding time.
4. Behavior of the oxide film on diffusion-welded interface can be classified to three types as follows:
A type-The oxide film on void surface and intimate contact zone disappears during the initial stage of diffusion welding. Titanium belongs to this type.
B type-The oxide film on void surface and intimate contact zone disappears, forming inclusions owing to spheroidizing of the oxide film. The amount of inclusions decreases owing to diffusion of oxygen during welding. Copper, iron and SUS304 stainless steel belong to this type.
C type-The oxide film on diffusion-welded interface does not disappear during welding. Aluminum belongs to this type.

Solid state bonding is genrally applied to joint of a dissimilar metal which is very difficult to bond by fusion welding techniques.
In this paper a diffusion bonding of aluminum bronze containing Si to austenitic stainless steel has been investigated. The investigations were performed by both method of a direct bonding without insert metal and bonding with insert metals of Cu, phosphor bronze and Ni foils under a restrained condition or a constant bonding pressure, 0.94 MPa. The results obtained as follows; (1) The bonding strength in the insert metal bonding was higher than that in the direct bonding and it was the highest in case of using Ni insert metal among others in the bonding temperature below 890°C. A maximum strength of 410 MPa was obtained at 710°C. (2) In case of using Cu insert metal, the bonding sterngth was much lower comparing to other cases, especially lower than the case of using Ni in the low temperature below 950°C, however, in the temperature of 960°C both bonding strength became almost the same. (3) At the temperature of 960°C and higher than that, the optimum results were obtained with melted down phosphor bronze, it is considered that a liquid phase bonding occurred then. (4) In case of using Ni insert metal, the effect of thickness differences of insert metals ranging between 10 and 60μm on the bonding strength was not recognized extending over all of the bonding temperature used. On the other hand, in the Cu insert metal bonding, the bonding strength in the case of 10μm foil was higher than that of 100μm foil and the strength difference in both foils was from 50 to 100 MPa. (5) The tensile fracture occurred at the boundary interface between the two base metals in the case of the direct bonding, it occurred too in the case of insert metal bonding between aluminum bronze and the insert metals such as Ni or Cu. For phosphor bronze insert metal bonding fracture occurred respectively at three places which are the interfaces between aluminum bronze and the insert metal, between the insert metal and a diffusion layer produced in stainless steel side and between the diffusion layer and stainless steel.

Fusion weldings of cast irons are very difficult to cause the cold cracking in the weld metal and heat affected zone. So, solid state bonding of ductile cast irons having the strength of 360, 390, 490 and 690 MPa and the ductile cast irons to a mild steel have been investigated using the diffusion welding technique at 730 to 980°C under restraint of test pieces or constant bonding pressures. The results obtained are as follows; (1) When the bonding of the cast irons were performed without a insert metal under the bonding temperature of higher than 820°C and for the bonding time of 1.8 ks, both graphites and ferrites precipitated at the bonding interface according to the transformation of γ→α+Gr on the cooling process. However, the fracture in the tensile test of the joints occurred in the base metal, and also in case of the bondings of the ductile cast irons using a insert metal of Ni foil of 10μm thickness, Cu foil of 10μm thickness or Fe foil of 100μm thickness the fracture in the base metal. Welding deformation of the joints was less than 1% in the diameter of the test piece. (2) The tensile fracture of the ductile cast irons bonded with the Ni foil at 820°C and for the bonding time of more than 0.24 ks under constraint of the test pieces or the bonding pressure of more than 0.8 MPa for 1.8 ks occurred in the base metal. (3) When the ductile cast irons and a mild steel were bonded by the diffusion welding using the Ni foil as the insert metal, the tensile fracture of the joint occurred in the base metal of the mild steel. However, in the bondings of the ductile cast irons and the mild steel with the Cu or Fe foil as the insert metal and without insert metal, the tensile fracture occurred at the bonding interface between the ductile cast iron and the insert metal. The tensile sterngths were about 400 MPa in all joints. (4) The phenomenon of the fracture at the bonding interface in the joints of the cast iron and the mild steel is considered that radial strain abruptly increases on the tensile axis because the hardness became discontinuity and the ductility of the cast irons is much smaller than that of the mild steel.

A copper plate 2 mm thick has been bonded to an alumina ceramic disk 20 mm in diameter and 7 mm in thickness using an insert metal of titanium foil 0.02 mm thick without external pressure to the bond interface. The bonding has been carried out in a vacuum at a bonding temperature of 1173K above the (lowest) eutectic temperature between copper and titanium.
The thermal stress which is generated in the cooling process after the bonding by the difference in the amount of thermal shrinkage between copper and alumina ceramics has been analysed by the finite element method. The analysis showed that the thermal stress in the ceramics was influenced strongly by the arrangement of the copper and ceramic specimens: the thermal stress in the ceramics became much lower when a couple of ceramic specimens were bonded to both sides of a copper plate (symmetric arrangement) than when a ceramic specimen was bonded to one side (asymmetric arrangement).
In fact, cracks initiated at the periphery of the bond interface were observed in the ceramics for the asymmetric arrangement. A joint of the symmetric arrangement could be cut into two joints (copper-bonded alumina ceramics) of the asymmetric arrangement without any crack. The copper-bonded alumina ceramics obtained could be brazed successfully to a mild steel using filler metal of BAg8 (JIS).

Proper welding conditions, welding defects, joint micrography and hardness are investigated, in order to apply electron beam welding for nuclear and thermal power plant components made by carbon steel and Cr-Mo low alloy steel. Materials used are SGV49, SB49 and SM41A as carbon steel, and SCMV3A
(I 1/4Cr-1/2Mo steel) as Cr-Mo steel. The results are summarized as follows;
(1) Electron beam welding is applicable for these steels up to 100 mm thickness. Proper welding condition range in heat input per unit thickness lies between about 8 and 25 kJ/cm².
(2) In order to prevent hot cracking, (P+S) content should be below 0.03% in steel whose C content is up to 0.35%.
(3) Porosity defect due to oxygen in steel did not occur in these commercial steels.
(4) Soundness of macro- and micrographies, and hardness of welded joint are confirmed.

A frequency band required for an oscillation technique on an electron beam welding was discussed by analyzing vibration frequencies of molten metal. Lower limits of beam scanning velocity not melting surfaces of work-pieces were also discussed by calculating temperature rise. A high-speed beam deflecting device to satisfy above conditions was developed and some applications of high-speed beam deflection were studied.
The conclusions obtained are summarized as follows:
(1) The vibration frequencies of the molten metal are less than 10 kHz and a beam oscillation with higher frequencies than 10 kHz is necessary to control distribution of an input power to a work-piece regardless of molten metal vibration.
(2) A beam scanning velocity of more than 10000 m/min is required not to melt a steel surface for a beam power of 10 kW
(3) A cutoff frequency of 60 kHz (-3dB) in small signal region and a maximum scanning velocity of 23000 m/min (deflection distance 300 nun) are realized using a deflecting device newly developed.
(4) The high-speed deflection technique is very effective for the new-field of electron beam application such as marking, patterning, simultaneous multi-beam welding, surface hardening, and so forth.

A new formula BL70 has been developed for estimating HAZ Hmax in a steel bead weld. A backward logistic curve Hmax=H_∞+K/(1+exp (a (Y-Y₅))) has proved to approximate with excellent accuracy to the measured relation between Hmax and Y(=log t_8/5, where t_8/5 is the cooling time between 800 and 500°C). The values of material constants K, a, and Y₅ were determined for each of seventy test steels (HT50 to HT100 of C%=0.017 to 0.33) by simple regresison analysis on measured values of Hmax and Y.
Stepwise multiple regression analyses were performed for the material constants, K, aK and Y₅ on eleven alloying elements as independent variables, and finally BL70 formula was obtained. The formula has proved to be more reliable than other formulae thus far proposed by Beckert, Yurioka, Düren, and Terasaki, respectively.
Relationship between carbon equivalents which are related with Hmax are also discussed. As the material constants can be expressed as linear functions of C% and Pcm%, a simplified formula BL70S also has been developed, in which only three variables, C, Pcm and t_8/5 are needed for estimating Hmax with satisfactory accuracy. It has satisfactory accuracy very close to BL70 formula.
For other Hmax data besides the BL70 steels, the simplified BL70S formula seems to be more accurate than the original BL70 formula. In conclusion, BL70S formula is more recommended than BL70 for future general use. However, for low-Mn-low-Pcm type TMCP (CLC) HT50 steels, BL70SM formula should be used.
A revised. formula CEm is proposed to substitute the IIW carbon equivalent CE; CEm=CE+(Si/24+ 15B).

The stress relief phenomena in pre-loading on butt welded plates and the changes of material properties owing to the pre-loading were investigated theoretically and experimentally. A over-matching welded joint composed of a 50kgf/mm² class weld metal by automatic CO₂ welding and a mild steel of 16 mm thickness was tested.
The strain distributions measured by strain gages showed the same tendency to those calculated by FEM, under pre-loadings in directions perpendicular and parallel to the weld line. The perpendicular loading was superior to the parallel loading in terms of reducing the plastic strains caused of the stress releaving.
Strain aging embrittlement of weld metal, HAZ and base metal obtained by 2 mm V notched Charpy test was quite proportional to prestrain, and significant difference among them could not be found. The equations for estimation of the brittle fracture initiation temperature were checked. The rise in the temperature caused by the pre-loading was less than that due to the weld residual stress. It was concluded that the effective pre-loading method recommended in the present paper was sufficiently safe.

When high speed submerged arc welding (SAW), was carried out using highly basic fluxes to decrease the oxygen content of weld metal of large diameter linepipes, a continuous shrinkage cavity happend to be formed on the bead surface and obstructured high-speeding of SAW in the past.
To raise the welding speed of SAW through the solidification process improvement, findings concerning the weld puddle shape and the liquid metal flow in the fundamental studies of Reports Nos. 1 and 3 have been effectively applied to the present investigation.
The method of preventing the shrinkage cavity has been experimentally examined from the aspects of weld puddle shapes and solidification structures. And it was elucidated that the shrinkage cavity occurred when the healing action of the liquid at the rear part of tear-drop typed weld puddle was insufficient, and would be prevented by positively forming the equiaxed dendrites at the bead center which approximate the rear part shape of the weld puddle to an elliptical one.

This study was made to investigate the effects of dissolved oxygen, strain rate and solution treatment temperature on the stress corrosion cracking (SCC) of the sensitized SUS304 stainless steel. For the specimens solution-treated at 1373K, the high IGSCC susceptibility was attributed to Cr-depleated zone and the segregation of phosphorus in the water with dissolved oxygen. However, in the deaerated water, the IGSCC susceptibility decreased. As the solution treatment temperature increased, the degree of sensitization evaluated by Strauss test decreased owing to a narrow Cr-depleated zone. In a slow strain rate test (SSRT), IGSCC susceptibility increased with decrease in the strain rate independent of the solution treatment temperature. But, for the specimens solution-treated at 1573K, SSRT results were not correspondent with Strauss results, that is, SSRT was a severe test compared with corrosion tests such as Strauss test. It was found that the electrolytic conductivity of high temperature water in SSRT was higher on sensitized specimens than on solution-treated specimens. This increase in the electrolytic conductivity seems to be related to the pit formation at the grain boundaries of the sensitized specimens.

Crack susceptibilites of weld crater and weld bead of synthesized-weld-metal Al-2%Zn-3%Mg alloy with and without zirconium up to 0.36% have been evaluated by means of GTA spot weld crater, Houldcroft type and Vatrah weld bead cracking tests.
It was proved that zirconium was beneficial additional element to reduce crack susceptibility of both weld crater and bead with more addition than 0.24%. This beneficial effect of zirconium was closely related with grain refinement of weld structure. The grain refinement of weld bead resulted in the increase in the cracking threshold in BTR, although BTR was not changed.

It appeared that electromagnetic stirring was very effective to improve the susceptibility of solidification cracking in weld metal of Al-2%Zn-3%Mg alloy with 0.24% zirconium. This beneficial effect of electromagnetic stirring was mainly due to the grain refinement of weld metal, but it was not seen for the alloys with less zirconium content than 0.24%.

Mechanism of ductility-dip cracking in weld metal of Invar (Fe-36%Ni alloy) reheated by following welding, pass was studied from the viewpoints of grain boundary precipitation and grain boundary serration. Main conclusions obtained areas follows: (1) Temperature dependency of amount of precipitates on the fracture surface, has good correlation with the characteristic of hot ductility. (2) The fracture surface of the specimen having low crack susceptibility has comparatively little amount of precipitates. (3): The features of intergranular fracture surface under different testing temperature and crosshead speed have intimate connectionn with that of grain boundary serration, and thus this suggests strongly. that grain boundary serration has an important role on'the ductility-dip crack. (4) Thus it is judged that both grain boundary precipitation and grain boundary serration cause cavity foi nation cooperatively accompanied by grain boundary sliding. (5) The mechanism of ductility-dip cracking was proposed synthetically from the viewpoints of grain boundary sliding, grain boundary serration, grain boundary precipitation, grain boundary migration, deformation stress and recrystallization.

The authors researched cold rcacking in electron beam (EB) welding by performing hardness and cracking tests on 28 different types of carbon and low-alloy steels.
Based on the test results, the welding paramteer (P_EBW), a factor used to estimate the cold cracking in EB welding, has been determined. The conclusions are summarized as follows;
(1) The maximum Hardness Value (H_Vmax) of the EB weld can be estimated by equation (1).
H_Vmax=(1020⋅C+268)±60…………………………………………………………………(1)
where C: the carbon content of the material (wt%)
(2) Restraint stress of EB weld can be estimated by equation (2)
σ_w=0.007R_F……………………………………………………………………………………(2)
where σ_w: restraint stress (MPa), R_F: restraint intensity (N/mm.mm)
(3) The cold cracking in EB weld can be predicted by equation (3)
H_Vcr=-0.16_rf+750…………………………………………………………………………(3)
where Hvcr: the critical hardness value of crack occurrence
rf: restraint coefficient (N/mm²·mm)
(4) The welding parameter (P_EBW) can be determined by equation (4)
P_EBW=C+l/6375·_rf…………………………………………………………………………(4)
when Q≤13500J/cm, without preheating; P_EBW=0.475

In this research, T-edge method and L_z method, in which inherent strains are dealt as parameters, are applied to measurement of three-dimensional welding residual stress distributions produced by electron beam welding (EBW). Accuracy of the measured residual stresses are discussed. As a result, characteristics of three-dimensional residual stress distributions due to EBW are clarified.
The main conclusions are as follows:
1) Validity and availability of T-edge method, developed in order to save. measuring time and expense, are demonstrated. by showing concretely the accuracy of three-dimensional welding residual stress distributions measured by this method.
2) Discussion on accuracy of residual stresses measured by T-edge method and L_z method suggested that residual stresses produced by EBW should never be uniform along the weld line, however continuous welding is applied for long weld length under the same restraint condition.
3) Residual stress in the welding direction, σ_x, produced in the weld metal distributed in the plate thickness direction varies from spot to spot in magnitude and distributes complicatedly. Since melting region is very narrow in EBW, dynamical restraint condition of the weld metal becomes so severe that the same large tensile residual stress as in HAZ is produced in the weld metal.
4) Residual stresses produced in the weld metal perpendicular to the weld line, σ_y, and in the plate thickness direction, σ_z, show characteristic distributions; i.e., σ_y is compressive on the top and bottom surfaces and tensile in the middle of the plate thickness, and σ_z is compressive in the plate thickness direction. The same stress distributions appear when instantaneous plane heat source is placed in the cross section of a plate. In case of EBW, it maybe assumed that instantaneous heat source is placed in the cross section, though no such assumption as moving plane heat source is placed continuously along the weld line can be made.

A new measuring method for axisymmetrical residual stress distribution which is uniform neither in radial nor in axial direction, is proposed for butt welded thick pipe.
The formulas to estimate residual stresses are first introduced following a maeasuring method that has already been conducted for axisymmetrical residual stress field in a thick plate. This method requires several numbers of thin arch-specimens normal to the pipe axis to be cut out along with a z-specimen parallel to the axis. In this case, the strain change in circumferential direction, that is, Δε_θ in an arch-specimen has great influence upon the accuracy of the estimated stresses. Therefore, its distribution along radial direction is next analyzed based on the elasticity, followed by the analytical investigation by finite element method using inherent strain distributions that may approximate residual stress distributions measured on a butt welded thick pipe.
As a result, it is clarified that the distribution of Δε_θ in radial direction may be approximated with a good accuracy by a function of A+B/r² where A and B are so determined as to pass through both measurements on inner and outer surfaces of each arch-specimen.
The final investigation on the accuracy for estimated residual stresses using the same inherent strain distribution concludes that the circumferential residual stress σ_θ can be estimated with a good accuracy by the first approximation although the relieved stress in radial direction in cutting a z-specimen, Δσ_r is ignored and at the same time the strain change in axial direction Δε_z is approximated by a straight line joining both measurements on inner and outer surfaces of z-specimen. Axial and radial residual stresses, however, should be estimated just to the second approximation to ensure the accuracy where the distributions of Δσ_r, and Δε_z, along radial direction calculated by making use of approximated σ_θ are utilized.

In order to make sound welds free from defects like weld cracking, characteristics of residual stresses around nozzle welds were experimentally investigated in an as welded condition for a thick-wall reactor vessel and an accompanying heat exchanger utilized in a petroleum refining process or a coal liquefaction process.
The residual stresses were measured throughout the thickness of the weldments by applying the measuring methods that had been newly developed for axisymmetrical stress distributions by the authors.
The main results obtained for set-in type nozzle, set-on type nozzle and simple butt welds of pipes are summarized as follows.
(1) Stress parallel to the weld line-The stress indicates the same distribution in the thickness direction regardless of the joints. That is, the maximum value appears just behind the final layer of welds and with descending to the initial layer it decreases. The stress gradient, however, differs from one joint to another, depending on the self-restraint intensity of the joint. Moreover, the peak value of simple butt welds of pipes is about 100 MPa smaller than the others.
(2) Stress perpendicular to the weld line-It indicates the similar distribution as the stress parallel to the weld line. The maximum value in the set-in type nozzle changes from 400 MPa to 600 MPa with the decrease in the diameter of welds from 600 mm to 300 mm. As for the set-on type nozzle, it reaches almost the same level due to the existence of channel as a stress raiser. The stress for the simple butt welds of pipes, however, indicates much smaller value of about 245 MPa and further the same order of tensile stress occurs in the inner surface as well as just behind the final layer of welds in the outer side.
(3) The stress in the thikness direction shows the distribution where the maximum value appears in the center portion of the thickness and it is a low stress level of about 150 to 200 MPa, regardless of the joints.

The main objective of this study is to investigate the possible role of residual stress redistribution on fatigue crack propagation behavior, caused by crack extention itself initially through residual tensile stress field. Experiments of both fatigue crack propagation and sawcut simulating a propagating crack were carried out on HT80 steel components on which melt running of electron beam welding was performed to induce residual stress field. Both experimental results were compared and reviewed from a viewpoint of residual stress redistribution. Clear differences between the results of both experiments were not observed, and the common feature recognized from the results was the residual stress concentration consecutively occurred in the vicinity of crack tip even the situation that the crack propagated to the region initially residual compressive stress existed. Prediction method that attempts to account for the observed results are proposed.

Some series of large-scale fatigue tests have already been conducted on the longitudinal groove weld, which is familiar at the corner welded joint of box sectional members. It was cleared in these tests that the fatigue crack originated from the root blowhole and that the fatigue strength was considerably lower because of the effect of the tensile residual stress. When disallowable blowhole is detected at fabrication, the reduction of the tensile residual stress seems to be useful to improve the fatigue strength instead of the repair welding. In this study, firstly, techniques for the local reduction of the tensile residual stress at the corner welded joint were examined, namely, stress relief annealing, low-temperature stress relief and bead welding. This research proved that low-temperature stress relief was most efficiently and economically method. Secondly, the reliability of low-temperature stress relief was confirmed on a large-scale box member. Finally, the fatigue test was carried out to investigate the effect of low-temperature stress relieving on the fatigue strength of welded joints. The result verified that stress relief could improve the fationv ctrnnoth

The development of high field conducting technique has created needs for new structual alloys and ensured joining methods. In this report, in order to examine the applicability of the explosively bonded joint for dissimilar materials to cryogenic machines, A5083P-SUS304L explosively bonded joint which had AllOOP, Ti and Ni layers as insert metals was prepared, and deformation and fracture behavior of its joint was investigated. The tensile test was carried out at 293K, 77K and 4.2K.
As the results, it was clarified that the deformation and fracture behavior of that joint was hardly affected by A1100P used for insert metal and that its strength was increased by restraint of deformation. Also it was recognized that the bond interface fracture was occurred in the region of lower temperature and harder restraint. But their ultimate tensile strength was twice as great as the offset stress of A1100P at least even at 4.2K.

Fracture performance and fracture toughness tests have been carried out to clarify fracture behaviors of welded joint of low yield-ratio HT80 (LY HT80) made by using TMCP process followed by Q'-quenching and tempering. Two kinds of HT80 steel, QT type conventional HT80 and CR-Q'-T type low yield-ratio HTRO which have the same chcmical composition, are prepared. Three point bending COD test, and narrow and wide plate tensile tests have been conducted for welded joints with a fatigued pre-crack along fusion line; through-thickness type notch or surface notch.
There is a considerable difference of fracture performance in welded joint between of low yield-ratio HT80 and of conventional HT80, regardless fracture toughness of crack tip region in both welded joints is almost the same. Fracture resistance such as critical COD of LY HT80 welds apparently increases due to remarkable plastic deformation in the vicinity of crack tip, and fracture ductitity of LY HT80 welds becomes larger than that of conventional HT80 welds at the temperature range at which general yielding fracture initiates. An appropriate decrease in yield strength and yield-ratio of HT80 steel is a proper treatment for obtaining sufficiently large fracture performance at the temperature range for actural use of HT80 steels.

The objective structure of the present study is LPG storage tank and numerous bending COD tests and tensile tests with through thickness notches have been conducted on the practical low-temperature steel weldments. Statistical approach has been applied for considering the scattering features of critical COD in the practical joints and correlation of critical COD obtained from both tests.
The bending critical COD scatters widely even in the practical welded joint due to the existence of local brittle zone. The bending critical COD (δ^B_c) and the tensile critical COD (δ^T_c) are influenced by the plastic constraint dependent on specimen geometry. In case the fracture stress is lower than yield stress (σ_y), δ^T_c≅δ^B_c. Nevertheless, when the fracture stress is higher than σ_y, that is fracture occurs in full yielding and general yielding. δ^T_c becomes considerably larger than δ^B_c due to the decrease of the constraint.
According to the above experimental results, consideration has been carried out on the procedure of the assessment of fracture toughness by using bending COD test. Accordingly, the following two cases should become acceptable in bending COD test for the welded joint of LPG storage tank.
(a)Critical COD values of three bending COD test specimens are satisfied required value; e.g. 0.25 mm according to API 1104.
(b)When one of three bending COD values is not satisfied the required value and min. value is more than critical COD of general yielding fracture in the practical welds; e.g. δ_c=0.00028σy (σ_y in MPa, δ_c≅0.1 mm in HT50), a second set of three test specimens may be performed. Five of the six tests must exceed the min. required critical COD. Furthermore, in that case, average value of δ^B_c is more than 0.5 mm.

Crack opening displacement (COD) testing is widely used to determine the fracture toughness of steel materials. However, COD tseting of thick weldments involves the problem of a non-uniform shape of fatigue precracking through the thickness due to the influence of welding residual stress. In this study, first, the residual stress distribution in a thick weld metal COD specimen was measured to clarify the cause of the irregular fatigue crack front. Then, initial machined notches of various shapes were tested to select suitable ones which would not affect the material toughness and would not have a non-uniform shape. As a result it was found that a partial arc notch and partial drill hole notch meet the above requirements.

The notch toughness of brazed joints with a gold-18%nickel alloy and an amorphous nickel-base filler metal was investigated. The base metals used were SUS304 stainless steel, SUS420J2 stainless steel, Inconel X-750 nickel-base superalloy and MAS-1 maraging steel. The specimen for the notch toughness test was a rectangular butt-brazed type which finished the standard 2 mm V-notch Charpy specimen after brazing. These specimens were tested at temperature ranging from -196°C to 600°C.
In the 304 stainless steel joint brazed with gold-18% nickel filler metal, the absorbed energy of the joint was low at temperature below room temperature because the strength of the base metal was considerably higher than that of the filler metal.
In the both 304 and 420J2 stainless steel joints brazed with nickel-base filler metal, a brittle eutectic liquated grain boundary formed at and near brazed zone due to the boundary diffusion of boron. The toughness of these joints was reasonably low.
Notch toughness of Inconel X-750 joint brazed with gold-18% nickel filler metal was extremely low over the whole temperature range from -196°C to 600°C. This was the reason why the strength of Inconel X-750 superalloy scarcely decreased up to the temperature of 800°C, compared to appreciably decrease in that of gold-18% nickel alloy. On the other hand, in the joint brazed with nickel-base filler metal for longer than 10 min, the brazed region revealed a homogeneous structure resulting from an isothermal solidification. Therefore, the joint indicated the fairly high toughness.

Effect of welding tensile residual stresses on fatigue crack propagation in low propagation rate region (low ΔK region) was studied experimentally. Material used was steel of 800 MPa in tensile strength. Specimens were center-cracked-plates and consisted of three kinds; specimens without weld, specimens with longitudinal weld (as weld), specimens with longitudinal weld (stress relieved). Major results obtained are summarized as follows:
(1) Tensile residual stresses decreased the threshold value of ΔK.
(2) When tensile residual stresses were large, the threshold value of ΔK was unique and independent of stress ratio.
(3) Effect of residual stresses could be dealed with as effect of mean stresses.

It is known that stress concentration and strain concentrtion occure at the root of welded joints with backing strip. Root radius is very important in relation to fatigue strength of the welded joints with backing strip. In this paper, statical equilibrium equation was derived from interfacial tension phenomena of molten metal in the welded joints with backing strip which was fabricated by flat welding or overhead welding. Influencing factors on root radius were studied from the statical equilibrium equation. Influence of torch angle on root radius was experimentally investigated.
The results obtained in this report are as follows;
1) The root radius became larger when misalignment was smaller at the welded joints with backing strip.
2) The root radius was larger when horizontal length between upper and lower fusion penetration point was larger for the welded joints with constant misalignment.
3) The root radius in overhead position was smaller than flat position due to the term of influence of gravity in the statical equilibrium equation.
4) The root shape was convex with solid wire and concave with flux cored wire. Accordingly, is is considered that root shape was influenced by slag.
5) The root radius was largest when torch angle was 90°C for flat position and 70°C for overhead position.