A study was performed to consider the effect of grain size and flux height on the bead appearance, especially on the bead ripples, based on the idea that variation of gas pressure in arc cavity may cause the molten pool motion. Welding was carried out under A.C. 700 amp. ×34 volt ×30 cm/min using SiO₂-CaO-MgO glassy flux with grain size ranging 8×12 to 150×D mesh and flux height ranging 20 to 100 mm.
Low flux height tend to result in fine ripple regardless of the grain size, but the high height result in characteristic appearance according to the grain size. The result of low flux height may be caused by relatively easier relaxation of gas pressure through a soft and thin slag layer. On the other hand, in case of high flux height, supposed reason for producing the characteristic bead appearance is as follows: Over the molten slag, a layer of flux grains is formed by adhering together with their own melt. Pressure relaxation is prevented by this layer, and irregular gas bursting causes irregular ripples. Thicker flux, or stronger bonding owing to smaller grain size promotes the irregularity. If the flux grain is fine enough, fluidization phenomenon prevents the formation of the adhered grain layer, with resultant fine ripples.

"Uranami bead" formation with DCEN polarity -as a part of a study on one side narrow gap MAG welding- was investigated in the previous papery, in which the ranges of suitable welding conditions for good "Uranami beads" and the variation of the "Uranami bead" shapes under various welding conditions were also reported.
In this paper, the effects of various factors such as blowing angle, blowing point and flow rate of control gas, wire oscillation etc., on the "Uranami bead" shape of root pass in DCEN I-groove narrow gap welding of thick plate (22 mm), were studied.
It was found that in one-side narrow-gap MAG welding with DCEN polarity, a stable "Uranami bead" with improved shape could be obtained by directing the control gas onto the surface of the molten metal, using a specially designed torch.

The authors have been carrying out practical studies on one-side narrow-gap MAG welding with DCEN polarity. In the previous paper, observation of "Uranami bead" shapes and determination of suitable range of welding condition for "Uranami Welding" were carried out and it was made clear that "Uranami beads" without defect could be achieved over a wide range of welding conditions with this welding method. However, in case of long weld line and changing groove width, it became clear too that fine adjustment of welding conditions (mainly adjustment of welding speed) would be necessary in order to achieve constant shapes of "Uranami bead" all along the weld line.
Consequently, the relationships between arc phenomena in "Uranami welding" and shapes of "Uranami bead" were observed, and it was found that the shape of the "Uranami bead" was dependent on the welding speed and the wave pattern of the welding current. Based on these results, a new "Uranami welding" control method was developed. In this method, a control signal which was obtained by monitoring the welding current with a micro-computer, was used to adjust the welding speed continuously in order to achieve a good "Uranami bead".
Application of this control system to one-side narrow-gap MAG welding with DCEN polarity confirmed that uniform "Uranami beads" could be obtained even if the root gap varies from 9 to 12 mm.

In this study, a pulsed current plasma arc is applied to the keyhole operating mode to control solidification of the molten pool. Further, the pulse time, peak current, and plasma gas flow rate are especially investigated.
A one-pulse/one-keyhole range of pulse times by low frequency is obtained at each of four positions (Flat, Overhead, Vertical-up, and Vertical-down). As a result, suitable pulse times for some combinations of the plasma gas flow rate (Q) and the peak current (I_p) are the peak time (t_p)=0.2-0.4 sec and the background time (t_b)=0.4-0.6 sec at welding speed (ν) of 120 mm/min in case of stainless steel pipe (JIS SUS304, 4 mm thickness).
Then fixed pipes are welded, and it is found that orbital welding is possible by starting from the 4:30 position and by control of the plasma gas flow rate, the peak current, and the welding speed according to the welding position at 6:00 or 7:30, 9:00, and 12:00 (o'clock) positions.
And effect of joint alignment error is studied, and it is found that permitted limits with cap pass welding are miss-matching=1.0 mm, root-gas=0.5 mm, arc-length=±1.0 mm, and arc-deviation=1.0 mm in case of mild steel pipe (JIS STPG38, 4 mm thickness).

Arc behavior of Rotating Arc Butt Welding were studied in order to extend the applicable pipe diameter and thickness ranges. Experiments were carried out for 216.3 mm O.D×5.8 mm wall steel pipes by using design of experiment in which gap, arc current and exciting current were selected as important experimental factors. Main results are summarized as follows;
(1) Arc rotating frequency is mainly affected by gap of pipes and arc rotating smoothness at the early stage of welding is affected by exciting current.
(2) The period of melting process and arc rotating process is influenced by arc current.
(3) Characteristic phenomena are observed by high speed photograph, such as showerlike spatter, small grain spatter, plural arc and dripping molten metal.
(4) It was observed much small grain spatters were dispersed at the small gap than large gap of pipes. The good quality of welded joint was not obtained at small gap condition.

It has been said that formation-process of bonded area in the solid state bonding is attained by two or more distinguishable mechanisms (plastic deformation, creep flow, interface-diffusion and volumediffusion). In the present paper, the diagrams of bonding mechanisms for copper are drawn, using three axes of bonding pressure P, bonding temperature T and percent bonded area S for the purpose of classifying the bonding processes. The diagrams allow the bonding process to be visualized by identifying the dominant mechanisms, resulting in the solid state bonding systematically divided into two processes of deformation and diffusion in terms of the dominant mechanisms. Also, the mechanism maps with contour curves are useful in planning the bonding process.

Press-Soldering is a method to make a joint at a relatively lower temperature and obtain a higher joint strength. Press-Soldering experiments at I kgf/mm² and 3 kgf/mm² pressure were done using Bi-Pb type solder with Ag and Zn additions to obtain a sound alloyed layer at a temperature below 350°C. Tensile strength, heat-resistivity and composition of the alloyed layers of press-soldered Cu to Cu joints were investigated. A joint which was press-soldered at 350°C with 3 kgf/mm² pressure using a Bi-42%Pb-3%Ag-2%Zn solder showed the best result. The joint had 8 kgf/mm² of tensile strength and the remelting temperature above 820°C. In this case, two alloyed layers were found in the joint interface. The composition of the alloyed layers were Cu-18.5%Zn-2.5%Ag and Ag-29%Zn-3.5%Cu respectively.

Joining of alumina (99.62mass% Al₂O₂) to copper was made using amorphous Cu-Ti (34-57 at%) brazing filler metals at 1173 to 1323 K in 2.7×10^-3 Pa. The amorphous filler metals, 1 cm in width and 45 μm thick, were produced by liquid quenching. The superior flexibility of amorphous filler provides simplicity to joining process.
The fracture shear strength of the joints was measured at room temperature. The strength of the joints brazed at 1323 K for 1.8 ks using Cu₅₀Ti₅₀ filler possesses 138.3 M Pa at room temperature. The strength of the joints at constant brazing time raises with an increase in brazing temperature. The strength increases slightly with an increase in titanium content in the filler. At 1298 K for 1.8 ks, the strength of joints at room temperature changes from 107.8 M Pa with Cu₆₆Ti₃₄ to 120.6 M Pa with Cu₄₃Ti₅₇. The strength of the joints was also investigated at high temperatures up to 973 K. The high temperature shear strength is almost constant up to 773 K, and deceases gradually at higher temperature.
Isothermal solidification process takes place during joining, and copper dissoves into Cu-Ti filler metal, and then, copper solid solution containing Ti precipitates from Cu-Ti filler metal. On the other hand, the formation of titanium oxide TiO_x and (Al, Ti)₂O₃ solid solution oxide is attributable to the strong joining between alumina and Cu-Ti filler.

The brazed joints are generally weak to impact loads; this results in the low reliability to the mechanical properties of the joints. Therefore, methods to increase the impact strength of carbon steel braze joints by utilizing dissolution and deposit of base metal was studied. The results were as follows:
(1) Charpy U-notch (GUN) impact strength of the low carbon steel (0.17%C)/high carbon steel (1.04%C) joint brazed with Cu filler metal is increased remarkably from 2.9 J/cm² to 27.2 J/cm² at room temperature, when both base metals are linked with the deposited columnar Fe-9 to 12%Cu-1 to 2%C alloy.
(2) When using Deposit-Filler metal (which consists of high carbon steel foil plated with Cu of 10μm thickness on both sides) for similar low carbon steel joint, CUN impact strength of the joint is advanced 89.4 J/cm² because of linking both base metals with the columnar phase through the foil.
(3) Due to carburizing-effective case depth of 0.2 mm-one of base metals prior to Cu brazing for similar low carbon steel joint, CUN impact strength of the joint is attained 89.2 J/cm² by linking, since the columnar phase deposits from the carburized low carbon steel.

Surface temperature near the rear end part of molten pool in girth TIG welding for 2 mm thick, 25.4 mm diameter stainless steel tube was measured with a PbS infrared radiation pyrometer. Effects of welding position and distance from arc center to a measuring point on the relation between the surface temperature and the reverse side bead width were investigated at down-hand, vertical down, over head, vertical up welding position of orbital TIG welding.
Temperature near the rear end, part of molten pool becomes high in order down-hand, vertical up, vertical down, over head welding position under the same reverse side bead width condition. Mutual correlation exists at each welding position between surface temperature at the position 5-10 mm away from the center of the arc and the reverse side bead width. It indicates the possibility of adaptive control of the reverse side bead width by monitoring the temperature near the rear end part of molten pool during the TIG welding.

An experimental model of a reverse side bead width control system which employs a fiber-optic thermal sensor and a microcomputer has been developed. The system is worked by monitoring infrared radiation emitted from the surface of the workpiece. In this system, infrared rays emitted from the surface of the weld are picked up and transmitted via optical fiber to a photodiode which converts them into electrical signals. The welding current is controlled by an 8-bit microcomputer in accordance with fluctuations in these signals so as to ensure the optimum surface temperature is obtained for a high quality reverse side bead.

This paper reports the possibility of detecting bead deflection defects during electron beam welding (EBW) using in-process monitoring by ultrasonics. Experiments were conducted using a 6 kW (150 kV) high vacuum type welding machine and a 5 MHz ultrasonic transducer. All welding was performed at a 30 mm penetration depth in the flat position on SM41 steel plates. A coupling method of the ultrasonic transducer to the plate at a high temperature in the vacuum chamber was investigated. A 1 mm gap was maintained by using 0-rings between the transducer and the steel plate surface and silicon oil was pumped into the gap for coupling and cooling.
The following points were concluded from the study.
(1) It is possible to monitor the bead deflection during EBW by detecting the time difference between the ultrasonic echo from the weld metal and that from the groove face.
(2) The detected voltage of the bead deflection is proportional to the measured bead deflection and the accuracy of detection is within ±10%.
(3) A bead deflection defect as small as 0.6 mm can be detected.
(4) The maximum echo height from the weld metal is obtained using the following conditions:
(a) the measurement depth equals the maximum depth of penetration,
(b) the measurement distance is 15 mm behind the weld cavity.
(5) The echoes from the weld metal are those arising from the root defects which occur at the bottom of the weld zone.

A numerical algorithm for the time-independent Stefan problem is presented. The heat conducting plate is assumed to be infinitely broad and the heat source of constant intensity is assumed to move with a constant velocity in the plate.
The shape of molten pool is determined by correcting successively the provisional shape of the molten pool given beforehand until it satisfies the Stefan condition. The isotherms showing the temperature field are determined successively by starting from the obtained fusion boundary. The algorithm is given for the two-dimensional case and for the axi-symmetric three-dimensional case. Some examples of the results are given for the case of a linear heat source and a point heat source of various intensities.

The influence of δ-ferrite content on notch toughness and elastic-plastic fracture toughness of Type 316L austenitic stainless steel weld metal was investigated. The weld metals were prepared by deposition of single pass weld using submerged arc welding. The amount of δ-ferrite was changed in the range of 0.3 to 11.1 % by addition of nickel and chromium.
In notch toughness test of as-welded specimens at 77 and 103 K, there was no simple correlation between the notch toughness and δ-ferrite content, differing from the results of other previous studies relative to this research. In more than about 3%δ-ferrite, the notch toughness increases with δ-ferrite content and reaches a maximum values at about 7%δ-ferrite.
The solidification structure indicated that in the 3%δ-ferrite specimen of minimum notch toughness, an eutectic δ-ferrite appeared and furthermore with increasing δ-ferrite content, the morphology alters to vermicular and then lathy ferrite. From these results, it was evident that the δ-ferrite morphology considerably influenced to the notch toughness.
On the other hand, the measurement of elastic-plastic fracture toughness Jrc, critical J and dJ/da revealed that these values related reasonably to the notch toughness.
In heat treating the specimens at 923 K for 100 h, the absorbed energy decreased dramatically with increasing prior δ-ferrite content. This was considered because δ-ferrite changed to σ phase and carbide by heat treating.

Effect of nitrogen addition on notch toughness of 2 1/4Cr-1Mo submerged arc weld metals was investigated. Nitrogen makes the grain size of weld metal smaller, and lowers the transition temperature of weld metal after stress-relieving and also after step-cooling. In stress-relieved weld metal, abundant and homogeneously distributed CrN precipitates are found, in addition to Mo₂C, M₇C₃ and M₂₃C₆, by means of TEM observation of extracted carbon replica. Larger reduction of hardness after PWHT results in lower vTr₄₀. From the results, the mechanism on notch toughness improvement in nitrogen added 2 1/4Cr-lMo weld metal is considered as follows. Nitrogen in weld metal makes the transformation temperature low and the grain size fine. The weld metal in as-welded condition is hard and brittle because it contains much nitrogen as solute atom. However the weld metal is softened by PWHT, and the softening is promoted by precipitation of CrN, in addition to Mo₂C, M₇C₃ and M₂₃C₆. Consequently, the notch toughness improvement is achieved by both the fine grain size and relatively soft microstructure.

It was previously reported that there was possibility for excellent improvement of the creep strength of TIG-welded joints by using the filler metals contained minor elements such as B, Zr and REM. The twenty-two laboratory heats of Hastelloy X filler metals contained minor elements of various amount lebels were produced for the weldability studies in this work, and mechanical properties of weld metals and weld crack sensitivity by TIG-welding method were examined. The main results are as follows:
1. A ductility of weld metals by using the filler metals contained minor elements was more excellent than that by using the filler metals without any minor element on the tensile test at elevated temperature.
2. Some fissures were observed in weld metal on the fissures bend test for welded joints by exceeding a suitable amount lebel of addition of minor element.
3. The filler metals contained minor element did not so much influenced on the weld hot-crack sensitivity, and there was no trouble in the practical welding performance using such filler metals.
4. A possibility was found that remarkable improvement of the creep strength of Hastelloy X TIG-welded joints was performed by using the filler metals contained minor elements, but its amount of addition had to be controlled sufficiently to keep the good weldability of Hastelloy X.

To improve the creep property of Hastelloy X TIG welded joints, some filler metals were developed with addition of several minor elements, such as, boron, rare earth metals and zirconium. The weld metals and weld joints obtained by the developed filler metals were subjected to the uniaxial creep testing at 900°C. Some of them showed excellent creep strength and rupture ductility. Among of the additions in the filler metals, boron was most effective for improvement of creep property, rare earth metal fairly effective, but zirconium little effective. Filler metals selected by the uniaxial creep testing were applied to weld the cylindrical specimens circumferentially and longitudinally, the weld metals were subjected to the creep testing with internal pressure at 900°C to investigates the integrities of the welded structure. As a result, the creep rupture life of some welded cylinders was longer than that of non-welded cylinder.

The local hydrogen concentration in the heat-affected-zone where the root cracking occurred has been investigated using the finite difference method. The effect of groove and weld metal shapes on the local hydrogen concentration was basically examined. It is clear that the local hydrogen related with the cold cracking is not determined by the time factor, but by the thermal factor, namely, throat depth does not become a main factor. The root face, the root angle and the bond line shape are important factors of the local hydrogen concentration.

Various forms of microcrackings are found in mild steel weld metal under same limited circumstances. A kind of microcrackings occurs on the surface after grinder dressing in the fillet weld of the lowest course shell plate and the annular plate in Oil Storage Tank. Compressive stress by grit blasting is effective for prevention of these crackings. Many efforts are made on the determination of optimum blasting conditions. They are as follows.
Exhaust pressure: 0.49MPa
Working speed : 1.5m/min
Grit : MGH30
Grinder dressing and grit blasting under optimum conditions are conducted on the surface of weld metal after welding. Microcrack susceptibility by MT of the latter is less than that of the former. Blasting is effective for prevention of microcrackings.

In welding for sealing purpose, blowhole are frequently formed by a small rise of the pressure within a space sealed by welding. The critical pressure for blowhole formation was studied in relation to the clearance in weld joint.
The clearance is simulated by the narrow hole provided in a block specimen. The inner space was filled with a gas and the outside of the hole was melted. The results are summarized as follows:
(1) The critical pressure for blowhole formation is determined by the apparent surface tension of the molten metal and the clearance in a weld joint.
(2) The relation between the critical pressure (Pcr) for blowhole formation and the diameter (D) of the gas hole was given by the equations,
Pcr-54.4=534/D (for air)
Pcr+192 =592/D (for argon)
(3) In welding for sealing purpose, it is effective for prevention of blowhole formation to decrease the pressure in a sealing space and to minimize the clearance in a weld joint.

In this study, the effect of rare earth metals (REM) addition on the solidification cracking susceptibility of electron beam weld metals in fully austenitic Type 316 stainless steel was researched.
The cracking susceptibility of this steel is put in order by REM and P or (P+S) content. Adding proper content of REM (about 0.25%REM for 0.025%P steel) is effective to improve the cracking susceptibility of this steel. However, the cracking susceptibility of this steel is deteriorated again by adding excess content of REM (about 0.37%REM for 0.025%P steel).
Free P and S atoms being fixed as phosphides, oxyphosphides, sulfides and oxysulfides by REM atoms is the main reason why the cracking susceptibility of this steel is improved by adding REM.

It was found out that the first layer overlay weld metal with an austenite/martensite duplex microstructure provided the overlay welds with between hydrogen-induced disbonding resistance than that with the conventional austenitic microstructure containing several percent ferrite.
The mechanical properties of the disbonding resistant overlay weld was studied in the previous report.
The purpose of the present study is to determine the metallurgical properties of the overlay weld with the 15%Cr-8%Ni duplex weld metal. The main results obtained in this study are as follows:
1) The Actl and Ac3 temperatures of the duplex structure are 510°C and 630°C, respectively.
2) The duplex structure consists of austenite and martensite after regular post-weld heat treatment (PWHT) at 700°C for 20 hours.
3) Coarse planar grains, which may affect the disbonding resistance of overlay weld, were scarcely observed in the overlay weld metal adjacent to the bond between the duplex weld metal and the base metal.
4) The width of carburized zone in the duplex weld metal is almost equal to that of the conventional weld metal after PWHT.
5) The width of transition zone of Cr and Ni concentrations at the bond of the weld with the duplex microstructure is larger than that of the weld with the conventional austenitic microstructure.

Susceptibility of stress corrosion cracking (SCC) of duplex stainless steel, SUS329J1 was investigated in 42%MgCl₂.aq. sol. (416 K). Results were as follows:
Effect of structural anisotropy for SCC were studied. It was made clear that SCC was more susceptible in the case of crack growth being parallel to main rolling direction than in other cases and crack tend to propagate along α phase other than γ phase. In addition, SCC susceptibility was remarkably high in welded plates, although 329J1 base metal had high resistance to SCC. This would be caused that the residual stress were higher for its high yielding strength of base metal.

A study has been performed to determine the effect of boron on the pitting corrosion resistance in HAZ of 25Cr-6Ni-2Mo duplex stainless steel. The effect of boron on the hot ductility of this steel has also been investigated. Boron had an effect to improve the pitting corrosion resistance of HAZ, but conversely to deteriorate it with being added more than about 100 ppm. Scanning electron microscopic observation for the specimens corrosion-tested for various periods demonstrated that the most susceptible site for the pit initiation in HAZ of the boron free alloy was the lamellar phase which consisted of the cellular M₂₃C₆ and the austenite, precipitated on the austenite/ferrite boundary during welding. The austenite existing among the cellular M₂₃C₆ in a lamellar phase was depleted in chromium content, which resulted in the preferential attack of this phase in a corrosion environment. Boron addition made the volume ratio of M₂₃C₆ to the austenite in a lamellar phase to decrease, resulting in the increase of the chromium content of the austenite in a lamellar phase to the level more than that required for passivation (12wt%). Consequently, the lamellar phase in the boron containing steel became more resistant, thus the pitting corrosion resistance in HAZ of the steel was improved. Hot ductility of 25Cr-6Ni-2Mo duplex stainless steel was also improved by addition of boron, especially for the alloy with high nitrogen content which had essentially the shortness of the hot workability.

The present study examines the mechanism of bending distortion during multi-layer welding of restrained butt joint in aluminium heavy plate. The experiments were performed to observe the reaction force, bending moment, deflection of plate and strain distribution during or after welding. Moreover, a theoretical consideration based on the theory of plates was applied to these results and effects of factor on the bending distortion were discussed. The main results are as follows.
The transient deflection during the welding of each pass changes according to the bending moment roughly in a such way that the deflection mode is concave at weld side during heating cycle and convex during cooling cycle. The bending distortion depends upon the reaction force, its eccentricity from the neutral plane of base plate and the plastic deformation. The effects of reaction force and of its eccentricity are remarkable for the first passes and the effect of plastic deformation unevenly distributed in the thickness direction appears for later passes. The bending deflection of plate decreases as the throat thickness increases and as the restrained length decreases. The plastic strain results from the restraint of thermal expansion of plate and from the bending deflection during welding, and its distribution is nearly linear from top face to bottom face. By applying the theory of plate to these specimens, the bending moment and inherent angular deformation of throat are calculated from the measured value of the reaction force, its eccentricity and bending moment of the welded plate, and the effects of eccentricity and inherent angular deformation upon the bending moment of the throat are predicted.

The transient thermal stresses of 13 Cr cast steels (SCS5) and structural carbon steels (SM41B) are investigated by round bar specimens of which both ends are fixed and subjected to thermal prestraining cycles to simulate welding process. The multi-thermal prestraining cycles tests are conducted in the case of fixed peak temperature (600°C and 1000°C) and increasing peak temperature (200°C-1000°C) and decreasing peak temperature (1000-200°C), respectively. The experimental results indicate that the effects of transformation on residual stress of SCS5 steels are remarkable, and the effects of multi-thermal prestraining cycles on residual stress are negligible in both SCS5 and SM41B. These results are compared with several weldments. For example, three dimensional residual stress distributions of 13 Cr cast steel weldment prepared with homogeneous weld metal are measured by Ueda-Fukuda's method. It is observed that the maximum compressive residual stress is -210 MPa at the bottom of the final weld pass and the maximum tensile residual stress is 310 MPa at the region out of the HAZ. It is shown that the both-ends fixed bar test subjected to multi-thermal cycles of incremental steps is available to simply estimate the residual stress characteristics.

The use of electrodes with a nominal strength lower than that of base metal has been considered to be effective in preventing weld cracks, although the use of electrodes with a nominal strength higher than that of base metal is in common. K. Satoh and M. Toyoda have analyzed the tensile behavior of undermatching welded joint and have demonstrated the appropriateness of this joint to heavy plates of HT 80 steel.
However, heavy plates of HT 60 steel are more common in heavy electric industrial structures, so that the strength of undermatching welded joint of HT 60 steel has been investigated in this paper. The results of investigations are summarized as follows:
(1) Undermatching welded joint of wide and heavy plates of HT 60 steel, where the weld metal has appropriate tensile strength, fractures in the base plate and exhibits significant ductility under tensile load even if the weld reinforcement is machined out, according to the increase in tensile strength of the joint caused by the hardening area in heat affected zone.
(2) Cylindrical pressure vessel has bursted at base plate region about 10 nun apart from the toe of the welded joint in the axial direction and it has been observed that the stress-strain curve for the weld metal of the joint locates in higher stress-strain region than that for the weld metal itself and it seems to approach to that for the base metal as the strain becomes larger.

The purpose of this study is to make clear the influence of the weld surface irregularity induced by the weld stop-start operation on the fatigue strength of longitudinal fillet welded joint. The materials used are 500 MPa and 600 MPa class steels. Weldings are done by manual arc processes and weld stop-start operations are carried out as leaving or filling the crater. The specimen is a flat plate with 16 mm thick and 60 mm wide. After the weld surface irregularities are measured, the specimens are tested under axial cyclic loadings. Then, the fatigue crack initiation and propagation behavior is closely observed. The test results indicate that the weld stop-start operation as filling the crater is preferable with respect to the fatigue strength.

Fatigue crack growth rates and threshold for crack propagation were determined for CCT and SEN specimens taken from the butt welded joints in 600 MPa grade steel. The crack opening stress has also been determined by using relation between the elastic compliance and loads. The compressive residual stresses resulted in the decrease of fatigue crack growth rates and the increase of threshold values. These behavior strongly depended on the stress ratio. While, the tensile residual stresses resulted in the increase of fatigue crack growth rates and the decrease of threshold values. These behavior scarcely depended on the stress ratio. The effects of welded residual stresses and stress ratio can be correlated by using the effective stress intensity range concept.

Crack Tip Opening Displacement (CTOD) tests were carried out to evaluate the Heat Affected Zone (HAZ) toughness. The present study was focused on the main factors to affect the critical CTOD values in the HAZ from viewpoint of microstructures.
It was observed that the fracture initiation properties of HAZ were related to the properties of the Local Brittle Zone (LBZ) near fusion line. The relationship between them was investigated from the viewpoint of austenite grain size, fracture facet size and fraction of Martensite-Austenite (M-A) constituent in the LBZ.
The results obtained were as follows:
(1) The crack initiation properties were strongly affected by the microstructures of the HAZ.
(2) The LBZ deteriorated the critical CTOD values in the HAZ because of both the coarsened grain size and existing M-A constituent.
(3) The fine ferritic microstructures improve the critical CTOD values in the HAZ, through reducing the fraction of M-A constituent.
(4) A good correlation was found between the critical CTOD values in the HAZ and fracture facet size, although the effect of fracture facet size on the critical CTOD values in the HAZ was different for the fraction of M-A constituent.
(5) The close correlation was found between the critical CTOD values in the HAZ and the fraction of M-A constituent.
(6) Griffith crack was observed at M-A constituent and propagated along M-A constituent/matrix interfaces, resulting in low critical CTOD values.

Correlation between Charpy absorbed energy and critical COD is investigated to get a useful method for estimating critical COD from Charpy V data. Round bar tension test, Charpy V impact test and static 3-point bending test with fatigue notched specimen are carried out using mild steel, 80 kgf/mm² grade high strength steel and A5083 aluminum alloy.
Correlations are found between W'c/σY and δc, as well as between EW'c/σ₂Y and Eδc/σY, where W'c is Charpy absorbed energy obtained by considering temperature difference between Charpy transition temperature and COD transition temperature. The symbols σY, δc and E are yield strength, critical COD and Young's modulus respectively. The correlations are established for various kinds of metals and over a wide temperature range including not only upper shelf range but also transition range.

The welding method for getting the joint performance capable of forming was investigated and the formability tests of the butt welded sheets in aluminium and its alloy were carried out.
Machining and wire-brushing of joints are effective to obtain the joint performance capable of forming. Especially, DCEN TIG welding method is more effective for this purpose when using helium shielding gas and shorter arc length. In the case of Al100P-O, Al100P-1/2H and A5052P-O, the formability of their butt welded sheet was improved up to nearly equal that of base material. But in the case of A5052-1/2H, softening by welding was so great that the formability of welded sheet was changed by width of softened zone.
In the case of severe forming, the formability can be improved by using the welded blank.

Among several hermetically sealing methods, we studied experimentally on Micro-Parallel Seam Joining (MPSJ) used for hermetic seal of high reliable ceramic package for semiconductor devices, particularly paying attention to the relationship between plated materials on the package and lid and its seam joining characteristics to verify the MPSJ phenomena.
Results obtained are summarized as follows.
(1) Cracks were observed on the ceramic substrate immediately after seam joining, that was considered to be caused by the difference in thermal expansion coefficient in the case where austenitic stainless steel, SUS-316 lid was used. Therefore, it will be required to use Fe-Ni-Co Alloy, KOVAR as a lid material which has a similar thermal expansion coefficient with the ceramic substrate (Al₂O₃).
(2) Seam joined area was consisted of both a mutually diffused layer at the joining boundary formed by plating material for the lid and seal frame and the fillet formed around the circumferential edge of the lid. No molten layer of both base materials existed.
(3) The main reason for the paragraph (2) is the fact that density of heat generation at the joining boundary will decrease because of current path area expands due to softening and fusion of plating material concurrent with the current flow. Therefore, both the contact resistance between the lid and roller electrodes and inherent resistance of the lid will be the majority for heat generation source. Consequently, lid and seal frame will be joined in the form of brazing under pressure of plating material.
(4) From the above results, it has been verified that plating material for the lid and seal frame plays an important role to secure specified hermeticity in the case of hermetic sealing of ceramic packages by MPSJ.

A CO₂ laser has been applied to fusion-welding and surface-melting of sintered zirconia ceramics fully stabilized with CaO, MgO or Y₂O₃, and chemical compositions and preheating temperature required for preventing cracks have been studied. The results obtained are summarized as follows:
(1) Although macroscopic crack in the base material is prevented at preheating temperatures higher than 1200°C, intergranular microcrack has been developed in the fusion zone in the zirconia ceramics other than CaO-stabilized zirconia containing 1.7%SiO₂, 1%MgO and 0.7%Al₂O₃ as impurities. A low melting point compound of SiO₂-MgO-CaO-Al₂O₃ formed at grain boundary plays an important role to release the shrinkage stress in the fusion zone, producing sound bead.
(2) The weld joint strength is decreased by porosities grown at HAZ at lower welding speeds, and at the bead center at higher welding speeds. As the welding speed becomes faster, both porosities are sup-pressed to grow, increasing weld joint efficiency up to 85-90%. Laser-melted surface layer has shown almost no porosities, providing bending strength and a Weibull modulus which are equal to or somewhat larger than those of the base material.

A new bonding technique of copper and SiC ceramics having high thermal conductivity was developed, and thermal characteristics of the bonded elements were evaluated.
SiC ceramics used in the experiments were hotpress-sintered material containing a small amount of BeO. It has high thermal conductivity (270W/m·k) and also high electrical resistivity (10¹³Ω·cm or above).
It was concluded from this study that,
(1) 50 mm×50 mm square SiC blocks were successfully bonded to copper blocks within 20 second without cracking by using both Cu-35wt%Mn foil as a brazing material and the copper-35vol% carbon fiber composite material of 2 mm in thickness as the compliant material to relieve the thermal stress.
(2) Manganese silicides (Mn₆Si) and manganese carbides (Mn₂₃C₆) were observed in the SiC/Cu bonded zone by using Cu-35wt% Mn brazing foil.
(3) Peel strength between a SiC block and 100μm thick copper foil was above 2 kgf/mm and fracture occurred in copper foil.
(4) Tensile strength of a bonded element of SiC block/Cu-35vol% carbon fiber compliant material/ copper block was about 2 kgf/mm². Fracture occurred in the composite material. The element can be heat resistant and used up to 300°C.
(5) Thermal reliability of the bonded element (using a 35 mm square SiC block) was evaluated by normal thermal cycle test and rapid thermal cycle test using a 1.5 MW splash test machine. Maximu allowable heat fluxes were about 500 W/cm² and 400 W/cm² respectively.