In order to investigate effects of metal vapor on the plasma state in welding process, electron temperature in helium Gas Tungsten Arc(GTA) plasma during welding was measured by using the laser scattering method. Furthermore, density distributions of metal ions were also measured by using the spectroscopic analysis. Our results showed that plasma during welding consisted of two plasma regions, namely, pure helium plasma region and quasi-metal plasma region.
Numerical analysis of bonded area at diffusion bonding was indicated by Takahashi at Quarterly Journal of the JWS (1985). Relationship between percent bonded area by this analysis and joint strength for diffusion bonding of ultra fine machined copper was evaluated. Bonding parameters are low temperature (473∼973 K) and extensive low pressure (0.005∼0.15 MPa). Details of bonding test were reported by authors at Quarterly Journal of the JWS (1997).Main results are as follows: (1) It was confirmed that percent bonded area with numerical analysis proposed by Takahashi were nearly equivalent to the ratio of joint strength/copper base metal strength. The ratio of joint strength/copper base metal strength is 95∼100% when percent bonded area with numerical analysis is 100% and 7∼14% over than percent bonded area with numerical analysis at another case. (2) This numerical analysis is effective for practical use at diffusion bonding of ultra fine machined copper by low bonding temperature and pressure.
Influence of oxide film, such as laser cutting edge or mill-scale, on CO2 laser welding phenomena and weld characteristics of mild steel was investigated. Laser welding of butt joints, which were different in location and thickness of oxide film, were performed. Examinations were made on the weld characteristics; which were bead geometry, defect and microstructure. Light emitted from laser induced plasma was also analyzed to understand welding phenomena. The results of these experiments lead to understanding that oxide film especially on joint face causes substantial spatters, under-filled surface, disappearance of nail-head, porosity formation and reduction of hardness. It is considered that these are attributed to oxygen which reduces surface tension of molten metal and react with carbon, manganese and silicon. These results suggest that reduction of oxide film thickness or killing oxygen activity by supplementary deoxidizing elements may produce acceptable weld.
High-power CO2 laser welding of A5083 and A6N01 aluminum alloys was performed with filler wires, and the optimum laser wire welding conditions were estimated by comparing surface appearances and fusion zone geometries of weld beads made at different wire feeding rates, welding speeds, shielding gas types and feeding directions. The high-speed shadowgraph imaging technique was applied to ascertain filler wire melting dynamics during laser welding under different conditions. A comparison of the surface appearances of weld beads produced in two wire feeding directions, i.e., leading-wire-feed and lagging-wire-feed, shows that a smooth surface can be obtained with the former. Sound beads were also attained at a relatively long defocusing distance and a high wire feeding rate, under the conditions in which the frequency of weld pool boiling decreased. Parametric experiments revealed that adding filler wire to the weld pool made A6N01 aluminum alloy more weldable than no use of a filler wire. Shadowgraphic observation of wire melting showed that humping beads were formed when large metal droplets were periodically supplied to the weld pool. It was also found that a shielding gas exerted a great influence on the melting dynamics. Especially, the type of the shielding gas affected weldability in terms of the maximum feeding rate and location tolerance of a wire. The highest weldability was obtained using N2 gas, which is attributed to the higher surface absorption of AlN film produced on the molten pool surface. When Ar or He shielding gas was used, a wine-cup-shaped fusion zone was produced, but when N2 shielding gas was used a bobbin-shaped fusion zone was formed. These results suggest that from the viewpoint of melting ability, N2 shielding gas is preferable in the case of CO2 laser welding of aluminum alloys using a filler wire.
Porosity is easily formed in a keyhole-type of deeply penetrated laser weld beads. In this study, therefore, high power CO2 laser welding utilizing a single spot or a tandem twin-spot beam was performed on thick plates of Type 304 steel with the objectives of investigating feasibility of porosity prevention and establishing the optimum conditions in tandem twin-spot beam welding. Porosity formation tendency was first investigated in a partial penetration weld made with a single-spot CO2 laser beam at low speed. Consequently, it was confirmed that many pores were present near the bottom parts of the weld beads produced without and with assist gas at a power exceeding 10 and 15 kW, respectively. On the other hand, it was revealed that the number of pores was drastically reduced by twin-spot laser beam welding. Especially, some weld beads produced at 30 kW indicated the best grade of the inspection test result according to JIS Z3106. Subsequently, the reason for the reduced porosity was investigated by using the microfocused X-ray transmission in-situ observation system with a high speed video camera. During high-power laser welding of Type 304 with the single-spot beam, it was confirmed that the keyhole fluctuated up and down and consequently a lot of bubbles were generated from the keyhole tip near the bottom part of the molten pool, leading to the formation of porosity. On the other hand, when an enlarged keyhole was produced by a twin-spot beam, few bubbles were generated from the keyhole tip. Moreover, even if the bubbles were formed, most of them did not move away from the formation location and were soon absorbed in the keyhole. This is interpreted by considering that no strong melt flow was induced near the bottom part of the molten pool just behind the tip of the widened keyhole. Consequently, the difference of porosity formation between single-spot and twin-spot laser beam welding is attributed to the differences of their bubble formation mechanisms, liquid flow near the bottom part of the molten pool, and the bubble disappearance via the keyhole.
In previous reports, an ultra-narrow gap CO2 gas metal arc welding process has been developed in order to produce excellent welded joints. In the ultra-narrow gap (less than 5 mm gap) welding process, as the welding wire melting behavior was controlled by low frequency pulsated arc current waveform, the arc generating at the wire tip was widely oscillated in the direction of thickness and a good weld bead with large throat thickness was formed. In such a higher efficient welding process, the strong arc force of CO2 arc is essential to obtain sound beads without lack of fusion at a groove bottom, however, the oxygen content of weld metal in CO2 arc welding has to be reduced to obtain high-toughness at weld metal. In this report, a newly welding process is proposed in which an arc is widely oscillated along groove walls and oxygen content of welded metal is reduced as periodically controlling composition of CO2 gas in shielding gas. As an arc length between the wire tip and groove wall is held constant in ultra-narrow I type gap joint, arc current at stable operating point of CO2 arc essentially decreases less than that at stable operating point of Ar arc in constant potential characteristics of power source. It was confirmed from the numerical simulations on both wire melting and electric welding circuit that pulsated current waveform was produced as periodically alternating CO2 arc with MIG arc. Using the developed welding torch locally introducing CO2 gas into Ar+2%O2 shielding gas, the welding was carried out under V-type groove with less than 30 degree in the bevel angle. It was shown that the arc was widely oscillated along V groove walls and sound bead was formed in one side welding. Furthermore, it was confirmed that the oxygen content of weld metal in the proposed process was reduced to the level of that in MIG arc welding, and absorbed energy of weld metal was the same level as the MIG arc welding.
GMA welding process with synchronized control system consisting of a pulsed arc current waveform with cyclical variation of wire feed rate has been developed on the base of computer simulation. In the ultra-narrow gap (less than 5 mm gap) welding process, as the welding wire melting behavior is controlled by low frequency pulsated arc current waveform, the arc generating at the wire tip can be widely oscillated in the direction of thickness. The arc heat pole is dominated by wire melting tip position in consumable electrode welding process, and the arc heat distribution can be changed by this wire tip behavior. The wire melting tip position is represented by wire extension which is a function of wire feed rate and wire melting rate. In order to more-adequately control the arc heat distribution, this report describes synchronized control process between wire feed rate and pulse current waveform in the ultra-narrow gap GMA welding. The important control factor in the newly proposed process is phase shift θ between pulse current waveform and wire feed rate pattern. Firstly, the non-steady state simulation of the wire melting behavior is developed in the case that both arc current and wire feed rate change periodically with low frequency. From simulation results of wire tip behavior under various phase sifts and wire feed rates and pulse conditions, the adaptive welding conditions can be found effectively. Secondly, the synchronized control of welding was carried out under adaptive welding conditions. From the experimental results, it was verified that the arc heat distribution was effectively controlled with the phase shift method expected by numerical simulation thus adequate penetration shapes were obtained in the proposed welding process.
The sizes of bead and weld bead appearance of weldments are expected to depend on the skills of welder operating such as semi-automatic welding machine. The visual examination is probably the one most widely used to evaluate the skills of welding operatives. The visual examination can appropriately judge the weld bead appearance, but this examination has the problems such as objectivity and unevenness of evaluate by differences among individuals. It is therefore desirable to understand the statistical properties of bead sizes such as bead width, bead height and bead appearance. This paper describes the statistical properties of sizes of bead width and bead height in butt welding joints with backing plates which are welded by welding operators having the specified welding skills using the semi-automatic MAG welding machine. The results obtained may be summarized as follows: (1) The distribution properties of mean bead width, maximum bead width and maximum difference in bead width are clarified, and they conform to a normal distribution. (2) The distributions of mean bead height, maximum bead height and maximum difference in bead height are made clear and also they adjust to a normal distribution. (3) 95% confidential intervals of mean value of populations such as mean bead width, and mean bead height in butt welding joints with backing plate welded by welding operatives having the license SN-2F (JIS Z3841) using MAG welding machine.
The hot cracking sensitivity was examined on a series of weld metals of which the carbon content was varied systematically from 0.07 to 0.95%. The cracking sensitivity of weld metal was compared by the cracking test with a self-restraint type specimen. A weld metal was deposited in a circular groove of 80mm in diameter by MAG welding. Weld metals of various carbon contents were obtained by combining each of twelve electrode wires (0.004 to 1.0%C) and three base metals (0.23 to 0.77%C). The cracking sensitivity of each weld metal was expressed by the total length of cracking. The cracking sensitivity increases abruptly when the carbon content of weld metal exceeds 0.20%. It increases to the maximum at 0.40%C and then reduces to the minimum at 0.65%C. This result quite differs from the generally accepted idea that the more is the carbon content of weld metal, the greater becomes its cracking sensitivity. This result was referred to the phase diagram of Fe-C-0.5%Si system in which the peritectic reaction; delta-ferrite+liquid→austenite takes place. For the alloys of 0.20 to 0.65%C, the liquid remains after the reaction. Phosphorus once dissolved in delta-ferrite will be rejected into the remaining liquid. It lowers the final freezing temperature of remaining liquid locating between the dendrites and increases the cracking sensitivity of weld metal. This action is likely to arise most intensively in the alloy of 0.4%C and least in that of 0.65%. Sulfur seems to assist the harmful effect of phosphorus.
Hardness distribution in heat affected zone (HAZ) of boron bearing 780 MPa grade high tensile strength steel was investigated. Welded joint was produced by submerge arc welding with 4.5 kJ/mm heat input. The Vickers hardness was 370 on HAZ 2mm distant from fusion line. The Vickers hardness decreased ti 300 on fusion line. In simulated HAZ test, raising the temperature from 1273 K to 1623 K decreased the Vickers hardness from 340 to 280. These results show that simulated HAZ test can simulate hardness distribution in HAZ in the welded joint. In simulated HAZ test at 1273 K, increasing the holding time from 5 seconds to 3600 seconds decreased the Vickers hardness from 370 to 300, although increasing the holding time at 1623 K did not change the hardness. The cause of the hardness change in simulated HAZ test can be explained in the following. McLean’s theory about equilibrium segregation shows that the raising the temperature decreases boron content at grain boundaries. In addition, rapid grain growth during heating causes a concentration of boron at grain boundaries (called dragging effect). In the area heated up to 1273 K, boron concentrates at grain boundaries over equilibrium amount by the dragging effect. After grain growth, boron content at grain boundaries decreases to equilibrium amount at 1273 K by diffusion. On the other hand, the area heated to 1623 K has a lower boron content than 1273 K at grain boundaries. The reason for this is that the high diffusive rate of boron cancels the dragging effect, and boron content decreases to the equilibrium amount at 1623 K lower than 1273 K. It is concluded that the Vickers hardness of 300 on fusion line increased to 370 at HAZ 2 mm distant from the fusion line, and this hardness distribution can be explained by the change of boron content at grain boundaries.
The application of the low-temperature transformation welding wire was introduced for the improvement of fatigue strength of boxing welded joints. This welding wire induces the compressive residual stress at the welded toe. In this study, the applicability of the low-temperature transformation welding wire was investigated to improve fatigue strength of non-load carrying type fillet welded joint. The measurements of residual stress and fatigue tests wetre carried out using test specimens. It is shown that the compressive residual stress at the welded toe is induced with the use of this wire. As a result, fatigue strength of welded joint is about twice lager than that of the convention joints.
A new method to evaluate the debonding behavior of thermal sprayed coatings is proposed. Type 304 stainless steel substrates were coated with 8% Y2O3-ZrO2 and 80Ni-20Cr by plasma spray. Then, GTA was applied as a moving thermal input to the substrate surface of a specimen while the surface strain of the coating was measured continuously by a non-contacting technique called the laser speckle strain meter. Strain curves were also recorded without thermal sprayed coatings and GTA current was varied from 40 to 120A. By comparing the strain behavior of the coating with that of a bare substrate, it was possible to determine the onset of debonding of the coatings. It was found that while 80Ni-20Cr coatings did not debond under the range of conditions tested, 8% Y2O3-ZrO2 coatings debonded when heat input was above a certain limit. With our method, it was possible to detect debonding beneath the coating surface, even when the debonding was not apparent on the coating surface, which was verified by mechanical testing afterwards.
In heavy industries, 304 austenitic stainless steel is the most popular material which is used for nuclear equipment, chemical vessels, vacuum vessels and so on. On the fabrication, not only a joint quality but also severe dimensional accuracy is required. To keep dimensional accuracy, considerable cost and efforts are requested, because the welding deformation of austenitic stainless steel is deeply depended on the physical properties of material itself. To decrease welding deformation, big jigs or water cooling method are commonly used which lead to the high cost. In general, the fusion welding by high energy density heat source results in less distortion. Today, laser welding technology has grown up to the stage that enables to weld thick plate with small deformation. The researches of welding deformation have been conducted intensively, but they are mainly concerned for arc welding, and studies for laser welding are very few. In this report, the authors will show the test results of deformation behavior in laser welding of 304 stainless steel. Also, they will discuss the deformation behavior comparing to that in arc welding. The main results of this study are as follows. 1. The angular distortion of laser welding can be unified by heat input parameter (Hp) which is used for arc welding deformation. 2. The angular distortion are same under the condition of Hp<6-9J/mm3 in spite of different welding method, however under the condition of Hp>6-9J/mm3 the angular distortion is quite different depending on the power density of welding method. 3. Pure angular distortion seemed to complete just after welding, but following longitudinal distortion took place for long period. 4. The critical value of longitudinal distortion can be estimated from heat input parameter. The transverse deformation can be also estimated by heat input parameter.
The purpose of this study is to weld long 304 stainless steel vacuum chambers without deformation. The requested accuracy is less than 1 mm per 4, 700 mm. To achieve this high accuracy welding, the authors compared suitable welding method among EBW, LBW and GTAW. According to this study, they decided that the LBW is the most appropriate method because of its highest welding speed and deep penetration ability that leads lower deformation. As there is less information of laser welding deformation, the present authors conducted some basic experiments on laser welding deformations of 304 stainless steel. In this report, following test results were obtained on the long chamber fabrication. 1. An equation of maximum deflection for 304 stainless steel laser welding is derived as δ=0.04[α·Q/(c·ρ)][L2·Z/I] 2. The maximum deflection has linear relationship to heat input Q (J/mm) and to the welding distance from neutral axis. 3. The compressed area of laser welding is fairy large compared to weld bead area and it expands almost all 10mm-plate thickness close to weld bead. 4. According to the information above, the authors could weld 4, 700 mm long vacuum chambers within the requested accuracy.
The bond strength and microstructure of a Cu (copper)/W (tungsten) joint friction-welded with an intermediate layer of Nb (niobium) foil 7-25 μm thick have been investigated in order to find proper thickness of the intermediate layer for obtaining a joint with high performances and slight microstructural changes in the weld. The tensile strength of the joint was increased more rapidly with the friction time, as the thickness of the intermediate layer was decreased. For all the joint obtained in this investigation, a number of Nb particles were incorporated into the Cu adjacent to the weld interface, forming a mixed region of Cu and Nb. The width of this mixed region and the number density of the involved Nb particle were decreased, as the intermediate layer thickness was decreased. In this mixed region and at the weld interface, significant amounts of porosities and uncontacted areas were observed at friction times shorter than 2 s, when the intermediate layer thickness was more than 15 μm. From these results, it can be concluded that in the Nb foil employed in this investigation thinner one is more favorable as an intermediate layer for the friction welding of copper to tungsten. The effect of the intermediate layer on the bond strength is also discussed considering the interaction of the Nb particle with W and O during the friction welding.
In order to make a thermal spray process highly controllable, it is essential to clarify the dominating factors of the process. Especially, the flattening behavior of the individual sprayed particle, which is the fundamental unit of the coating formation, has to be investigated precisely. In this research, the effect of substrate surface condition after preheating on the flattening behavior of the thermal sprayed particles was investigated. The surface roughness of the substrate before and after the pre-heating was measured by atomic force microscope (AFM), and the relation between the roughness and the splat morphology was investigated. The results obtained in this research are summarized as follows: (1) The surface roughness of each substrate material was increased by the preheating, and splashing seemed to be induced by the surface roughness. (2) The microstructure on the bottom surface of the splat changed from porous to dense with an increase of the substrate temperature. From this fact, it is indicated that the adsorbent on the substrate surface seems to affect the splashing.