In the case of CO2-O2 arc welding of rimmed steel, it was previously shown that the impact values were increased by addition of a certain amount of oxygen to the supplied gas. But if the electrode wire contained a large amount of non-metallic inclusions, the amounts of those in weld metals also increased and consequently, the impact values of weld metals decreased considerably. In this paper, to confirm this effect of oxygen in supplied gas, some experiments were carried out about impact values of weld metal of semi-killed steel, killed steel, 60 kg/mm2 high tensile strength steel and 3.5% Ni steel. As a result of experiments, impact values of weld metals of these steels were increased by addition of oxygen to the supplied gas up to about 25%. But if electrode wire contained a large amount of in-clusions, such tendency was not recognized. Such effect of oxygen in supplied gas on the fatigue strength of weld metal was also examined using various electrode wires and base metals. But the fatigue strength of weld metal was not considerably changed by addition of oxygen to the supplied gas up to 10-25%.
From the previous results reported by authors, it seems that the microstructure and the compositions of steels would exert a serious influence on the hydrogen embrittlement. In this report, first, the effects of the kinds and quantities of alloying elements in steel on the sensitivity of hydrogen embrittlement were studied. Results obtained are as follows: (I) In the case of Ti or Nb being the alloying element, it is confirmed that steels are embrittled remarkably, even if these elements are contained only in small quantity in steels. (2) In the case of Mo, Mn or Si being the alloying element, steels are hardly etnbrittled. (3) In general, it can be said that the sensitivity of hydrogen embrittlement depending upon the content of each alloying element is a similar tendency to the consolidation of ferrite matrix influenced by these elements forming solid solution in steel or iron. (4) Commercial chromium steels are observed to have less sensitivity of hydrogen embrittlement than mild steel. This phenomenon might be attributed not to the behavior of Cr but to the lower carbon content of Cr steel.
On weld cracking, especially on the cold cracking during cooling and after cooled to room temperature, intensity of restraint of weld joints will be one of the important factors. Research was made to investigate the relationship between the intensity of restraint and weld cracking under controlled contraction. In the previous report, a new test method named "Rigid Restraint Weld Cracking Test (RRC-test)" was developed for the first-pass weld cracking as related to external restraint of a weld joint, in this method the length of a specimen (or the restraining gauge length l) is kept constant during cooling (See Fig. 1). Critical intensity of restraint Kcr for weld cracking was obtained by this method. In this report, effect of plate thickness and welding heat input on the development of reaction stresses and the critcal intensity of restraint were investigated for the weldments of a mild steel (M) and a high strength steel of 80 kg/mm2 tensile strength level (H 8). The results obtained are as follows. 1) In the RRC-test, reaction force increases with cooling and the tendency of its development is almost similiar to the contraction process of weld joints during cooling, as related to the restraining gauge length, plate thickness and welding heat input (Figs. 2-4). 2) The magnitude of final reaction stress in weld (σw)t=∞, in the range that the intensity of restraint K is small and plate thickness h is larger than critical plate thickness hcr is proportionate to K independently of heat input and plate thickness (Eq. 8). The proportionate factor m is (3.9-5.0) × 10-2 in steel weldments (Eq. 9). As K-value increases, (σw)t=∞ approaches a constant value which is almost equal to the tensile strength of the weld metal. 3) The critical intensity of restraint Kcr for given materials and welding conditions can be obtained from the RRC-test results. Kcr of the mild steel weldments is independent of plate thickness and welding heat input in the case of h>hcr (Figs. 14 and 15). Kcr of the high strength steel weldments decreases with increase of plate thickness in the case of a constant heat input (Fig. 17). 4) The intensity of restraint had been measured for weld joints of actual structures such as ships as well as some weld cracking test specimens by one of the authors and others. These values were compared with the Kcr obtained. The posibilities of cracking in those weld joints were discussed (Fig. 18).
Weld crack is one of the most difficult problems encountered in welding processes of high strength steels. It is classified into heat affected zone crack and weld metal crack. This report clarified the correlation of chemical composition, hardness and structure of weld metals to weld metal cracks. The basic electrodes were prepared so that the main effects of C, Si, Mn, Ni, Cr, Mo and the interactions between C and the other elements for the weld metal cracks could be analyzed using orthogonal array table L16(215). Weld metal cracking tests were carried out with U groove cracking test specimens. From the results of the weld metal cracking tests, the carbon equivalent equations for the crack sensi-tivities of weld metals were calculated and the correlation between the values of the carbon equivalent equations and weld cracks was obtained. The conclusions are as follows: 1. The carbon equivalent equations for the crack sensitivities of weld metals are expressed by the following equations. Surface cracking percentage: +0×Cr ..................0.03%≤Cr≤0.53% Ceq(f)=C+0.14Si+0.10Mn+0.04Ni+0.11Mo+0.27(Cr-0.53) .........0.54%≤Cr≤1.49% Section cracking percentage: +0×Cr ................0.03%≤Cr≤0.53% Ceq(s)=C+0.15Si+0.11Mn+0.05Ni+0.12Mo+0.24(Cr-0.53) .........0.54%≤Cr≤1.49% Root cracking percentage: +0×Cr ................0.03%≤Cr≤0.53% Ceq(r)=C+0.15Si+0.12Mn+0.05Ni+0.13Mo+0.26(Cr-0.53) ........0.54%≤Cr≤1.49% 0.10%≤C50.18%, 0.12%≤Si≤0.53%, 0.70%≤Mn51.38% 0.03%≤Ni≤1.17%, 0.01%≤Mo≤0.49%, 0.03%≤Cr≤1.49% 2. It was confirmed after testing more than ten basic electrodes including commercial high tensile strength electrodes that the tendency of crack sensitivities of basic weld metals conld be well represented by the carbon equivalent equations as above mentioned. 3. In order to prevent the initiation of weld metal cracks, the upper limit of weld metal hardness must 270Hv. The chemical composition and cooling rate must also be selected so that ferrite may precipitate in the structure of weld metals.
Heat content of the droplet detaching from the wire tip in MIG arc welding is measured calorimetrically under a special electrodes arrangement shown in Fig. 1, 2. Ozawa already reported on this problem and we tested further to understand the temperature characteristics for wires of low and high conductivities. Fig. 4 shows the results for copper and aluminum wires and Fig. 6 those for steel and stainless steel wires. For aluminum wire, the heat content Qm expressed in cal/g increases as the current is increased in globular transfer range and reaches a nearly constant high value. The temperature is estimated to be nearly equal to the boiling point for the spray transfer range. The characteristic curve of Qm-I under a constant wire extension for steel or stainless steel wire (Fig. 6) is shown in Fig. 7. Qm increases as the current I is increased for globular transfer range, AB, which is similar to that of aluminum shown in Fig. 8. After reaching a maximum value, Qm decreases for further current increase (curve BCD). The decrease of Qm is related to the pencil-like forming of the wire tip, which is the well-known characteristic for wires of low thermal and electric conductivities, and this suggests that the droplet can detach more easily when the wire tip takes the pencil-like form. Qm-I curves in Fig. 6(a) can be explained clearly from the above mentioned idea. Fig. 8 shows a Qm-I curve for aluminum wire for an extraordinarily long wire extension, and the curve shows the same tendency as that of steel wire. Now we understand that the droplet temperature decreases to as low as the melting point of the wire material when the wire is preheated sufficiently by joule's loss, and it rises to as high as the boiling point when the joule's heating is negligible and the wire is heated abruptly by arc of sufficiently large current density. Figs. 9, 11 show the equivalent melting voltage calculated from Qm and the specific melting rate m (mg/Amp⋅sec) shown in Figs. 4-8. From these figures we can clearly estimate the joule's heating and the anode heating.
In discussing the delayed failure of steel concerning the hydrogen embrittlement, it is important to understand the behavior of hydrogen, especially the occlusion and the permeation controled by the diffusion. As to the diffusion of hydrogen in steel, many results of investigations have been reported, but many of them are obtained at high temperature. And there are a few investigations which are made at room tem-perature or below. Especially, the systematic investigations of the effects of the compositions and the structures have been rarely made. In this report, considering these points, the hydrogen diffusion at the occlusion and the permeation at the room temperature is treated. As to the hydrogen diffusion in steel, Fick's law has been admitted Then, using the mild steel, the volumes of the occluded and the permeated hydrogen were determined by cathodically electric charging for various thicknesses, and comparison was made of the relation between the experimental results and the values which can be calculated using Fick's law and the diffusion coefficient offerd by Hobson. The results of authors experiments are that the occluded hydrogen and the hydrogen distribution at the occlusion are comparable with the theoretical value, and in the case of permeabilty, the experimental value is a little greater than that of the theoretical value. And authors made sure that the hydrogen solubility for the mild steel, used in those experiments and under the condition of cathodical charging, was 43 cc/100g Fe.