This paper describes the results of a study on the determination of the most favorable welding procedures of submerged arc welding for an 80 kg/mm2 high strength steel. When a basic bonded flux containing Mn-Ni-Mo alloying elements was used for submerged arc welding, visible cracks frequently appeared in weld metal. In this case, the lower the welding heat input, the higher the cracking susceptibility of weld metal. This would be due to the fact that an excessive nickel content of weld metal remained as the heat input was decreased. On the other hand, in the transverse tension test of weld joint of HT-80 welded by using Mn-Ni-Mo bonded flux with the welding condition of more than 45, 000 Joules/cm in heat input, the fractured surfaces of H.A.Z. near the bond of weld were observed to be of cleavage type and moreover in this case the H.A.Z. had a poor notch toughness. When using the Mn-Mo basic bonded flux (not containing Ni), although no visible crack was found in the weld metals deposited under various welding conditions, the notch toughness of H.A.Z. was also impaired if the welding was carried out with the heat input of over 45, 000 Joules/cm regardless of the alloying element of weld metal. Through microscopic examination it was clarified that when the structure of H.A.Z. showed bainite more than 70% in area, the H.A.Z. resulted in a poor notch toughness. In this connection it was observed that the smaller the cooling rate of welding, the larger the austenitic grains in H.A.Z. and also that the grains mainly contained a bainitic structure in the matrix and finally that such coarse grains seemed to cause a poor notch toughness of H.A.Z.
We have derived equilibrium formulas for oxygen gas, FeO-Fe2O3-CaO-SiO2 molten slag and Fe-Si-O molten iron by the statistical thermodynamic method, made their approximation to evaluate the concentration of Fe2O3 in the molten slag and the oxygen pressure Po2 equilibrating with the system, and evaluated the concentration of silicin and oxygen dissolved in the molten iron by the electronic computer. Details can be summarized as follows. (1) The Fe2O3 quantity in the total FeO is a negligible quantity in the numerical calculation, when no Fe2O3 is added to the molten slag, FeO-Fe2O3-CaO-SiO2. If Fe2O3 is forcibly added, however, the oxygen pressure is greatly increased. (2) With a decrease in FeO, the oxygen pressure Po2 equilibrating with the composition of the T.FeO-CaO-SiO2 molten slag and the molten iron decrease. Where FeO is low, the minimum value shifts to the CaO side. (3) In relation to the steel-making reaction of the basic open hearth, the temperature is low, and the silicon reduction is little. In the welding exposed to high temperature, however, the reduced silicon greatly increases in relation to the composition of the slag of much SiO2 content, so discloses the calculation, too. (4) According to the conventional equilibrium study of chemical metallurgy of the Fe-Si-O system, if silicon increases, O decreases; but O increases, if silicon decreases. The statistical thermodynamic analysis has, however, disclosed that O also increases, if silicon increases beyond a certain extent. It seems to be attributable to the fact that silicon and O of some bond force are dissolved in the molten iron. (5) The oxygen quantity in the welded metal is far more than that in the steel-making reaction, because the welding is a high temperature reaction, and still the molten slag composition is rich in SiO2. In the reaction at high temperature, therefore, it is necessary to make such choice of the slag composition that the oxygen quantity may be minimum.
Six graces of stainless steels, wrought round bars of types 347, 304, 304L, 316, 310 alloys and Croloy 16-8-2 Cr-Ni-Mo alloy, were first studied about the melting temperatures, and then subjected to the R P I hot ductility test by the authors' apparatus in order to obtain the relationship of the hot ductility at 1200°C during the cooling portion of weld thermal cycle with the change of maximum heating temperatures (1220°-1380°C) of the thermal cycle, and the following results were obtained: (1) The melting temperatures of stainless steels were almost independent of the R P I hot ductility, but those of both grades of stainless steels, type 347 and 310, which indicated a poor value of R P I hot ductility, were lower than those of the other stainless steels. (2) R P I hot ductility of 18-8-2 Cr-Ni-Mo alloy which showed a higher value during the cooling portion from 1340°C of the thermal cycle exhibited a significant decrease when measured at 1200°C during the cooling portion from various maximum temperatures beyond 1340°C, and that of type 347 which showed a lower value exhibited a significant increase below 1340°C. (3) Electron microscopic observation indicated that the grain boundary structures of both grades stainless steels, type 347 and 16-8-2 Cr-Ni-Mn alloy, subjected to R P I hot ductility test were markedly different ; that is, precipitates were dotted clearly on the grain boundary of type 347 and electron diffraction showed that these precipitates were CbB, while no precipitates were observed in 16-8-2 Cr-Ni-Mo alloy. (4) The investigation showed that the degree of embrittlement due to grain boundary liquation of type 347 was far greater than that of 16-8-2 Cr-Ni-Mo alloy, because the eutectic temperature due to reaction of γ-Fe and CbB was 1315°C. R P I hot ductility measured at 1250°C during the cooling portion from 1340°C of the thermal cycle was closely related with the chemical compositions of the tested stainless steels in authors' reports I and II except martensitic stainless steel and the. following formula obtained. Hot ductility Index (I hd) =100-(500+100 Ta+50 Cb-10 Mo)%. This formula must be examined with all the other 'stainless steels, especially, about the other chemical compositions (P. S. Si. Mn etc.) of stainless steels omitted in the formula. But in general R P I hot ductility of stainless steel would be evaluated by this formula, and also the formula may be a great aid to improve the weldable stainless steel with a low hot weld cracking sensitivity, namely a higher value of R P I hot ductility.
In using the brazed joint at high temperature, we have thought that the strength of brazed joints is related to the mechanical properties of brazing filler alloys at elevated temperatures. In this study, we adopted nine kinds of brazing filler alloy composed of Ag-P-Cu, Ag-Cu-Zn-Cd and Cu-Zn systems that were used for brazing copper and copper alloys, and tested their tensile strength, elongation and hardness in the range from the room temperature to 500°C. The test results are summarized as follows; 1) Ag-P-Cu systems With an increasing content of Ag, the mechanical properties at the room and elevated temperatures improved, and a filler alloy of 5%Ag is stronger than one of 15%Ag at the temperature under 250°C, but at over 250°C, the opposite is true. 2) Ag-Cu-Zu-Cd systems No definite relationship was found between Ag content and mechanical properties at elevated temperatures. 3) The mechanical properties of Ag-P-Cu systems are stronger than those of Ag-Cu-Zu-Cd systems at the temperature under 250°C but at over 250°C, the opposste is true. 4) The influences of work hardening on the strength of Ag-P-Cu systems are eliminated at 250°C, but on that of Ag-Cu-Zn-Cd systems they remain a little at 500°C. The mechanical properties at elevated temperatures have been improved by work hardening and the effect of the latter is remarkable with an increasing Ag content.