CA1133992A - Welding process - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE
This invention relates to a welding process using an eutectic alloy wire suitable for welding a super low-temperature steel such as 9% nickel steel. It is characterized in that, in welding a base metal comprising 3.5% - 9.5% nickel by weight, less than 100 ppm oxygen and less than 100 ppm nitrogen through the use of a welding wire comprising 8 - 15% nickel by weight, 0.1 - 0.8% manganese by weight less than 0.15% silicon by weight, less than 0.1% carbon by weight, less than 0.1% aluminum by weight, less than 0.1% titanium, less than 0.0006% boron by weight, less then 100 ppm oxygen and less than 100 ppm nitrogen, the sum of the oxygen content of the wire and the double oxygen content of the base metal is less than 200 ppm and the sum of the nitrogen content of the wire and the double nitrogen content of the base metal is less than 200 ppm.

Description

:~33~9392 BACKGROUND OF THE INVENTION
This invention relates to a welding process using an eutectic alloy wire suitable for welding a superlow-temperature steel such as 9% nickel steel.
9% nickel steel is a high tensile steel which may be used at a superlow--Lemperature up to -196C. The tensile strength o the 9~ nickel steel is defined to be on the order of 70.3 - 84.4 kg/mm according to the ASTM standard, A353 (NNT
material) and A553 (QT material~, and the yield point (0.2%

yield strength~ higher than 52.7 kg/mm2 and higher than 59.8 kg/mm2 according to A353 and A553. The ASTM standard also re-quires that thQ impact value thereof be greater than 3.5 kg-m at -196 C. A fuxther requirement of the ASTM standard, case 1308-5, when a building construction is made by welding the 9% nickel steel is that the tensile strength of a joint including a base metal material be higher than 66.~ kg/mm2 and lower than that of the base metal material per se in order to assure joint perform-ances when annealing is not carried out as weld condition for the removal of stress.
In recent years, howeyer, there have been strong des;res for the development of joints of a tensile strength well above the standard value as defined by the case 1308-5 and welding Materials of a strength not less than that of the base metal material for increasing stress at ti~e of design for weld-ing. As is obvious from the AST~ standard, proper strengt~ and low-temperature toughness of the ~% Ilickel steel is o~tainable from heat treatment but in the case o~ a huge building con~truc-t~on, for example, a storage tank such heat treatment is substant~ally impossible after the building of the construction.

To this end the construction is made serviceable as weld condition.

~ .~

~335~2 1 While it is most desirable to use a welding wire whose composition is identical to that of the base material for welding the 9% nickel steel, high nickel alloy wires as defined ~y the AWSA standard, 5.11 ENiCrFe l - 3, etc. are very often actually used for welding because there are difficulties in obtaining stable low-temperature toughness of the 9% nickel steel wire.
While joints made through using the high nickel welding wire exhibit excellent toughness at a temperature of -196C after welding, they undergo very small tensile strength (particularly, 0.2~ yield strength) as compared to that of the base metal material. No matter when the ~% nickel steel or a 70 kg/mm2 high tensile steel is used, the strength of the joints is low so that stress should be low at the time of design for welding and the overall construction welded be thick. The conventional welding methoa failed to take full advantage of the strength property of the ~% nickel steel and, in fact, suffered rrom two-old economical expenditures, an increased thickness of the construction welded and an increased amount of the expensive high nickel alloy welding wire consumed. I~elding by the high nickel alloy was further disadvantageous of undergoing hot cracks and thermal fatlgue due to a difference between the coefficients of thermal expansion and requiring laboursome welding procedures.
For those reasons the ~% nickel steel is severely limited in application while showing excellent performances as a superlow~temperature steel.
OBJEC~S OF ~HE XN~ENTION
W~th the foregoing in mind, it is an object of the present invention to provide a welding process using a welding wire having sta~le low-temperature toughness comparable with that of the conventional high nickel alloy welding wire and streng-th ~2-~399Z

1 comparable with that of 9% nickel stee] ensuring more stable performances.
BRIEF DESCRIPTION OF THE DRAWINGS
For a more complete understanding of the present in-vention and for further objects and advantages thereof, reference is now made to the following description ta~en in conjunction with the accompanying drawings, in which:
Fig. 1 is a block diagram of an embodiment of the present invention constructed for controlling arc length;

Fig. 2 is a circuit diagram of an example of an arc voltage control method;
Fig. 3 is a circuit diagram of another example of the arc voltage control method;
Fig. ~ is a graph showing the input versus output characteristics of the example of Fig. 3;
Fig. 5 is a perspective view of a drive section;
Fig. 6 and 7 are perspective views of the concept of a magnetic arc blow;
Fiq. 8 and 9 are schematic cross sectional views of a weld zone;
Fig. 1~ and 11 show an e~bodiment of the present in-; vention wherein Figs. 10 and 11 are side ~iews;
Fig~ 12 is a wave~orm diagram of pulsating current;
Figs. 13 and 14 are graphs showing the relationshipbetween the sum of the oxygen cont0nt and nitrogen content of a wire and double oxygen content and double nitrogen content of a base metal and the V notch absorption energy of a weld metal.
Fig. 15 is a plot of the impact resistance of a joint against the boron content of the wire;

~L~33~92 Figs. 16 and 17 are graphs showing the relationship between the oxygen content and nitrogen content of a wire and the "V" notch absorption energy of a weld metal;
Fig. 1~ is a graph showing the relationship between the cool period and the breaking strength of a final layer during TIG welding;
Figs. 19 and 20 are photographs showing ~he joint welded according to the present invention; and Fig. 21 is a representation of groove shapes used with the embodiment of the present invention.
Fig. 22(A), (B) and (C) shows various arc deflection states when the DC voltage sources are connected as in Fig. 1.

DETAILED DESCRIPTIO~ OF THE PREFERRED EMBODIMENTS

The present invention achieves the above discussed objectives by providing a wire of which composition is as follows, wherein ~ means ~ by weight: The welding wire according to the major features of the present invention essentially includes ~ - 15% of nickel and 0.1 - 0.8% of manganese or essentially includes, in addition to those ingredients, less than 0.15% of silicon, less than 0.1~ of carbon, less than 0.1% of aluminum and less than 0.1% of titanium. It further includes less than 0.0006~ of boron, less than 100 ppm of oxygen and less than 100 ppm of nitrogen. The following detailed description will first set forth the wire and then a welding process using the wire.
Although the wire embodylng the present inYention is applicable to the TIG welding method and the TIG plasm arc welding method, they may be called simply as "the TIG welding method", "TIG

welding wire", etc. hereinafter.

1~33~9~

1 The TIG welding wire according to the present lnvention is less expensive than any high nickel alloy welding wires and free from the various problems with the high nickel alloy wires as discussed above, thus providing joints which are excellent in low-tempexature toughness, tensile strength, etc. This makes it possible to reduce substantially the thickness of an overall construction welded, ta]ce full advantage of the inherent proper-ties of the 9% nickel steel and e~pand applications of -the 9%
nickel steel. ~lthough the foregoing has discussed the problems in welding the 9% nickel steel, a typical example of superlow-temperature steels, it is understood that the present invention is applicable to not onl~ the ~% nickel steel but also lower-grade nickel steels such as 5.5~ nickel steel and 3.5% nickel steel.
~, Since as stated above the wire according to the present invention is required to exhibit excellent low-temperature toughness in welding the superlow-temperature steel such as the 9% nickel steel, the content of such a deoxidizer as Al, Ti, ~In and Si is severely limited. In the case of welding materials ~ containing a very small amount of the deoxidizer, more than lQ0 ~ ppm ox~gen in the weld metal leads to the possibility of such .~ weLd defects as blow holes and has adverse effects on low-temperature toughness. On the otr~er hand, oxides in fluxes are generall~ reduced in the shielded arc weld;ng method and the su~merged arc welding method and an active gas (CO2 or 2~
which ~s sllghtly mixed ~nto a shield gas for arc stabili~ation is also reduced in the MIG welding method. In an~ case it is difficult to limit the oxygen content of the weld metal below 100 ppm. However, since the TIG weldi,ng method uses neither oxides as weldiny material nor active gas in the shield gas, it -5~

~339g2 1 can provide a weld joint which is free of joint de~ects at a superlow-temperature of -196C and excellent in low-tempera-ture toughness and other mechanical stren~ths, by using the welding wire and the base metal material which will be detailed with respect to the compound thereof.
As described briefly above, the welding wire according to the present invention includes ~ - 15~ nickel by weight and 0.1 - 0.8% manganese by weight, less than 0.15% silicon, less than 0.1% carbon, less than 0.1~ aluminum, less than 0.1%
titanium, less than 0.0006% boron, less than 100 ppm oxygen and less than 100 ppm nitrogen.
Nickel is essential in ensuring low-temperature tough-ness as in the case of the high nickel steels used with the wire of the present invention. Less than 8% of nickel results in failure to afford sufficient low-temperature toughness to the joints. More than 15% of nic]el, on the other hand, makes the mechanical strength. of the joints too high and brings forth a remarka~le reduction in duct;`lity ~ith.the results that an unsta~le residual austinite i.s developed and then transformed into the martensitic structure at a s.uperlow~temperature to thereby decrease low-temperature toughness. Wh~.le ~qn is very effective in lmproving weldahili.ty and as a deoxidizer and a sulfur captor, a less than 0.1% amount of ~n impai.rs greatly welda~ility and tends to develop blow holes, etc. in the joints due to lack of deoxidization. Accordi.ngly, in this case the effects of ~In i5 not expected. For ~n in excess of ~.8~ there is the trend to develop an unstable residual austinite and deterioxate low~temperature tou~hness to a great e.xtent.
The silicon content should ~e smaller than 0O15~ since si.l~con ~mproves weldability and serves as a deoxidizer but on ~33~92 1 the other hand lowers low-temperature toughness and increases remarkably susceptibility to hot cracks. While only a small amount of carbon is enough to enhance tensile strength, the content should be less than 0.1~ not to decline low-temperature toughness. Aluminum and titanium are both required to be added at less than 0.1% since both are effective as a deoxidizer and in preventing the occurrence of blow holes, etc. but the former impairs significantly crack resilience and the latter accompanies a substantial decline of low-temperature toughness due to precipitation hardening of titanium carbite~
The results of the inventors' experiments indicate that boron is very detrimental in ensuring excellent low-temperature toughness at a superlow-temperature when the welding wire of the above defined compound is used. If the boron content exceeds O.QOQ6%, then the wire is more susceptible to hot cracks, easier to harden and more tough at low-temperatures. For the purpose of the present invention it is most preferable that the boron content be zero and as a practical matter, the boron should be at least less than O.OOQ6%. It îs well known that boron is ~0 mixed as an impurit~ into iron system materials such as electro-lytic iron, one of the chief ingredients of the wire and its content may sometimes exceed Q.02% with the electrolytic iron containing the least amount of impurities. In the case where a su~stantial amount of ~oron is mixed the material, the vacuum degassing solution method would ~e unsuccessful in removing the boron. Pursuant to the teachings of the present invention, the boron content of the starting mater~al should be severly goYerned and the starting material ~e selected such that the boron content of the welding wire does not exceed O.OOQ6%, preferabl~ 0.0004%. It was not until the inventors' findings ~33992 1 that such adverse effects of boron were unveiled. ~ven the ingredients other than boron are within the above defined ranges, it is by no means easy to achieve the objects of the present in~ention so long as the boron content fails to meet the require-ment.
Since oxygen causes oxides to be deposited on a grain boundary or the like, it is necessary to control the oxygen content of the welding wire such that oxygen amounts to less than ` 100 ppm within the weld metal, while it is therefore recommended to keep the oxygen content of the welding wire below 100 ppm.
Since the oxygen in the weld metal is correlated with not only the oxygen in the welding wire but also the counterpart in the base metal, the oxygen content of the base metal should be as small as possible for the purpose of the present invention. The results of the inventors-' experiments also indicate that the oxg~en content of the base metal should be less than 100 ppm and a total of the oxygen content of the wire and the double oxgyen content of the base metal ~e less than 200 ppm in order to attain the objects of the present invention. Why the oxygen content of the base metal should be less than lOQ ppm is due to the fact that the oxg~en in the base metal is hardly affected by the deoxidiz-ing activitv of the deoxidizer contained within the welding wire and dif~icult to remove in the progress of the welding process.
Finall~, nitrogen has the propertie~ o~ precipîtating nitrides ~n the weld metal and deteriorating significantly low-temperature toughness. It leads to that the nitrogen content of t~e welding wire should ~e smaller than lQ0 ppm. Since the nitrogen in the weld metal has a correlation with both the nitrogen in the welding wire and in the ~ase metal, the nitrogen content should be as small as possible for the purpose of the -8~

1~3399;2 1 present invention. The inventors' experiments proved that the nitrogen content of the base metal should not exceed 100 ppm nor the sum of the nitrogen content of the wire and the double nitrogen content o~ the base metal exceed 200 ppm in order to attain the objects of the present invention~
Figs. 13, 14, 16, 17 are ~raphs showing that the "V"
notch absorption energy falls below 80J at -196C in the presence of more than 100 ppm oxygen and more than 100 ppm nitrogen and the oxygen and nitrogen contents are needed to be smaller than 10~ ppm and when the sum o each gas content of the wire and double each gas content of the base metal is more than 200 ppm.
Since as noted above the welding wire according to the present invention is allowed to contain only extremely small amounts of o~ygen and nitrogen along with a very small amount of the deoxidizer, it is most desirable to apply the vacuum degass-ing solution ~ethod to prevent the mixture of oxygen and nitrogen. It is evident from the foregoing that the present invention is aimed at using superlow~temperature steels as the base metal material and the most significant advantages of the present invention are assured when low-temperature steels containing nickel in the range of 3.5 - q.5%, for example, 9%
nickel steel, 5.5% nickel steel and 3.5~ nickel steel, are used as the base metal material, Supposing that the TIG welding method is carried out, the present invention makes it pos-sible to provide joints bearing tensile strength and lo~--temperature thoughness comparable with those of lo~-temperature steels such as 9% nickel steel be defining the compound of the welding wire and more particularly the ceiling contents of boron, oxygen and nitrogen. The welding w;re accord~ng to the present invention has substantially the ~9~

~133'~92 1 same compound as the base metal material and thus provides the joints which are free from the problems such as thermal fatigue aue to differences bet~een coefficients of thermal expansion and hot cracks and exhibit very high mechanical strengths. It is accompanied by providing an economical construction welded which i5 designed with permissible lowest stresses with taking advantage of the properties of the low-temperature steels.
The foregoing has set forth in detail the compound of the welding wire, the oxygen and nitrogen contents of the base metal material which are defined to make sure the performances of the welding wires, and the critical values thereof in taking the oxygen and nitrogen contents of the wire into consideration.
If those requirements are fulfilled, then joints which are excellent in both low-temperature toughness and tensile strength are available anywhere in a weld metal zone, a bond zone and a heat-affected zone (~IAZ2 through the TIG welding method or the TIG plasm arc welding method. A welding process according to the present invention will now be described together with its welding conditions.
Shield gas is of great importance in carrying out the TIG welding method or the TIG plasm welding method. A pure inert gas such as pure Ar or pure He is employed in the welding process of the present invention as is in the con~entional procedure.
Since the oxygen and nitrogen contents of the wire and the base metal material are limited as discussed a~ove according to the present inYention, the advantages of the use of the pure inert gas are enjoyable to the greatest extent. Since the welding process according to the present invention belongs to the TIG
welding method or the TIG plasm welding method, the wire embody~
ing the present invention may be defined in terms of a filler ~10--1 material throughout the following descrip-tion and the append~d claims.
Automatic adjust~ent for the length of an arc developed between a nonconsumable electrode and the base metal material will be set forth as a first condition for the TIG welding method.
In order to gain a homogeneous welding result, the automatic arc ~elding method of the nonconsumable electrode type is needed to maintain a constant arc length at all times irrespective of torch electrode moving method and groove shape and hold in a homogeneous liqui~ pha~e the welding material being supplîed automatically. Since, when the nonconsumable electrode type automat;c arc welding method is to be conducted in every positi,on, it is desira~le to ~eave the torch in such a manner as to make even the surface of beads and minimize internal defects, failure to control the arc length.accurately would result in concave-convex configurations in the grooves or in underla~ing weld ~eads ;n the casa of multi-layer welding and misalignment between the weaving movement and th.e grooYeS. Consequently, the arc length. i5 varied and, ~hen the arc becomes too short, the nonco~sumable electrode may be short-circuited with the base metal w;th.the resulting accidents $uch as the destruction of theelectrode and the mixture o~ th~ electrode material into the weld metal. Moreover, variations in the arc length, that is, variations in the current densit~ o~ an arc column and in the area occupied ~y the arc column w~.thin a molten pool lead to not onl~ lack of penetrat;on but als.o an uneven bead configuration due to failure to gain a homogeneous. molten pool.
While the filler material i~ automatically conveyed in the nonconsum,able electrode type automatic arc welding method, a sligh.t variation in the arc length causes a variation in the ~33~9~:
1 melting speed of the filler wire. Under the circumstance the ~eads ~ecome uneven and the molten pool is not held at a constant temperature, resulting in insuffi.cien~ or uneven penetration of the filler wire into the molten pool or premature penetration.
In the latter case, molten globules fail to move into the molten pool in the normal way in vertical position, overhead position, etc. Particularly, in the case where the nonconsumable electrode type automatic arc welding method is conducted with high alloy steels such as low-temperature steels and stainless steels and non~errous metals, the above discussed pro~lems are more severe because the shape o~ the molten pool is easily variable upon even a slight variation in the arc length in conjunction with the melting point of the ~eld metal and the melting speed of the filler material, Therefore, in the case of the nonconsumable electrode type automatic arc welding method and particularly overall position weaving ~elding and welding with high. alloy steels and nonferrous metals, it is necessary to keep the arc length at the optimum value. very accurately and a measure to control the arc lengt~ ~s necessary and indispensable.
In the past, an attempt to keep substantiall~ constant the arc length. ln the nonconsuma~le electrode type automatic arc weIding method was made by sensing and amplifying the arc voltage and moving ~orward and ~ackward the electrode. The attempt was intended to obvlate motor hunting b~ giving the linear relationship between a motor supply voltage enabling the weld electrode and the arc voltage a specific arc voltage range where the motor is not responsible.

~ithin the specific arc voltage range or a blind zone where the motor is not responsible, the motor comes to a stop in ~L33~92 1 different positions between when the arc voltage returns to a stable point in the decreasing process and when in the increasing process. The stable point is dependent on the amplitude of the varying arc voltage and the movement range of the motor is also dependent on the applied voltage thereto, thus presenting difficulties in stopping the motor in a desired fixed position.
If it is unknown ~Jhere the stable operating point is located within the blind zone, t~en d;fficulties are experienced in adjusting the arc voltage and response relevant ~o variations in the arc voltage declines by the voltage range of the ~lind zone.
In view of the foregoing, the conventional way of controlling the arc length in the nonconsumable electrode type automatic arc welding method is particularly unsatisfactory for the various welding processes which require accurate arc lengths to achieve uniform fusion of the wire being automatically fed and high ~ualit~ weld zones, for instance, fine welding with high allo~ steels and nonferrous metals, weaving welding and overall position welding.
The same applicant as this application has studies on those practical problems, devised improved welding machines and applied for patents therefore. Throughout th.e specification there are disclosed two representative ways. of controlling automaticall~ the arc length.as follows, whether the TIG welding method or the TIG plasm ~elding meth.od:
(~ Through the use of an integrator element~ a dif~
rerential voltage bet~een a detected arc voltage and a preset reference Yoltage is proportionally integrated or multiplied and the resulting signal is us.ed to energize an electrode drive motor, thus controlling automatically-precisel~ the arc length between th2 noncomsua~le electrode and the weld metal; and ~33~9Z

1 (B) As an alternative, there is provided an arc vol-tage detector including an integrator element, a re~erence voltage setting section, an arc voltage control including an integrator or a multiplier, a motor control including an operator and a polarity decision element, and a drive section for moving the nonconsumable electrode by the motor. The difference between an output voltage of the arc voltage detector and the counterpart of the reference voltage setting section is stabilized tnrough the arc voltage control, controlling appropriately the arc length between the nonconsumable electrode and the weld metal.
A specific embodiment of the present invention will be set forth by reference to the accompanying drawings. The illustrative embodiment of the present invention (Fig. 1~
comprises an arc voltage detector 1 including an integrator element 11, a reference voltage setting section 2, an arc voltage control 3 for comparing an arc voltage and a reference voltage or calculation, a motor drive control a and a drive section 5 for moving forward and ~ackward a nonconsumable electrode 43 according to the arc length by the action of a motor.
The arc voltage C-Ea) is sensed by the arc voltage detector 1 and stabilized by the integrator element 11 hav;ng a time constant greater than its hîgh frequency component and the response rate of the motor. The integrator element may be i~plemented ~ith.an CR integrator or integration operational amplifier of an appropriate gain in relation to an input thereto.
The reference voltage setti.ng section 2 divides a DC
constant voltage C~E~ through a variable resistor, a desira~le arc voltage or a desirable arc length ~eing determ;ned by the positi.on of an arm of th.e varia~le resi.stor.
The arc voltage control 3 is adapted to linearly -14~

1~33992 1 integrate and amplify the differential voltage (hereinafter referxed to as "error voltage") between the output voltage of the arc voltage detector 1 and the output voltage of the re-ference voltage setting section 2, whi.ch con~rol 3 includes a linear integration amplifier consi.sting of a resistor 12, an operational amplifier 13, a capacitor 14 and a gain adjusting variable resistor 15, but the last two elements in its feedback circuit. See Fig. 2 The linear integration amplifier achieves the integrat-ing and amplifying operation according to the error voltage andthen provides the next-stage motor drive control 4 with a signal for regain;ng the desirable or appropriate arc length in response to only a small variation of the error voltage, thus making sure that welding i5 effected at the optimum value of the arc voltage.
As des:cribed above, the electrode drive motor 2a is ~raked not to operate in the vicinity of its optimum operating point with.an overload o~ work through the utilization of the inte~rator element 11 and the arc vcltage control 3. The problem with.hunting is thus altogether avoided.
~ As indicated i.n Fig. 3, a pair of multipliers 21 and 22 within th.è arc voltage control 3 may bear the "n"th. (n=2, 3, 4...l power correlation bet~een th.e input and output thereof.
The control 3 includes the two se~ially connected multi.plier~ 21 and 22 and a coefficient potentionmeter 26 next to one of th.e ~ultipliers 22, the potentionmeter 26 comprising a resistor 23 t a operational amplifier 24 and a varia~le resistor 25.
Through. the control 3 the error voltage (input2 and the arc length.control si~nal Coutputl applied to the next-stage motor dri.ve control are correlated as depicted by the cubic curve in Fi~ 4.

~L~33~

1 In this manner, the higher the error voltage the greater the output signal applied to the next-stage motor drive control 4. As a result, welding i5 carried out in the vicinity of the optimum arc length more quickly. The braking torque increases as the optimum arc length is approached. ~ventually the arc length rests on the optimum value. Since no excessive output signal i5 applied to the motor drive control 4 when the error volta~e i5 lo~, the electrode drive motor 20 becomes operative ~ithout hunting in the event that the time-related linear inte-gration met~od is not relied upon.
The motor drive control 4 amplifies the output signal from the arc voltage control 3 and keeps the electrode drive motor from being overloaded. The electrode drive motor is reversi~le according to the polarity of the output signal. The control 4 comprises an operator 28 and a next-stage polarity decision element 3Q.
The operator 28 includes an operational amplifier 31, a feedback element 32 and a tachogenerator 32 for generating an output voltage în proportion to the number of revolutions of the electrode drive motor 20, the output of the tachogenerator being ne~ativly ~ed back via the feedback element 32 to the input of the operational amplifier 31. The feedback element 32 is to reduce variations in the output of the motor 2Q caused by a varyin~ load on the eIectrode driye motor 20.
The polarity decision element 30 includes an npn transistor Trl and a pnp transistor Tr2, base terminals of the transistors Trl and Tr2 being connected to an output terminal of thQ operational amplifier 31, a collector terminal of the transistor Trl bein~ connected to a terminal b of the electrode driye motor 20 via a power source 34 and the equivalent of the ~16-~L~L33992 1 transistor Tr2 being also connected to the terminal b via a different power source 35. Emitter terminals of both the tran-sistors Trl and Tr2 are connected to a grounded terminal a of the electrode dri~e motor 20. ~hen a positive signal is applied to the polarity decision element 30, the transistor Trl is conducting so that the electrode drive motor 20 rotates in a positive direction upon current flowing from the terminal a into the texminal b of the electrode drive motor 20. Contrarily, iL
a negative signal is applied to the polarity decision element 30, then the other transistor Tr2 is conducting so that current flows rom the terminal b into the terminal a of the motor 20 to revert the revolu~ion direction of the motor 20. The drive section 5 includes an electrode section 40 and an electrode driving assembly 41, cf. Fig. 5. The electrode section 40 contains the nonconsuma~le electrode 43 and an insulator 44 supporting the electrode 43, the nonconsumable being connected to a weld cable 45 via a lead conductor extending wi.th;n the insulator. The electrode driving assembly 41 comprises electrode support arms 46, a guide lever 47 for quidin~ the arms 46, a screw 48 for thrusting ~orward and backward the arms 46 and a frame 4~
support~ng the guide lever 47 and th.e screw 48. The electrode supporting arms 46 has three arms 7 the first supports the electrode section 4Q and supports a welding guide chip 5Q at an appropr;ate angle with respect to the nonconsumable electrode 43; the second having a sli.de slot of an appropriate dimension wherein a ~u~de ~ar 47 is slida~le; and the third and last carry-ing a male.screw in mesh.with the screw 4~. The scre~ 48 is coupled with a rotation shaft of the electrode driver motor 20..
The filler ~i.re 51 trayels withi.n the Eiller wire guide chip 50.
~ weaying mechani.sm S5 îs coupled with the Erame 49 to weave the --11~

~3399%

1 nonconsumable electrode 43 to the left or right via the frame 49 In this manner, the drive section 5 travels along a weld line with a suitable traveling device.
As stated above, the arc voltage is sensed by the integrator element having the time constant greater than the high frequency components thereof and the response rate of the motor and the diferential voltage between the output of the integrator element and the preset reference voltage is applied to the linear inte~rator or the multiplier and derived therefrom as the motor drive signal so that the electrode drive motor operates without hunting in such a manner as to settle the arc length.at the optimum point Accordingly, the arc length can ~e preset as desired and settled quickly at th.e preset value in the noncon-sumable electrode type automatic arc weldin~ method irrespective of an uneven ~eld zone and the configuration of the grooves.
Thi~ protects the electrode material, gains a homogeneous fusion of the filler wire, guarantees high quality of the weld zone and makes possi~le oYerall position weldi.ng or precise nonconsumable electrode type automatic arc ~eldi.ng with high.alloy steels, nonferrous metals, etc.
The ab.ove described method enables automatic control of the arc length. The following description will set forth.
deyices for preventing any magnetic blo~ when the DC TIG welding method i5 carried out at a high.speedr These devices are un-suitable for the TIG plasm ~eIding method and applicable only to the TIG welding method in a narro~ sense.
The TIG weldin~ methbd is di.sadvantageous as follows:
. (.lL The TIG welding method is mainly intended to fuse the weId metal into the base metal due to heat conduction. The TIG arc itseIf develops- a~out the molten pool (hi~h temperature ~18-~3L33~9;~

1 portion) without difficulties. If the melting speed is too high, insufficient preheating will cause inferior association (wetness) of the weld metal with the base metal and imperfect fusion of the desposited metal into the base metal.

(2) When the TIG welding method is effected with a DC voltage source, the TIG arc is very sensitive to variations in the surrounding magnetic filed caused by magnetization and varied shape of material to be welded and the weld disable state is brought owing to a magnetic blow. By way of example, Figs. 6 and 7 show in a perspective view the concept of magnetic blow states wherein Fig. 6 is an example of the magnetic blow due to magnetization of steel sheets 61 and 61', ~he base metal material and Fig. 7 is an example of the magnetic blow due to variations in the shape of the steel sheets 61 and 61'. A tungsten electrode 62 (hereinafter referred to as "electrode") extends within grooves in the steel sheets 61 and 61' and the steel sheets 61 and 61' are respectively magnetized with the "N" and "S`' poles, developing a magnetic field within the grooves.
When, for example, a DC constant voltage source is interposed between the electrode 62 and the steel sheets 61 and 61', current flows in a direction normal to the magnetic field. In the case that the current has a positive polarity, an electromagnetic force is developed in the arrow direction f pursuant to the Fleming's left hand law to deflect an arc column 63, a fle~ible conductor, as depicted in the drawings. In Fig. 7 the steel sheets 61 and 61' are not magnetized and the electrode 62 is located near the edges of the steel sheets 61 and 61'~ In this case the eIectromagnetic force is primarily oriented toward the steel sheets 61 and 61' to deflect the arc column 63 in the arrow direction f. Through Figs. 6 and 7 depict the few examples of the magnetic blows. Figs~ 8 and 9 illustrate actual 1~339~Z

1 situations in a weld spot wherein Fig. 8 is a cross sectional view showing upward oriented vertical position welding ~W: the welding direction) and Fig. 9 shows lefthand Elat position ~elding. In any case the arc column 63 is deflected toward the side where the amount of the steel material is large (that is, opposite to the welding advance direct;on~. Under these circum-stances the steel on which welding is about to be performed is hardl~ affected by the arc. As discussed above, preheating and melting are imperfect and inferior fusion occurs between the groove face of the ~ase metal ma~erial and the deposited metal.
The arc is developed on previousl~ ~ormed beads 64 as shown in Figs. 8 and ~ so that the beads 64 are locally fused and become uneven in shape. With the upward oriented vertical position or the overhead position as depicted in Fig, 8, the weld metal may burn through due to overheating of the beads 6~, thus disabling the next succeeding weldiny procedures.
~ ith the foregoing in mind, the inventors have studies on high`speed weldiny conditions free of the above discussed problems and concluded that it i5 desirable to deflect the arc in the welding advance direction in an attempt to make good use of the magnetic blo~ phenomena. The inventorsl attempt is summarized as follows:
In the DC TIG welding method;
(1~ DC voltage sources are connected between the non-consumable eIectrode and the base metal material and between the filler and the base metal ~aterial, respectively;
C2 L flows of current therebetween are (al same when the ~elding material is ahead of the electrode along the welding advance direction; and ~339~32 1 (b) opposite when the welding makerial is behind the electrode along the welding advance direction; and

(3) the arc is directed toward the welding advance direction.
A few embodiments for meeting all the above require-ments will be described by reference to the drawings.
Referring to Fig. 10 showing a side view of a first embodiment, the filler wire 66 is located ahead of a shield gas 1 n cap 65 along the welding advance direction and supplied in the arrow direction Y. As soon as the tip of the filler wire is immersed in the molten pool, it enters into the arc 63 being deflected. Fig. lO is depicted with a straight polarity wherein the base metal 61 serves as an anode and the electrode 62 as a cathode. Conduction current flows through the base metal 61 with the same polarity as the electrode 62 (the base metal 61: an anode and the filler wire 66: a cathode). If the flows of current through the electrode 62 and the welding material 66 are identical to each other in this manner, there are developed two magnetic fields which are attractive to each other so that the flexible arc column 63 is deflected toward the filler wire 66 and hence the welding advance direction as viewed from Fig. 10. The intensity of the magnetic field developing around the filler wire 66 and the degree of the deflection of the arc column are made variable by varying the amplitude of the conduction current into the filler wire 66.

Figure 22 (A), (B) and (C) shows various arc de-flection states when the DC voltage sources are connected as in Fig. 1. The tungsten electrode is conducting with 250A and 15 3~ and the filler wire is conducting with OV (A), lOOA and 4V (B) 1~33~2 1 and 16~A and 6V (C), separately. When the conduction current into the filler is zero (the normal condition of the TIG arc welding method), the arc is not deflected. In this case, the greater the amplitude of the conduction current into the filler wire, the greater the deflection angle of the arc.
Fig. 11 shows another embodiment wherein the filler wire 66' is supplied from behind with respect to the welding advance direction and the filler wire 66' is supplied with con-duction current with the filler wire 66' as an anode and the base metal 61 as a cathode, although the electrode 62 is of a positive polarity as is the case in Fig. 10. Therefore, the direction of current into the filler wire 66' is opposite to that of current through the electrode 62 so that the two resulting magnetic fields repel each other to steer the arc column 63 away from the filler wire and thus toward the welding advance direction.
The closer the supply position of the filler wire is placed w;.th respect to the nonconsumable electrode, the more the influence of the magnetic fields comes into effect. The above advantages may be expected with a very small amount of the con-duction current.
The welding process according to the present inventionas described above is advantageous as follows:
(1) the TIG arc can be directed ahead of the weld line;
(2) the directing force is easily adjustable by vary-ing the amplitude of the conduction current into the -filler wire;
(3) the region in front of the weld line is properly heated and comes to the molten state or nearly the molten state, thus completing the fusion of the filler wire;

(4) overheating of the deposited metal is avoided without damaging the appearance of the beads or burning through ~33g92 1 the deposited me-tal in overhead position welding or upward oriented vertical position we]ding; and

(5) the temperature gradient in the weld zone increases gradually from before the arc to the arc point and decreases gradually rom a fused metal region to a solidified metal region, thus enabling high speed welding without humping the beads.
Although the present invention overcomes the major pro-blems inherent to the TIG welding method, the inventors' effort has further been devoted to make sure the advantages of the present invention. In other words, improvements have been deemed necessary to enhance facility in fusing the weld metal into the base metal in welding with various welding positions and various stee]s and eliminate possible blow holes in high speed welding.
Since the welding process is not the so-called hot wire method per se and the welding material is not heated, if the tip of the filler is moved away from the fused poo~ for any reason, then the filler wire will be thereafter conveyed onto the solidified beads to discontinue further welding procedures.
One powerful approach to solve the above probIem is to weave the arc, but is still disadvantageous as follows:
(1) in the case of a mechanical method wherein a weaver is installed about a welding hèad, the overall construc-tion is massive and bulky and difficult to carry and apply within a narrow space due to installation of the weaver, a motor, a slide base, etc.
(2) the above mechanical method generally needs a proper relative distance between the arc point and the tip of the filler wire. The welding torch and a filler wire guide are therefore mounted integrally on the slide base within the weaver but the relative distance therebetween varies unavoidably due -23~

~33992 1 to vibrations in the weaving operation. In some cases it becomes impossible for the filler wire to enter into the molten pool in a proper position.
(3~ for another approach to develop a magnetic field through an electromagnet, it is required that the electromagnet be located as close to the arc point as possible. In the case of welding with thick sheet steels, the end of the electromagnet should be exposed into the grooves and exhibit an extreme high heat resistance since magnetic force may be centered on a steel of a good magnetic permeability. Satisfaction to these requisites is possible to a limited extent and the overall construction is large sized as set forth in the paragraph (1) even when a water coolins scheme is used at the same time.
In view of the foregoing, investigations have been conducted into the inherent characteristics of the TIG arc in the search fora new weaving method. The results of the investigation indicate that the pinch effect is little since the nonconsumable eIectrode used with the TIG welding method is generally thick (say, 4mm~) to reduce eIectrode consumption toa minimum and ~0 current density is lower than that in the MIG welding method (generally, approximately lmm~). In addition, since the rigidity of the arc is small in comparison with that of the MIG welding method (for example, an inert gas and a metal plasm), the TIG
arc has great flexibility not comparable with that of the r~IG arc.
To take advantage of these inherent characteristics of the TIG
arc, the magnetic fields used in the above disclosed welding process are varied in a fixed or variable rhythm by pulsating the condùction current into the filler wire, steering the TIG arc from a position a little ahead along the welding advance direc-tion to a position beneath the nonconsumable electrode and vice ; -24-,.

~IL33~Z
1 versa. In this instance it is only necessary to pulsa~e the conduction current into the filler wire so that weaving welding needs no large sized and complex peripheral devices about the torch and is applicable to a narrow space. A similar technique for the MIG welding method is disclosed in Japanese Patent Publication 45/39931, for example. This technique employs a current carrying wire conductor other than a consumable electrode, feeds the conductor from behind the consumable electrode and decides flows of current through the consumable electrode and the conductor for deflecting forward the MIG arc along the welding advance direction. ~s noted earlier, the MIG arc is much less flexible than the TIG arc and thus more difficult to deflect forward, as a matter of prac~ice. However, some difficulties are supposed in weaving the arc by the application of the pulsating current. The MIG welding method requires a consider~
able amount of the conduction current into the filler wire because of its high rigidity of the arc if it is desired to deflect the arc by conducting current into a portion of the filler wire being fed near the arc. Under such high current condition it is necessary to increase the feed speed of the filler wire or decrease the current density through the use of a filler wire thick in diameter; otherwise the filler wire becomes fused or an arc develops about the filler wire until the filler wire reaches the fusion pool. Whereas the MIG welding method suffers from the occurrence of the arc but is never unable to go on the welding operation, the nonconsumable electrode is contaminated with metal vapor to an extent to disable substantially the weld-ing operation. Therefore, measures to increase the feed speed of the filler wire and use the filler wire of a great dimension are still available but an increase in the amount of the weld ~39~Z

1 metal leads necessarily to insufficient melting with the MIG
welding method. The above described measures are difficult to adopt with the MIG welding method since the main arc penetrates deeply. In this way, with the MIG welding method it is greatly difficult to deflect the arc and in the case of the TIG welding method various conditions are carefully considered.
Although the -Figures 22 are depicted with 250A of the conduction c~rrent into the nonconsumable electrode, it is generally desirable that the current amplitude be 500A ox less since an excessive amount of current causes an increase in the current density and the rigidity of the arc to thereby mate the deflection and weaving impossible. The electrode is convention-ally supplied with constant current and, if desirable, is energized to develop a pulsive arc. The characteristics of the TIG arc per se are not intended to limit the scope of the present invention.
As is in the conventional combined MIG and TIG welding method or plasm MIG welding method, the conduction current into the ~iller wire should be low enough to avoid the operating ~0 situation where an arc develops from the filler wire and an operating state like a hot wire. Preferably, the conduction current is 200A or less and a voltage at a porjection of the welding wire is lower than the TIG arc voltage; otherwise the magnetic field is too intense and the TIG arc is blown off or blown out. In order that the operating state like the hot wire is avoided and the weIding wire is certainly short-circuited with and brought into contact with the fusion pool, a higher wire feed speed is required. Furthermore, the problem with excessive deposited metal should be avoided.
As discussed above, the present welding process makes ~L3~31t92 1 the weaving action by supplying the pulsating current to the filler wire as shown in Fig. 12. Fig. 12 depicts waveforms of the pulsating current on the left side (A)-(E) and the arc deflecting states on the right side (A)-(D). In the examples (A)-(C) the conducting period alternates with the non-conducting period and particularly in the e~ample (C) -the nonconducting period is zero. In the examples (D) and (E) the filler wire is always supplied with the welding current and high current (Ah) alternates with low current (Al) to form the pulsating current.
(Th) represents the time period where the high current flows and (Tl) the time period where the low current flows. It is under-stood that the deflection state of the arc column in each step is dependent on the amplitude of current. The example (E) indicates that current varies slightly for both the conducting period and the non-conducting period and the present invention is also applicable to this example. The weaving width (weaving angle) and the weaving cycle are freely selectable by a proper selection of the various values (Ah), (Al), (Th) and (Tl) and the weaYing progress and the behavior at both ends of the weaving ~0 amplitude are freely adjustable by varying the amplitude of the current. For example, when butt welding is carried out with the circumference of a pipe in a sequential fashion in vertical ~
horizontal ~ flat positions, the direction of gravity varies with respect to the molten pool so that the most desirable weav-ing pattern may be selected from time to time. This is one of the major advantages of the present invention.
` If unusual situations such as a variation in the groove root gap and an error in the root face are met along the weld line in the conventional one side backing welding method, then the amplitude of the TIG arc current is varied and the arc . -27-~33~Z

1 temperature and shape and the si~e of the molten pool are also varied. ~'o this end it is required to var~ the melting rate of the filler wire and bring the TIG arc current into synchronism with the feeding rate of the filler wire. Such adjustment is rather laboursome but the welding process according to the present `invention can cope with such unusual situations by merely varying the amplitude of the welding current into the filler wire, several measures are listed below:
(1) for example, when welding goes on in the pattern shown in Fig. 12 (A) and the root face becomes thicker, the back bead at that zone is difficult to go out as it is. If the amplitude of the high current (Ah) is increased for this reason, then the forward direction angle of the arc rolumn increases and the arc acts directly on a groove root in a forehand non-deposited metal zone. As a result/ the melting of the root becomes suf~icient and the back bead is completely formed;
(2) in conjunction with the above measure, if the con-duction pexiod of the filler wire is extended, the forward direction period of the arc also becomes longer enough to gain sufficient penetration;
(3) the above measures (lJ and (2) are combined;
(4) the pattern is modified as viewed from Figs. 12 (D) and 1~ (E) and if necessary the current ~Ah) is increased;
(S) the measures (2) and (4) are used together;

(6) the measures (4) and (5) are used together; and

(7) several other measures are available through fine adjustment of those factors.
The welding process according to the present invention is successful in obtaining a subtle weaving pattern by applying the pulsating current in the case that welding is effected with ~.339gz 1 the DC straight polarity and the filler wire is fed from ~ehind the nonconsumable electrode. It is only necessary to make identical the directions of the conducting currents in the case that the filler wire is fed from ahead of the nonconsumable electrode. Moreover, in the case of reverse polarity welding the directions of the conduction currents may be opposite to those of straight polarity welding. The present welding process is also applicable when the filler wire is fed before and after - the nonconsumable electrode.
Satisfactory results are given in low-temperature toughness, tensile strength, etc. as long as the above require-ments are fulfilled. One way to evaluate the mechanical strength of the resulting weld joints is to use small sized specimens such as the charpy test. As long as such evaluation method is traced, there is no problem with the low-temperature characteristics of the joints made according to the requirements at all. However, a few problems still remain with the joints in the` event that they are evaluated through the COD test which has proved itself an appropriate method for evaluating the brittle ~O breakdown characteristics of building constructions welded. The inventors have revealed through extensive studies that such problems are attributable to the heat history of the final layer when multi-layer welding is carried out through the TIG welding method or the TIG plasm weldiny method. It has been concluded that the final layer is also to be given sufficient heat history.
This is achieved by, following multi-layer weIding, cooling the weld bead surface of the final layer below 150C and re-melting the final alyer with the arc generated from the noncomsumable eLectrode while the final bead surface is shielded with an inert gas. Further details thereof will be set forth beIow.

1 The results of the inventors' experiments indicate that, when multi-layer welding is carried out on such a superlow-temperature steel as 9% nickel steel through the use of the welding wire including 8 - 15% of Ni, a central portion of the groove, that is, lower layers are influenced by the effects of heat treatment due to heat cycle during welding of upper layers, the effects of such heat treatment being effective in enhancing the low-temperature toughness of the lower layers. However, the final layer does not enjoy the benefit of such heat treatment and, as a result, the low-temperature toughness of the overall weld metal is decreased remarkably. This tendency is significant when the effects of heat treatment are extremely great as is in eutectic alloy welding with Ni containing ferrite steels such as 9~ nickel steel (grains become greater without difficulties because of Ni contained therein). If the bead surface of the eutectic alloy weld zone of the Ni containing steel is molten again with the nonconsumable electrode, then residual stress is remarkably reduced from the final layer and the low-temperature toughness of the overall weld metal is greatly improved.
A distinguished feature of the welding process of the present`invention, an increase in low-temperature toughness can be evaluated by the COD test which has proved ~o be more appropriate than the conventional Charpy test for evaluation of toughness-at low-temperatures or fracture toughness.
Although the welding process of the present invention using a Ni containing steeI as the base metal will be set forth below, it is obvious that the present invention is also applicable to weld other low-temperature steels.
According to the present weIding process, a joint of a superlow-temperature steel containing Ni is multi-layer welded ~30-~L~3.~
1 through the utilization of an eutectic alloy steel material containiny 8 - 15% Ni by wei~ht and subsequently subjected to re-fusion treatement.
The re-fusion treatment is intended to remove residual welding stress from the final finishing layer in the multi-layer weld zone and give the weld metal low-temperature toughness through the treatment thereof. This treatment is accomplished by the arc heat from the nonconsumable electrode. The depth of penetration during the re-fusion treatment shculd be equal to or less than the depth of the final finishing layer. Otherwise, excessive penetration reduces the effects of the re-fusion treatment. As stated hereinbefore, the object of the re-fusion treatment is to eliminate the residual welding stress on the final finishing layer and increase low-temperature toughness.
For the purpose of the re fusion treatment is is desirable that the depth of penetration during the re-fusion treatment be equal to or less than the depth of the final finishing layer. In the event that penetration is deeper than the final finishing layer during the re-fusion treatment, the re-fused beads become larger than the previous to suppress the effects of the re-fusion treatement~ The re-fused zone is preferably more than half as wide as the final finishing layer in order to allow the whole of the joint to enjoy the desirable effects of the heat treatment.
In the case that there-fused zone is more than 1.3 times as wide as the final finishing layer, heat input becomes excessive and excessive influence of the heat occurs on the base metal.
In carrying out the re-fusion treatment the bead surface of the multi-layer weld zone should be air~ or water-cooled below 150C. In the event that the re-fusion treatment is carried out with a bead surface temperature over 150C, the cooling ~l339g;i~
1 rate decreased in the re-fused zone so that grains on the bead surface becomes coarse with a resulting reduction in low-tempera-ture toughness. If the bead surface is cooled below 150C once after the completion of welding, then it is allowed to be coded through heat release for a relatively brief period of time after the re-fusion treatment so that the bead surface shows a fine crystalline structure with an excellent low-temperature toughnessO
The reason why the cooling rate should not decrease after the re-fusion treatment is evident from analysis of Fig. 18. The beads are cooled gradually through heat release subsequent to the re-fusion treatment. In particular, when the length of time ; where the beads are cooled from 800C ~o 500C extends over 100 seconds, the COD value (the fracture toughness of the most brittle portion at -162C~ becomes lower than 0.1. It is there-fore p~eferable that the beads be cooled from 800C to 500C
for a period of time of about 50 seconds. In other words, an excessive amount of weIding heat input during the re-fusion ~ treatment expands the heat affected zone, prolongs the cooling 3 period and renders the crystal grains coarse, thereby preventing low-temperature toughness from increasing. While the re-fusion treatment goes on under the conditions that the nonconsumable eIectrode is made up of tungsten and the re-fused zone is ' shieIded with an inert gas such as argon and helium, the amount of the shield gas supplied is preferably within the range of 10 -100 ~/min. A supply amount less than 10~ /min leads to various troubles due to lack of shield and on the other hand an amount in excess of 100~/min causes a flow of the shield gas to be disturbed and involved into the re-fused zone, resulting in the occurrence of joint d~fects such as bits.
A well known technique similar to the welding pro~ess ~.

~L~33~9~

1 Of the present invention is disclosed in Japanese Patent Laid-open specifications ~9.55538 and 49/66548. The former teaches '1an attempt to prevent brittle fracture after welding by reheating both the heat-affected zone and bond zone with the heat radiated from the TIG arc" and is considered to be similar to the welding method in that post-treatment is carried out by the TIG arc heat after welding. However, both are totally different from each other in the following aspects:
(1) Whereas the object of the technique disclosed in laid-open publication 49/55538 is achieved by re-heating the heat-affected zone and bond zone, the welding process according to the present invention re-fuses the final finishing portion of the weId metal itself. The distinction results from that the welding process according to the present invention is to enhance the low-temperature toughness of the weld metal zone itself.
In other words, according to the present invention using a super-low_temperature steel as the base metal, it is necessary to increase`the low-temperature toughness of the weld metal in order to gain joint performances comparable to the base metal. The heat treatment on only the heat affected zone and the bond zone as disclosed in publication 49/55538 is, however, not instru-mental in achieving such object.
(2) The technique disclosed in publication 49/55538 achieves its major objects by mereIy re-heating without bringing the weId metal into the molten state, while the re-fusion treatment on the final finishing layer of the weld metal is essential for the welding process of the present invention. The second distinction is due to the background that present invention is applied to the welding of superlow-temperature steels and also the inventors' findings that the object of the present ~33~9Z
1 invention is impossible to achieve unless the re-fusion treatment is effected when a Ni containing steel is employed as the base metal material and the weld metal. In addition, publication 49/
55538 suggests nothing about the concept of the present invention that the weld beads should be kept at a less than 150C tempera-ture in the progress of the re-fusion treatment. ~his fact reflects essential differences in object and technical solution between the present invention and the disclosure of publication 49/55538.
On the contrary, laid-open publication 49/665~8 suggests "an attempt to make smooth the bead surface by re-fusing the same with a TIG welding torch after MIG welding". It also sets forth that the technique disclosed herein prevents in-sufficient solution of the weld metal zone and joint deficits such as blow holes and undercuts and enhance joint strengths.
There is further a suggestion that such attempt is applicable to the welding of a Ni containing steel. The only object that the attempt achieves is to make smooth the bead surface and there is no relevanc~ to technique for enhancing the low-temperature toughness of the weId metal æone.
The primary object of the present invention, on the other hand, is to enhance the low-temperature toughness of the final finishing layer in the weId metal zone. The present invention achieves its primary object by limiting the surfacial temperature of the weld beads below 150 C and carrying out the re-fusing treatment. A variety of desirable conditions for the re-fusing treatment are also defined by the present invention, There is no disclosure about those criteria in publication 49/~6548.
Eventually, the welding process of the present -3~-~33~2 1 invention has the following advantages throuyh the re-fusion treatment of the final finishing layer of the weld metal subsequ-ent toeutectic alloy welding. A significant feature of the present invention resides in extended applications of superlow-temperature steels.
(1) The resulting joint and the base metal material exhibit substantially the same low-temperature toughness, thus enhancing the low-temperature toughness of the overall construc-tion welded;
(2) The welding wire is economical because of no need to use a high Ni steel;
, (3) Both the joint and the base metal material are substantially the same in chemical composition and coefficient of thermal expansion, thus unifying mechanical strengths of the whole construction such as 0.2% strength and hot crack resistance without thermal fatigue due to varying temperature;
(4) As a consequence, the whole construction is relatively thin and light in weight;
(5) It is only necessary to re-fuse the bead surface ~ so that the welding procedures are simple and less expensive;
(6) The effects of the re-fusion treatment are surely attainable by merely keeping the surface temperature below 150C
during the re-fusion treatment.
Although specific examples of the present in~ention will be described in detail, it is not intended to limit the present invention thereto. It is to be understood in view of the whole disclosure that many changes and modifications may be made and are intended to be included within the scope of the present invention.

:~33~9~

1 Example 1 Base metals whose composition are shown in Table 1 were prepared and provided with 60 de~rees grooves through gas cutt,ing.
A~ter the removal of scales ~rom the grooves with a grinder, the TIG welding was carried out ~nder the conditions of Table 3 through the utilization of welding wires of which the composition is enumerated in Table 2. Welding was conducted in such a manner that the front side was first welded and, subsequent to arc air gouginy on the groove root, the xear side was welded. An automatic TIG welding machine with an automatic arc control scheme was employed.
Table 1 9% nickel steel (sheet thickness 200mm) , . . ...................... ~
Symbol C . Si _ S Ni O N

A 0.06 0.42 0.21 0.012 0.004 9.2 20 ppm 25 ppm B 0.06 0.34 0.31 0.006 0.004 8.9 115 ppm 130 ppm C 0.04 0.31 0.25 0.005 0.006 9.0 100 ppm 110 ppm Table 2 ~O Symbol a _ _ . __, d e . - 1- --C 0.08 0.02 0~06 0.07 0.03 Mn 0.45 0.64 0.24 0.72 0.42 Si 0.12 0.03 0.13 0.05 0.06 P 0.010 0.009 0.006 0.010 0.004 S 0.008 0.006 0.004 0.008 0.003 Ni 10.4 12.8 13.4 9.8 11.0 Al 0.04 0.06 0.01 0.01 0.01 Ti 0.01 0.02 0.008 0.01 0.01 B 0.0002 0.0003 0.0008 0.001 0.0002 O 70 ppm 60 ppm 140 ppm 220 ppm 75 ppm 60 ppm 80 ppm 250 ppm 150 ppm 60 ppm Example Example Comparison Comparison Example . _ Examp1e Example __ ~3~99~

1 Table 3 (Welding conditions) Welding position Vertical lIorizontal Current (DC.SP) 250A 350A
Voltage 12V 14V
! Welding rate 5.5 cm/min 15.0 cm/min Weldi.ng heat input 32.7 KJ/cm 19.6 KJ/cm Shield gas argon argon Bath temperature Max. 150 C ~ax. 150 C
difference Weldability was satisfactory in welding in both verti-I cal and horizontal positions.
After welding all the examples were subject to tension test (JIS-Z-3112. A2; measured at room temperature), impact test (JIS-Z-3112, 4; measured at -196C) and side bend test (JIS-3122), the results thereof being listed in Table 4.

~33g9~

_ _ o ~ a) ~ Oo o ~ ~D ~D ~ O M
O 11~ 07 ~1 ~C
.~
~ ~ ~ O ~ O
CO ~ ~ 1` CO ~ O ~ Z
a).~
, . ____ _ ~ U~
1~ ~ n o u~ O ~ O
~ ~ O ~ Z
~C.) P~
_.. _ ... _ .
' O
N ~ ~ U~
rl r~ o ~ co ~ o O u~ Q) Q F~ h ~ D ~ O ~1 :>-~0~ ~
__ .~ . oo ~
~ h ~ u) o l_ ~ ~ ~o ~ ~ O
- a) In C) m ~ O ~ z . ~1 ~
0 . ~ .
E~
N ~) ~r ~ 1 u7 o ~ ~ ~ ~ ~ u~ O
o~l CO O 1~ O ~ Z
I ~ ~
2 0 __ . ___ ___ __ ____._ _ __ ___ _ _ . ~ _ ~ Q ~ o oo ci~
~ o a) ~ o _ . _ u~ O
m h~l o o co co ~ o ~ Z;

~g ~ o ----- - -l ~ ~
~1 ~ o\O . ~ <U ~ U~
q Q) ~ o u~ ~ ~ ~
O ~ o o u~ a:\ ~ O ,~
~ --~ -----------aJ
S~
~1 ~ O ~ rl 3 (~ a) ~ ~
. ~ rl ~ ~ ~ h ~ rd O
3 0 O ~ ~ ~ rl ~z: ~ J to o ~ to ~ ~ ~
~ a) o ~ ~ a) ~:
~o ~ o~
ol ,~ O ~ ~ ~ ~ ~ 0 O (d O rl X ~ rl O O ~ ~) E3 ~ rl I r~ ~ ~
~: al 3 -IJ O 3 ~ O E-l U~ U~ H U~ U~ ~ ~) 1:1~ H
_____ _ _ _ _ _ _, . . - -I

-38~

~3399Z

1 The results of Table 4 can be analyzed as follows:
Nos. 1, 3, 6 and 9: examples within the requirements of the present invention were excellent not only in mechanical strengths such as tensile strength and impact strength but also in the results of X-ray radiation examination.
No. 2: This example (comparison example) contained considerable amounts of the oxygen and nitrogen in the weld metal wherein both the oxygen and nitrogen content thereof are in excess of 100 ppm. The impact strength (low-temperature tough-ness) thereof was very poor and the side bend strength and the X-ray radiation results were also poor.
No. 4: the boron content of the welding wire (com~
parison example) exceeded 0.0006~ with a relativeIy very low fracture strength. The nitrogen content of the wire was also too much.
No. 5: the boron content of the welding wire was too much'and the oxygen and nitrogen contents of the weld metal were both in excess of 100 ppm (comparison example) with unsatisfac-tory results in fracture strength,` stretch'and side bend ~0 strength along with poor outcomes of ~-ray radiation.
No. 7: while the oxygen contents of the filler wire and the base metal met the requirements of the present invention, the'sum of the oxygen content t70 ppm) of th~ welding wire and the double oxygen content (100 x 2 = 200 ppm) of the base metal was-in excess of 270 ppm (i.e., 70 ~ 200 - 270 ppm).
Unsatisfactory res'ults were given in fracture strength, side bend strength'and X-ray examination.
No. 8: this example (comparison example) contained the oxygen content of the filler wire in excess of 200 ppm and 3~ the boron content thereof in excess of 0.0006~ by weight. The !

~33~9'~

fracture strength, the side bend strength and outcomes of X~ray ' radiation examination were unsatisfactory.
Subsequently, the impact strength of the resulting weld joint was measured at ~196C when the 9% nickel steel as denoted by the symbol A in Table 1 was used as the base metal material and the boron content of the welding wire was varied.
Fig. 15 is the results of such measurement indicating that the impact strength of the weld joint greatly decreased with the boron content in excess of 0.0006%. The welding wire exhibited a very high impact strength when the boron content was 0.0006%
or less particularly less than 0.0004~.
The following will discuss various exemplary welding $ conditions as required by the present invention. Unless provided otherwise, a wire used comprised the wire "a" of Table 2 in ~ Example 1 and a base metal material used the Ni steel "A" of ; Table 1.
Example 2 A bead-on-plate was formed with the conditions of Table 5. An excellent appearance was obtained with a high speed of ~0 60 cpm according to the present invention, whereas the conven-tional way resulted in a humping bead at a low speed of 40 cpm.

~3399Z

~ Table 5 .__ ___ Conventional way Present invention . ._ . . __ Welding current 300A (DCSP) Same Welding voltage 12V Same Welding rate 40cm/min60 cm/min Filler wire 1.2 mm~ Same ~ire feed direction behind in rela- Same tion with welding advance direction Wire conduction Non-conductionConduction (60A):

Opposite polarity Shield gas Pure Ar25~/minSame Tungsten 3.2 mm~ Same Base metal 12 mmt Same .

Example 3 Upward vertical position welding was effected with the conditions of Table 6. Fig. 19 shows a bead cross sectional macroscopic structure at an end portion of a welded sheet accord-ing to the conventional way, while Fig. 20 shows the counterpart according to the present invention. The conventional way made a significant convex bead configuration, whereas an excellent bead configuration was provided by the present invention.

~3~92 1 Table 6 . .~ _ . .
Welding conditions Conventional way Present invention . . _ _ _ _ . __ .
Welding current 28OA (DCSP) Same Welding voltage llV Same Welding rate 4 cm/min 7 cm/min Filler wire 1.2 mm~ Same Wire feed direction behind in rela- Same tion with weld-ing advance direction Wire conduction Non-conduction Conduction (60A):
Opposite polarity Shield gas Pure Ar. 25~/min Same Tungsten 3.2 mm~ Same Base metal 25 mmt "V" groove Same Example 4 Welding was carried out in all welding positions on a tube hàving a groove shape as shown in Fig. 21 under the con-ditions of Tables 7 and 8. While-the conventional ~ay provided a convex shàped bead in a flat position and a concave shaped in a vertical position, the present invention made a hbmogeneous and clean bead in all positionsO

339~Z
.

1 Table 7 _-- ~
Welding conditions. . Conventional way .. Present invention Welding current 20OA (DCSP) Same ; Welding voltage llV Same Welding rate 7 cm/min Same Filler wire 1.2 mm~ Same Wire feed direction behind in rela- Same tion with weld-ing advance direction 10 Wire.conduction Non-conduction Conduction (Table 8) Shield gas Pure Ar 30~/min Same Base metal 50 mmt "U" groove Same ,: . : . , . - -. _ _ Table 8 Tube position Time .
1:30 Conduction period in one cycle 0.4 sec _ . 4:30 Non-conduction period in one cycle` 0.6 sec 4:30 Conduction period in one cycle 0.. 6 sec S
7:30 Non-conduction pe.iod in one:c.ycle.. . Ø4.. sec .
207:30 Conduction period in one cycle . . .. 00.4 sec S
10:30 Non-conduction period in one cycle .. Ø6.. sec _ . _ __ 10:30 .Conduction period in one cycle .-S _ Non-conduction 1:30 Non-conduction period in one cycle Example 5 Horizontal direction welding was carried out under the conditions of Table 9. While tubes made by the conventional way were located in JI5-lst class, 2nd grade (blow holes) from the i ::

~3399Z

t outcomes of X-ray transmittance test, the tubes according to the present invention were free of any deficits (the groove shape and the procedures of forming deposited metal are viewed from Fig. 20).

Table 9 Welding conditions Conventional way Present invention _ . .
Welding current 350A (DCSP~Same Welding voltage 12V Same Welding rate 13 cm/min21 cm/min ~elding wire 1.6 mm~ Same Wire feed direction behind in relat- Same tion with weld-ing advance . direction Wire conduction Non-conductionPeriodic conduction Shield gas Pure Ar 50~/min Same Deposition proce- Straight beadStraight + arc dure weaving Base metal 25 mmt Same ExampLe 7 A commerical available 9% Ni Steel of 20 mm thick was provided with a 60 "V" shaped groove and the front surface side of the groove was multi-layer welded. Subsequently, arc air gouging was effected on the groove root and the rear surface side of the groove was welded. The chemical compositions of the 9%

~L~33~92 1 nickel steel used (base metal material) and the ~lelding wire used are illust~ated in Table 11 and the welding conditions in Table 12.

Table 11 _ C Mn Si S Ni Ti Co B O N

Filler ~ _ _ (ppm) (ppm) wixe _ _ _ _ __ _ _ _ ___ A 0.03 0.55 0.25 0.008 0.005 9.5 _ 2.4 0.0004 75 60 Base 0.04 0.68 0.05 0.004 0.004 12.3 0.03 3.5 0.0005 80 50 10Metal 0.06 0.38 0.25 0.007 0.008 9.5 _ _ _ 30 50 (Unit: % by weight) Table 12 Welding Welding Shield method position Current Voltage Date gas Polarity _ Automatic Vertical 300A lOV 70cm/min Ar DC-SP
_ 25~/min _ _ After the bead surface had been cooled below 100C upon the completion of welding, it was subjected to re-fusion treat-ment through the utilization of the TIG arc. The conditions forthe re-fusion treatment are listed in Table 13, wherein the cooling rate is the length of time where temperature falls from 800C to 500C.
Table 13 Condi- ¦ Cooling _ Width of re- Depth of re-'tion rate Shield gas fusion zone fusion zone : .. , ,_ a 50 sec Ar. 25Q/min 1/2 W 1/2 t b 120 sec Ar. 30 1.0 W 1/2 t c70 sec Ar. 30 " 2/3 W 1/2 t d40 sec Ar. 30 " 1.3 W 1/3 t ::: .::;: : .. .. __~

~33~2 1 Table 13 continued W: the width of the final finishiny bead t- the depth of the final finishing bead The resulting weld joints were subjected to impact test (JIS-Z-3112; with a 4th Charpy test specimen and at -196C) and three-point bend COD test (BS standard DD-l9; with fatigue notch added and at -196C), the results thereof being indicated in Table 14.
: Table 14 Weld- . . . COD .
ing Welding Re-fusion vE-196 value wire method conditions (kg.m) (mm) Remarks _ . , B TIG a 13.0 0.25 Exam~le B .. b 12.5 0.05 CoOnmpari-B .. c 15.0 0.25 Exxammple B .. d 18.0 0.30 Example B .. Non re-fusion9.0 0.08 CsOnmpari-: ~ ~ _ ~
Blank (base metal~ 12.0 0.23 .

Analysis of the results of Table 14 reveals that, through the outcomes of the Charpy test are not necessarily in agreement with the COD evaluation values, the low-temperature toughness of the joint subject to the re-fusion treatment -~6-~33~

1 according to the present invention was comparable with that of the base metal (examples X-Z) and relatively very high unlike comparison example Z (non re-fusion treatment). Comparison example X was made when the cooling rate was slow after the re-fusion treatment (Table 13). In this case the COD values indica-tive of low-temperature toughness were very low no matter how excel]ent the Charpy test results were. Comparison Y was made when-a steel containing a large amount of Ni (the Ni content:
17.4%, Table 11) was emplo~ed as the welding wire and could not be expected to have the advantages of the present invention.
Example 8 After welding had been carried out with the same base metal, the same filler wire A, the same groove formation method and the same TIG welding conditions as Example 7, the re-fusion treatment was effected with varying the bead surface temperature (heat input: 45 KJ/cm, shield gas: Ar. 30~/min, width of re-fusion zone: 2/3W, and depth of re~fusion: 1/2t). The results of the Charpy test and the COD test on the resulting joints are depicted in Table 15.
Table 15 Bead surface temp. vE-196 COD value Remarks (C) (k~/m) (mm) .. . _ 100 15.0 0.25 Example Y

150 14.3 0.22 Example O

300 13.8 0.04 Comparison Example O

450 4.7 0.02 Comparison Example P

It is obvious from Table 15 that the bead surface temperature during the re-fusion treatment had great influences on low-temperature toughness. In other words, the present ~47-9~ ' 1 invention remained in effect when the temperature was 150C or less but a considerable reduction in low-temperature was experienced when the temperature exceeded 150C ~comparision examples O and P).

Claims (24)

The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:

1. A welding method for welding a base metal comprising 3.5% - 9.5% nickel by weight, less than 100 ppm oxygen and less than 100 ppm nitrogen through the use of a welding wire compris-ing 8 - 15% nickel by weight, 0.1 - 0.8% manganese by weight less than 0.15% silicon by weight, less than 0.1% carbon by weight, less than 0.1% aluminum by weight, less than 0.1%
titanium, less than 0.0006% boron by weight, less than 100 ppm oxygen and less than 100 ppm nitrogen, said welding method characterized in that the sum of the oxygen content of the wire and the double oxygen content of the base metal is less than 200 ppm and the sum of the nitrogen content of the wire and the double nitrogen content of the base metal is less than 200 ppm.

2. A welding method according to Claim 1, characterized in that the wire is fed into an arc column developing between a nonconsumable electrode and the base metal or into a molten metal and an arc atmosphere is shielded with a pure inert gas.

3. A welding method according to Claim 2 wherein welding is effected in a TIG welding manner.

4. A welding method according to Claim 2 wherein welding is effected in a TIG plasm welding manner.

5. A welding method according to Claim 3 or 4 wherein an electrode drive motor is driven by a signal obtained through linear integration or multiplication of a differential voltage between an arc voltage sensed via an integrator element and a preset reference voltage t thereby automatically controlling the arc length between the nonconsumable electrode and the weld metal.

6. A welding method according to Claim 3 or 4 wherein there are provided an arc voltage detector including an integrat-or element, a reference voltage setting section, an arc voltage control including an integrator and a multiplier, a motor control including an operator and a polarity decision element and a drive section for driving the nonconsumable electrode by a motor, a differential voltage between an output voltage from the arc voltage detector and an output voltage from the reference voltage setting section is stabilized via the arc voltage control for automatically controlling the arc length between the nonconsumable electrode and the weld metal.

7. A welding method according to Claim 3 wherein both the nonconsumable electrode and the wire are connected to a DC
voltage source for DC TIG welding and the directions of current flowing through the wire and the nonconsumable electrode are same when the wire is ahead of the nonconsumable electrode along a welding advance direction and opposite when the former is behind the latter and an arc is deflected forward in the welding advance direction by the influence of the resulting magnetic fields.

8. A welding method according to Claim 7 wherein welding is effected with a DC straight polarity.

9. A welding method according to Claim 7 wherein the arc is weaved forward in the welding advance direction by pulsating the current flowing through the wire.

10. A welding method according to Claim 9 wherein a conduction period of the pulsating current into the filler alter-nates with a non-conduction period thereof.

11. A welding method according to Claim 9 wherein the period where a relatively large pulsating current flows into the filler material alternates with the period where a relatively small pulsating current flows.

12. A welding method according to Claim 10 wherein the current value into the wire is varied during the conduction period.

13. A welding method according to Claim 10 wherein either or both of the conduction period and the non-conduction period are varied.

14. A welding method according to Claim 11 wherein either or both of the large current conduction period and the small current conduction period are varied.

15. A welding method according to Claim 10 wherein a peak value of the conduction current is varied.

16. A welding method according to Claim 11 wherein either or both of a peak value of the large conduction current into the filler and the counterpart of the small conduction current are varied.

17. A welding method according to Claim 9 wherein a ceiling of the conduction current into the filler is 200A

18. A welding method according to Claim 9 wherein a ceiling of the conduction current into the nonconsumable electrode is 500A.

19. A welding method according to Claim 9 wherein the voltage applied to the wire is lower than the voltage applied to the nonconsumable electrode.

20. A welding method according to Claim 2 wherein multi-layer welding a weld bead surface of the final layer is cooled below 150°C and then re-fused through the use of the nonconsum-able electrode while being shielded with an inert gas.

21. A welding method according to Claim 20 wherein the flow rate of the inert gas used for the re-fusion is within the range of 10 - lOO?/min.

22. A welding method according to Claim 20 wherein the depth of penetration is equal to or less than the thickness of the final layer during the re-fusion.

23. A welding method according to Claim 20 wherein a re-fusion zone is 0.5 to 1.3 times as wide as weld beads including the weld bead center.

24. A welding method according to Claim 20 wherein heat input for re-fusion is selected such that the temperature of the beads reaches from 800°C to 500°C within 100 seconds.