Title of Invention

METAL-BASED FLUX CORED WIRE AND WELDING METHOD AND METHOD OF FORMING HIGH FATIGUE STRENGTH WELD JOINT WITH LITTLE AMOUNT OF SLAG

Abstract

The present invention provides metal-based flux cored wire able to reduce the amount of production of slag and to secure a good coatability and a gas shielded arc welding method and a method of formation of a high fatigue strength welded joint resulting in a small amount of production of slag using the same, that is, metal-based flux cored wire for gas shielded arc welding comprised of a steel outer skin in which a flux is filled, containing, by mass% of the wire as a whole, C other than graphite and other than SiC: 0.001 to 0.20%, graphite: 0.10 to 0.7%, Si other than SiC and other than Si02: 0.05 to 1.2%, and Mn: 0.2 to 3.0%, restricted to P: 0.03% or less and S: 0.02% or less, and containing one or more of Si02/ A1203, Na20, and K20 in a total of 0.05 to 0.40% and the balance of iron and unavoidable impurities, the graphite and the one or more of Si02, A12 O3, Na20, and K20 being contained as at least the flux.

Full Text

DESCRIPTION METAL-BASED FLUX CORED WIRE AND WELDING METHOD AND METHOD OF FORMING HIGH FATIGUE STRENGTH WELD JOINT WITH LITTLE AMOUNT OF SLAG TECHNICAL FIELD The present invention relates to the art of arc welding used for an arc welded joint etc. in the automotive field, more particularly relates to metal-based flux cored wire enabling use of metal-based flux cored wire instead of solid wire by reducing the amount of slag of the weld bead surface produced when welding using metal-based flux cored wire and a welding method and method of formation of a high fatigue strength welded joint using the same. BACKGROUND ART Due to the coating process after the welding ends, arc welded joints in the automotive field are formed using solid wire with little amount of production of slag during welding. This is because when a weld zone is covered by slag, the slag is coated over and therefore a problem arises in the adhesion of the coating film and weld zone. On the other hand, due to the rising awareness of environment issues etc., even in the automotive field, lighter weight is being promoted from the viewpoint of improvement of the fuel economy etc. For this reason, the trend is to use higher strength steel material to reduce the thickness, but at this time, a major problem is the fatigue strength of the weld zone. That is, even if using a high strength steel material, the weld zone fatigue strength does not become higher in proportion to the steel material strength. When designing a joint by the fatigue strength, there is the problem that there is no longer any merit in using a high strength steel material. As one means for solving this problem, the method has been proposed of designing the ingredients to lower the transformation temperature of the weld material and reduce the residual stress of the weld zone and thereby improve the fatigue strength (see Japanese Unexamined Patent Publication No. 11-13829 and Japanese Unexamined Patent Publication No. 2004-1075) . Note that after this, this weld material will be referred to as a "high fatigue strength weld material". This method does not require any particularly new production process and enables a high fatigue strength by just replacing the conventional weld material, so may be said to be an efficient method. However, this method also has problems. That is, a high fatigue strength weld material is designed in ingredients so as to include a large amount of alloy elements, so the manufacturing costs increase. Use of this weld material (welding wire) for all of the arc welding in the automotive field is not preferable economically. For this reason, it is necessary to limit use of a high fatigue strength weld material to only parts where fatigue becomes a problem and reduce the amount of consumption of the high fatigue strength welding wire as much as possible. However, while solid wire is superior economically compared with flux cored wire when the amount of consumption of the wire is large, when the amount of consumption of wire becomes smaller, there are the problems that it is not possible to adjust the ingredients by changing the design of the flux like with metal-based flux cored wire and, once preparing the wire-making rods, the design of the ingredients cannot be subsequently changed. When the amount of consumption of the wire is small or in the case of short run use, the economicalness conversely becomes inferior to that of flux cored wire. As a material enabling economic realization of adjustment of ingredients to give weld metal having a high fatigue strength with a small amount of consumption of wire, there is flux cored wire. However, ordinary flux cored wire for arc welding is filled at the inside of the steel outer skin with flux ingredients for maintaining a good welding work efficiency and wire workability in addition to the alloy ingredients. For this reason, when using this for the automotive field to produce an arc welded joint, slag ends up being produced and problems arise in the coating process after welding. This problem is solved by new capital investment for the process of removing the slag after welding, but in this case, an increase in cost due to the capital investment cannot be avoided, so this is not preferred. Various arts have been proposed up to now for reducing the amount of slag of flux cored wire. For example, Japanese Unexamined Patent Publication No. 2000-197991 provides, for wire for welding using a mixed gas of an inert gas and carbon dioxide gas, art for suppressing the amount of production of slag in the case of horizontal fillet welding of wire containing, by mass%, C: 0.08% or less, Si: 0.7 to 1.5%, and Mn: 1.0 to 3.0% and having a filling rate of flux limited to 10 to 30%. However, while this invention has less of an amount of production of slag than the amount of the slag of ordinary flux cored wire, it is not art suppressing the amount of slag to an extent comparable to solid wire. Further, the art of Japanese Unexamined Patent Publication No. 2001-179488, Japanese Unexamined Patent Publication No. 2001-287087, and Japanese Unexamined Patent Publication No. 2003-94196 is art for keeping the flux filling rate low and raising the ratio of the steel outer skin. That is, this method is art making flux cored wire as close as possible to a solid wire. If considering the fact that if the amount of production of slag of the solid wire is small, this art could probably be used to suppress the amount of the slag. However, this method gives rise to a problem similar to that of solid wire due to the smaller amount of flux in the wire, that is, the difficulty of designing the ingredients of the wire as a whole by just adjusting the flux ingredients, and thereby ends up sacrificing the all important economicalness of the flux cored wire. In the above way, according to the prior art, when using metal-based flux cored wire to suppress slag production to the level of solid wire, the filling rate has to be sacrificed. Conversely, in this case, it becomes difficult to secure the amounts of addition of the alloy elements for improving the fatigue strength. That is, in the prior art, no high fatigue strength welded joint using a metal-based flux cored wire resulting in a small amount of production of slag has yet been formed. In view of this problem of the prior art, a method of formation of a high fatigue strength welded joint using a metal-based flux cored wire resulting in a small amount of production of slag has been desired. DISCLOSURE OF THE INVENTION In view of these problems in the prior art, the present invention has as its object the provision of flux cored wire with a much smaller amount of production of slag compared with conventional flux cored wire and a welding method and a method of forming a high fatigue strength welded joint using this. The inventors, from the above viewpoint, took note of the relationship between the flux ingredients and the amount of production of slag and engaged in intensive research on the relationship and slag reduction method. Further, they discovered that by controlling the flux ingredients, it is possible to greatly reduce the amount of slag compared with conventional flux cored wire. The present invention was made based on this research and has as its gist the following: (1) Metal-based flux cored wire for gas shielded arc welding comprised of a steel outer skin in which a flux is filled, characterized in that the metal-based flux cored wire contains, by mass% of the wire as a whole, C other than graphite and other than SiC: 0.001 to 0.20%, graphite: 0.10 to 0.7%, Si other than SiC and other than Si02: 0.05 to 1.2%, and Mn: 0.2 to 3.0%, restricted to P: 0.03% or less and S: 0.02% or less, and containing one or more of SiO2, Al2Os, Na2O, and K20 in a total of 0.05 to 0.40% and the balance of iron and unavoidable impurities, the graphite and the one or more of SiO2, Al2Oa, Na20, and KaO being contained as at least the flux. (2) Metal-based flux cored wire as set forth in (1), characterized in that the metal-based flux cored wire further contains, by mass% of the wire as a whole, SiC: 0.05 to 0.6%. (3) Metal-based flux cored wire as set forth in (1) or (2), characterized in that the metal-based flux cored wire has a flux filling rate of 10 to 20%. (4) Metal-based flux cored wire as set forth in any one of (1) to (3), characterized in that the metal-based flux cored wire further contains, by mass% of the wire as a whole, one or more of Ni: 0.5 to 12.0%, Cr: 0.1 to 3.0%, Mo: 0.1 to 3.0%, and Cu: 0.1 to 0.5% in a total amount of 0.2 to 12.5%. (5) Metal-based flux cored wire as set forth in any one of (1) to (4), characterized in that the metal-based flux cored wire further contains, by mass% of the wire as a whole, B: 0.001 to 0.03%. (6) Metal-based flux cored wire as set forth in any one of (1) to (5), characterized in that the metal-based flux cored wire further contains, by mass% of the wire as a whole, one or more of Nb, V, and Ti in a total amount of 0.005 to 0.3%. (7) Metal-based flux cored wire as set forth in any one of (1) to (6), characterized in that the metal-based flux cored wire further contains an arc stabilizer other than an oxide type, by mass% as a whole, in an amount of 0.05 to 0.5% as the flux. (8) Metal-based flux cored wire for gas shielded arc welding comprised of a steel outer skin in which a flux is filled, characterized in that the metal-based flux cored wire contains, by mass% of the wire as a whole, C other than graphite and other than SiC: 0.01 to 0.20%, SiC: 0.6 to 1.2%, Si other than SiC and other than Si02: 0.05 to 1.2%, and Mn: 0.2 to 3.0%, restricted to P: 0.03% or less and S: 0.02% or less, and containing one or more of Si02, A1203, Na20, and K20 in a total amount of 0.05 to 0.4% and the balance of iron and unavoidable impurities, the SiC and the one or more of Si02, A1203, Na20, and K20 being contained as at least flux in the steel outer skin. (9) Metal-based flux cored wire as set forth in (8), characterized in that the metal-based flux cored wire further contains, by mass% of the wire as a whole, graphite: 0.05 to 0.4% as at least the flux in the steel outer skin. (10) Metal-based flux cored wire as set forth in (8) or (9), characterized in that the metal-based flux cored wire further contains, by mass% of the wire as a whole, one or more of Ni: 0.5 to 5.0%, Cr: 0.1 to 2.0%, Mo: 0.1 to 2.0%, and Cu: 0.1 to 0.5% in a total amount of 0.5 to 6.0%. (11) Metal-based flux cored wire as set forth in any one of (8) to (10), characterized in that the metal-based flux cored wire further contains, by mass% of the wire as a whole, B: 0.001 to 0.015%. (12) Metal-based flux cored wire as set forth in any one of (8) to (11), characterized in that the metal-based flux cored wire further contains, by mass% of the wire as a whole, one or more of Nb, V, and Ti in a total amount of 0.005 to 0.3%. (13) Metal-based flux cored wire as set forth in any one of (8) to (12), characterized in that the metal-based flux cored wire further contains an arc stabilizer other than an oxide type, by mass% of the wire as a whole, in an amount of 0.05 to 0.5% as the flux in the steel outer skin. (14) A gas shielded arc welding method characterized by using metal-based flux cored wire as set forth in any one of (I) to (13) for welding steel sheet. (15) A gas shielded arc welding method as set forth in (14) characterized by using a shielding gas containing C02 in an amount of 3 to 25% and comprised of a balance of Ar gas and unavoidable impurities. (16) A gas shielded arc welding method as set forth in (14) or (15) characterized by using a shielding gas further containing 02 gas in an amount of 4% or less. (17) A gas shielded arc welding method as set forth in any one of (14) to (16) characterized in that the thickness of the steel sheet is 1.0 to 5.0 mm and the tensile strength is 440 to 80 MPa. (18) A method of formation of high fatigue strength welded joint producing a small amount of slag characterized by using a gas shielded arc welding method as set forth in any one of (14) to (17) for welding steel sheet. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a conceptual view showing the relationship between the amount of graphite in metal-based flux cored wire and the amount of the oxides and the amount of the slag produced when using the wire for welding. FIG. 2 is a view showing the relationship between the amount of graphite in metal-based flux cored wire and the amount of C in weld metal when using the wire for a welding test. FIG. 2 is a view showing the relationship between the amount of addition of SiC in wire in metal-based flux cored wire and the C content and Si content in weld metal when using this wire for welding. FIG. 4 is a view showing the relationship between the amount of SiC in metal-based flux cored wire and the amount of oxides and the amount of the slag produced when using the wire for welding. FIG. 5(a) is a plan view showing the shape of a fatigue test piece of a welded joint. FIG. 5(b) is a side view showing the shape of the fatigue test piece of a welded joint and the fatigue load application direction. FIG. 6 is a view for explaining the method of formation of a welded joint and the method of obtaining a Charpy test piece. BEST MODE FOR WORKING THE INVENTION Below, the present invention will be described in detail. Arc welded joints in the automotive field are subjected to a coating process after finishing being welded, but at that time the problem becomes the slag present at the weld bead surface. To secure the adhesion of the coating film and joint surface, it is desirable to reduce the slag as much as possible. For this reason, in the prior art, solid wire has been used. In general, solid wire has much less of an amount of production of slag than flux cored wire. For this reason, so long as using solid wire, it is possible to proceed to the coating process without particularly considering a process for removal of the slag. However, since the joints applicable to are limited to ones of high fatigue strength weld material, when the amount of consumption of wire is small, solid wire has the problem of being inferior to flux cored wire from the economic viewpoint. To simultaneously solve these problems, it is necessary to form welded joints by flux cored wire reduced in the amount of production of slag to the level of solid wire. Flux cored wire comes in two types: an ordinary type and a metal-based flux cored wire containing a large amount of metal powder. Ordinary flux cored wire is designed to secure a predetermined amount of slag ingredients in the flux so as to improve the bead shape and enable all position welding. As opposed to this, metal-based flux cored wire has less slag ingredients due to the large content of metal powder, so all position welding becomes difficult, but slag production can be suppressed. However, with either type of flux cored wire, in the range of the prior art, the amount of production of slag becomes far greater compared with solid wire. Solid wire does not have flux filled in the wire, so the welding position is limited. However, if considering coatability, in the range of the prior art, selection of solid wire is unavoidable. The problem of the welding position was solved by adjusting the position of the steel sheet to be welded. When welding by flux cored wire and analyzing the slag produced on the welding bead, it is learned that almost all of it is comprised of oxides such as SiC>2, MnO, A1203/ and Fe203. Further, this trend remains the same even when adding Ni, Cr, Mo, and other alloy elements into the flux. For this reason, to reduce the amount of production of slag of the flux cored wire to the level of solid wire, it is necessary to suppress the production of oxides during the welding as much as possible. For this reason, it is necessary to reduce the amounts of oxides in the wire. The wire handled by the present invention is limited to, particularly in flux cored wire, metal-based flux cored wire filled with flux with large metal ingredients for this reason. The reason why conventional metal-based flux cored wire could not be suppressed to the level of solid wire in the amount of production of slag was that the wire ingredients causing the amount of production of slag were not sufficiently understood. The art of suppressing the amount of production of slag in the prior art is art keeping the filling rate considerably lower than 10% (for example, Japanese Unexamined Patent Publication No. 2001-179488, Japanese Unexamined Patent Publication No. 2001-287087, and Japanese Unexamined Patent Publication No. 2003-94196). A commensurate effect can be expected, but with this method, the amount of adjustment of the wire ingredients ends up being limited. Reduction to the extent of solid wire has not been achieved. In the art disclosed in Japanese Unexamined Patent Publication No. 2000-197991 having a filling rate of 10% or more, the relationship between the amount of slag ingredients, that is, oxides, contained in the flux and the amount of the slag produced after welding is not sufficiently grasped, so far greater slag is produced compared with solid wire. The problem of reducing the amount of production of slag to the level of solid wire is solved in the present invention by the following methods. That is, the first method is the method of greatly reducing the amount of Si02, KzO, Na20, Al2Oa, or other oxides contained by the flux present in the wire to low slag of the level of solid wire, the second method is the method solving the problem of the increase in the wire drawing resistance caused by the reduction of the amount of the oxides in the flux by using graphite or SiC or in accordance with need both, and further the third method is the method of causing the C in the graphite or the SiC to react with oxygen to form CO or Co2 and enable the source of the slag production, that is, the oxygen itself, to escape from the weld zone. First, the first method will be explained. The inventors first analyzed the ingredients of the slag formed at the weld bead surface. They discovered that almost all of the slag was formed by oxides. Therefore, they considered that if reducing the amount of oxides present in the flux inside the steel outer skin, slag production could be suppressed. They discovered that with wire actually designed in ingredients in this way, slag production was small. The reason why such wire ingredients have not been designed up until this is that the relationship between the oxides in the flux and the amount of production of slag after welding was not clear. The inventors clarified this point and invented metal-based flux cored wire with little amount of production of slag. Next, the second method will be explained. While the above first method was discovered, with just reducing the amount of the oxides in the wire, the problem of the wire drawing resistance arises. That is, during wire production, in particular, drawing, the problem of breakage occurs. This is because the oxides which had been added in the flux acted to increase the fluidity of the flux and keep the resistance to the wire drawing low. That is, the oxides acted as a lubricant. Therefore, the inventors decided, instead of reducing the amount of the oxides, to make up for this by adding into the flux either graphite or SiC, or both, having an action of a lubricant similar to oxides. That is, graphite or SiC, like the oxides in the flux, acts to increase the fluidity of the flux and lower the wire drawing resistance, so the inventors decided to utilize this. Next, the third method will be explained. As explained in the first method, the slag formed at the welding bead is almost all oxides. For this reason, it is desirable to reduce the amount of oxygen as much as possible. The first method is the method of reducing the amount of oxides themselves in the flux, but in addition to this method, as a third method, in the present invention, the method is employed of causing the C in the wire to chemically react with oxygen to enable it to escape outside of the weld zone in the form of CO or C02-CO or C02 is a gas, so even if this is formed, it is allowed to escape outside the weld zone, so will not form slag. Due to the action of reducing the oxygen, a slag reduction effect can be expected. The above were the slag reduction methods in the present invention. Next, the basic ingredients in the wire, that is, the graphite and SiC, in the present invention will be explained. In the present invention, the graphite and the SiC are ingredients utilizing the effects of their both acting as lubricants for suppressing resistance during wire drawing, acting to react with oxygen to enable the oxygen to escape outside the weld zone in the form of CO or C02, and acting to securing the amount of C ingredient in the weld metal and have the same background of art. However, due to their individual characteristics, a person skilled in the art could consider those characteristics and judge whether to use welding wire containing graphite, use welding wire containing SiC, or welding wire containing both. Graphite is an ingredient comprised solely of C and is a convenient ingredient for controlling C. On the other hand, SiC results in the addition of Si in addition to C, so viewed from the perspective of setting the ingredients, graphite is easier to utilize. However, graphite is small in grain and therefore gives rise to the problem of aerial dispersal during preparation of the flux. Production facilities designed to prevent aerial dispersal are not particularly impossible with the state of art, but there is the problem of an increase in the investment cost. On the other hand, SiC has the problems that it ends up adding the excess element Si compared with graphite, has a weaker action as a lubricant compared with graphite so the amount of addition tends to be greater than graphite, etc. A person skilled in the art should select the wire ingredients considering these features. The addition of graphite and SiC acts to increase the C of the weld metal. This acts to degrade the joint characteristics. However, the inventors discovered that the actual amount of C in weld metal becomes smaller than the amount of C in the weld metal estimated from the amount of addition to the wire. The reason for this is believed to be that C reacts with the oxygen and escapes in the form of CO or C02. Therefore, first, the action of graphite will be explained in more detail. FIGS. 1 and 2 are conceptual views showing this action. FIG. 1 plots the amount of graphite in the wire on the abscissa, while plots the amount of oxides in the wire and the amount of the slag produced after welding on the ordinate. In FIG. 1, the relationship between the amount of graphite and the amount of oxides is shown by the broken line, while the relationship between the amount of graphite and the amount of the slag is shown by the solid line. The region above the broken line is the region where the production efficiency of the wire does not fall. As will be understood from the broken line of FIG. 1, if increasing the graphite, it is possible to reduce the amount of oxides without reducing the wire production efficiency. From the fact that the reduction of the amount of oxides is preferable from the viewpoint of suppression of slag production, the effect of addition of graphite will be understood. The solid line of FIG. 1 shows the relationship between the amount of graphite and the amount of the slag produced after welding. The solid line of FIG. 1 shows the following. That is, the smallest amount of oxides (B of FIG. 1) where the wire production efficiency does not fall when a certain amount of graphite is added (A of FIG. 1) is determined from the broken line. Further, the amount of production of slag (C of FIG. 1) when welding by wire containing this amount of graphite and amount of oxides is shown by the solid line. Therefore, even if adding graphite in exactly the amount shown in A of FIG. 1, if not reducing the amount of oxides (for example the amount of graphite and the amount of oxides shown by D of FIG. 1), the amount of the slag does not become the amount shown by the solid line. A greater amount of the slag is produced. The effect of the addition of graphite lies in the ability to thereby reduce the amount of oxides causing slag production (amount shown by the solid line of FIG. 1) without reducing the wire production efficiency. However, FIG. 1 shows that addition of graphite has a greater effect. In the region with a small amount of addition of graphite, in FIG. 1, the solid line and the broken line substantially match, but as the addition of graphite becomes greater, the solid line becomes positioned at the bottom. That is, this means that there is an effect above that of reduction of the amount of the oxides. This phenomenon occurs because the C of the graphite bonds with the oxygen to become CO or Co2 and reduces the amount of the oxygen itself, so as a result production of slag, that is, oxides, is suppressed. In the above way, there are several advantages to graphite. The reason why it has not been used up to now is the belief that the demerits as seen from the mechanical characteristics of the welded joint are too great. For this reason, up to now, addition of graphite has only been attempted in very small amounts. The inventors were suspicious of this conventional common sense. That is, they thought that even if graphite is added and the amount of C of the welding wire increases, if considering the fact that the C may be discharged outside of the welding bead as CO or C02, the C would not become that high as a weld metal ingredient after actual welding. Therefore, the inventors investigated the relationship of the amount of graphite and the C ingredient in the weld metal test. FIG. 2 is a view of the results. FIG. 2 plots the amount of graphite added to the flux (mass% with respect to the wire as a whole) on the abscissa, while plots the amount of C in the weld metal test (mass%) on the ordinate. Note that FIG. 2 shows the case of use of welding wire comprised of metal-based flux cored wire having a content of C in the steel outer skin of 0.05% with respect to the total mass of the wire. Further, the amount of graphite plotted on the abscissa shows all of the amount of graphite contained in the flux. As will be understood from FIG. 2, even if graphite is added in an amount of 0.4%, less than 0.2% remains as a weld metal ingredient. With this degree of amount of C, there is no particular demerit in the mechanical characteristics of the joint. On the other hand, if the amount of addition of graphite exceeds 0.7%, the amount of C remaining as a weld metal ingredient becomes 0.40% or more, in some cases exceeds 0.5%, so the mechanical characteristics are liable to be affected. For this reason, in the present invention, in metal-based flux cored wire, the upper limit of the amount of addition of graphite contained as at least flux has to be 0.7%. The reason is that making the amount of addition of graphite less than 0.7% enables the wire production efficiency to be maintained well while reducing the oxides in the wire and as a result the amount of production of slag during welding is sufficiently reduced and the C content in the weld metal can be kept from sharply increasing. Note that in FIG. 2, even if the amount of addition of graphite is 0, the amount of C of the weld metal does not become 0 since the steel outer skin contains C. Next, the action of SiC will be explained in more detail. Regarding the addition of SiC as well, there is the possibility of the problem arising of increasing the C in the weld metal and causing deterioration of the joint characteristics. Therefore, the inventors investigated the relationship of the amount of addition of SiC in the flux and the amount of C and the amount of Si in the weld metal. The results are shown in FIG. 3. FIG. 3 plots the amount of addition of SiC (showing the amount of addition with respect to the total weight of the wire by %) on the abscissa and plots the amount of C and the amount of Si in the weld metal on the ordinate. Even if the amount of addition of SiC is 0, the amounts of C and Si in the weld metal will not become 0 since the steel outer skin contains 0.05% of C and 0.2% of Si. From FIG. 3, it is understood that there is less C present in the actual weld metal than the C in the weld metal as calculated from the C in the wire. This means that even if adding SiC in the wire, in the same way as graphite, not all of the C in the SiC will necessarily be introduced into the weld metal. That is, this is believed to be due to the C bonding with the oxygen and escaping outside the weld zone as CO or Co2. For this reason, in the present invention, the method is also employed of adding SiC instead of reducing the oxides in the flux. Note that FIG. 3 also shows the amount of Si in the weld metal. If increasing the SiC in the wire, not only C, but also Si-is introduced into the weld metal, but the amount of Si introduced is greater than C. In a system of ingredients in the present invention, in particular a system of ingredients mainly comprised of SiC, as explained later, the upper limit of SiC is limited to 1.2%, but the reason is to avoid an excess of Si of the weld metal. The second slag reduction method in the present invention utilizes the effect of the C in the SiC bonding with oxygen and escaping outside in the form of CO or C02. This phenomenon not only prevents a remarkable increase in C in the weld metal, but also has the effect of reducing the oxygen ingredient, so as a result it is also possible to suppress slag production. That is, addition of SiC is a method having the effect of slag reduction by the two effects of enabling a reduction of the amount of addition of oxides in the wire and the effect of enabling oxygen to escape outside as a gas ingredient. FIG. 4 is a view plotting the amount of addition of SiC on the abscissa and plotting the amount of production of slag and the amount of oxides in the flux in the wire on the ordinate. In FIG. 4, the relationship between the amount of SiC and the amount of oxides is shown by the broken line, while the relationship between the amount of SiC and the amount of the slag is shown by the solid line. The oxides in the flux and the amount of production of slag after welding have a good correlation. For this reason, while different in scale, they can be plotted on the same ordinate. The region above the broken line is the region where the wire production efficiency does not drop. As will be understood from the broken line of FIG. 4, if increasing the SiC, it is possible to reduce the amount of oxides without a drop in the wire production efficiency. If reducing the amount of oxides, it is possible to reduce the amount of the slag. The effectiveness of SiC addition can be understood. For example, the smallest amount of oxides (B of FIG. 4) where the wire production efficiency does not fall when a certain amount of SiC is added (A of FIG. 4) is determined from the broken line. Further, the amount of production of slag (C of FIG. 4) when welding by wire containing this amount of SiC and amount of oxides is shown by the solid line. Therefore, even if adding SiC in exactly the amount shown by A of FIG. 1, unless reducing the amount of oxides (for example, the amount of SiC and the amount of oxides shown by D of FIG. 3), the amount of slag will not become the amount shown by the solid line. A greater amount of slag is produced. The effect of SiC addition lies in the ability to thereby reduce the amount of oxides causing slag production (to the amount shown by the solid line of FIG. 4) without reducing the wire production efficiency. From the fact that the reduction of the amount of oxides is preferable from the viewpoint of suppression of slag production, the effect of addition of SiC will be understood. However, FIG. 4, in the same way as FIG. 1, shows that there is a greater effect in SiC addition. In the region where the amount of addition of SiC is small, in FIG. 4, the solid line and the broken line substantially match, but if the SiC addition becomes greater, the solid line becomes positioned at the bottom. That is, this means that there is an effect above that of reduction of the amount of the oxides. This phenomenon occurs because the C of the SiC bonds with the oxygen to become CO or C02 and reduces the amount of the oxygen itself, so as a result production of slag, that is, oxides, is suppressed. Above, in the present invention, three methods using graphite and SiC were used to realize metal-based flux cored wire reducing the amount of production of slag after welding to the level of solid wire. Next, the reasons for limitation of the numerical values in the present invention will be explained. First, the reasons for limitation of the numerical values of the ingredient elements in the metal-based flux cored wire will be explained. In the present invention, wires of two groups of types of ingredients are shown, that is, a type of ingredients having graphite as an essential ingredient and a type of ingredients having SiC as an essential ingredient. In the present invention, these types of ingredients will be called "graphite ingredient-based wire" and "SiC ingredient-based wire". First, the reasons for limitation of the graphite ingredient-based wire will be explained. For the C other than the graphite and other than the SiC, the lower limit of the amount of C, by mass% with respect to wire as a whole, was made 0.001%. With an amount of C not reaching this, it becomes difficult to secure the strength of the steel outer skin and the problem of breakage is caused at the time of production of the wire, so the lower limit value was made this value. Further, the upper limit of the C other than graphite and other than SiC was made 0.20% because with an amount of C over this, since graphite is separately added to the flux in the metal-based flux cored wire of the present invention, the amount of C of the weld metal becomes excessive, so the upper limit was made 0.20%. Note that as C other than the graphite and other than the SiC, there may be C contained in the iron powder added to the flux. In this case, considering the hardening during wire drawing, it is desirable to set the C of the steel outer skin at 0.15% or less and make up for the balance by the C in the iron powder. In the present invention, graphite is added so as to function as a lubricant of the flux and so as to react with the oxygen and form CO or CC>2 so as to allow the oxygen to escape outside. These are aimed at reduction of the amount of the slag, but they further keep the C in the weld metal suitable and keep the transformation start temperature low so as to obtain the function as a high fatigue strength weld material. One of the objects of the present invention is to secure the coatability due to the fact that one of the fields of utilization of the present invention is the automotive field. The amount of C of steel sheet used in this field is usually comparatively low. It is 0.05% or less, in some cases, 0.01% or less. In this case, it becomes difficult to maintain a suitable amount of C in the weld metal when considering dilution of the base material as well. This amount of C may be introduced from the steel outer skin as well, but inclusion in the wire as graphite enables both the actions of slag reduction and lubricant to be given, so in the present invention, the upper limit of the amount of C in the steel outer skin is kept low. It is necessary to add the lowest amount of graphite to that extent. The lower limit of the graphite in the present invention was made 0.10% for the two reasons of securing the lowest limit of amount of C in the weld metal and securing slag reduction and a lubricant effect. The upper limit of 0.7% was set because with an amount of graphite over this, the amount of C in the weld metal increases, the weld metal becomes too hard, and other problems arise in the joint characteristics. Note that to secure the action of graphite as a lubricant, preferably the lower limit of graphite is set to 0.15%. The Si in the wire may be divided into Si of the oxide Si02 and SiC and the Si other than that of SiOg and SiC. Si02 and SiC are mainly contained in the flux. The Si contained in the steel outer skin is almost entirely Si in solid solution in the steel. In the present invention, the Si02 in the steel is an unavoidable impurity. The Si02 and other oxides contained in the flux are also contained, in addition to mica, in the binder used when granulating the flux to be filled. While referring to the wire as being metal-based, so long as assuming flux cored wire, it is not possible not to add a binder. For this reason, as explained later, in the present invention, the total amount of the oxides is prescribed. However, the Si other than the Si02 does not have to be defined that much from the viewpoint of slag production since no oxygen ingredient is contained. However, since it is necessary to achieve a lowest limit of the deoxidation, the lower limit was made 0.05%. On the other hand, excessive Si addition is not preferable from the viewpoint of the joint characteristics since it causes hardening of the weld metal, so the upper limit was made 1.2%. Mn is an element necessary for securing the strength. The lower limit of 0.2% of Mn was set since if lower than this, it is difficult to secure the weld metal strength. On the other hand, if the amount of addition becomes too excessive, deterioration of the toughness of the weld metal is caused, so the upper limit was made 3.0%. P and S are unavoidable impurity elements. In the present invention, if these elements are present in large amounts in the weld metal, the toughness deteriorates, so the upper limits of the contents of P and S were set to 0.03% and 0.02%, respectively. Si02, A1203, Na20, and K20 are called "slag materials". The reasons for adding these materials are that they function as binders when granulating the flux ingredients before production of the metal-based flux cored wire, they function as lubricants reducing the resistance of the flux in the process from filling the inside of the steel skin to drawing to the predetermined wire diameter, etc. In the process of granulation of the flux, the flux content of the wire becomes uniform, so this process is essential for production of good quality metal-based flux cored wire. On the other hand, the action as a lubricant is given to the graphite in the present invention, these oxides are mainly added for the granulation of the flux. However, these are all oxides. From the viewpoint of reducing the amount of production of slag, it is preferable not to add them. However, without these ingredients, the flux cannot be granulated, so the lowest limit of amount has to be added. The lower limit of 0.05% was set because if below this, the above effect is no longer obtained and problems arise in the wire quality and production efficiency. The upper limit of 0.40% was set since with an amount of addition over this, the amount of production of slag after welding becomes greater and a problem arises in the coatability. In the graphite ingredient-based wire in the present invention, it is possible to add SiC in accordance with need. In this case, since graphite is added for securing the C of the weld metal or reducing the resistance during wire drawing, it is not necessary to add this to the extent of the later explained SiC ingredient type. The lower limit of 0.05% of the amount of addition of SiC was set as the lowest limit of the value by which addition will have the effect of improving fatigue strength and reducing slag. On the other hand, the upper limit of 0.6% was set since with an amount of addition over this, since graphite is already sufficiently added, the amount of C in the weld metal would become excessively high and deterioration of the joint toughness would be invited. In the present invention, Ni, Cr, Mo, and Cu are elements added mainly for the purpose of improving the tensile strength or fatigue strength of the welded joint. The amounts of addition of these elements should be selected according to the object of use of the weld material. These elements are also added to increase the strength and lower the transformation start temperature of the weld metal to increase the fatigue strength. However, while these elements have the same actions, their effects per percent are not necessarily the same, so this range was set for each element. Ni is an element lowering the transformation start temperature and improving the strength, toughness, or other joint characteristics. The lower limit of 0.5% of Ni was set as the lowest limit of the value by which addition results in improvement of the strength or toughness. The upper limit of 12.0% was set because with an amount of addition over this, the cooling may end without the weld metal transforming and with it remaining as austenite and thereby no improvement in fatigue strength being able to be expected any longer. Cr and Mo are elements added in the present invention for raising the strength and hardenability of the weld metal. To improve the fatigue strength of the welded joint, it is necessary to make the structure a martensite or other structure with a low transformation temperature. For this reason, hardenability must be secured. Cr and Mo are elements added to improve the strength and more easily secure the hardenability. For this reason, the lower limit 0.1% of each of these elements was set as the lowest limit of the value for obtaining the effect of improving the strength and securing hardenability. On the other hand, Cr and Mo, like Ni, may be added to reduce the transformation start temperature, but unlike Ni, they are not as preferable as with Ni addition in the point of improvement of the toughness of the weld metal. For this reason, the upper limits of these elements have to be set lower than Ni. The upper limits of 3.0% of these elements were set since with addition over this, problems arise in the joint characteristics. Cu, in the same way as Cr and Mo, is an element having the effect of reducing the transformation start temperature, improving the strength, and securing hardenability. However, if adding too much, there is a risk of causing Cu cracks in the weld metal, so the upper limit value has to be set lower than Cr or Mo. The upper limit of 0.5% was set to eliminate the risk of Cu cracks. On the other hand, Cu can be used for mainly plating wire to secure a current carrying ability. The lower limit of 0.1% of Cu was set as the lowest limit of the value required for the effect of improving the strength and improving the hardenability and for securing a current carrying ability. In the present invention, the total value of these four elements Ni, Cr, Mo, and Cu is also limited. These elements have the effect of reducing the transformation start temperature, improving the strength, and securing hardenability and act in the same way. However, if these elements are overly added, the weld metal structure becomes an austenite structure, that is, there is no longer a transformation in the cooling process after welding, so the effect of improvement of the fatigue strength disappears. Further, if the amount of addition is small, the effect of improvement of the tensile strength can no longer be expected. For this reason, the total of these elements also has to be limited. The lower limit of 0.2% was set since with an amount of addition below this, the effect of increase of strength can no longer be expected. The upper limit of 12.5% was set since with an amount of addition over this, the weld metal becomes a structure mainly comprised of austenite, the transformation expansion in the cooling process during welding becomes insufficient, and no improvement in the fatigue strength can be expected any longer. Note that when these elements are added for the purpose of only improving the tensile strength, the upper limit of the amount of addition is preferably set to 4.0% and the later explained Nb or V is preferably added together in terms of economy. Further, when added for the purpose of improving the fatigue strength of the welded joint, the lower limit of the total amount of addition of these four elements Ni, Cr, Mo, and Cu is preferably set to 2.0%. This is because when the amount of addition is below this, the transformation start temperature of the weld metal will not become lower and improvement of the fatigue strength will become difficult. To more reliably improve the fatigue strength, this lower limit value is preferably set to 3.0%. B is a hardenability element. For addition of B to secure the hardenability of the steel sheet, addition by wt% of 0.001% or so is sufficient, but in the case of weld metal, the oxygen content is higher than steel sheet. The B bonds with the oxygen to end up robbing it of that effect, so it has to be added in a greater amount compared with the case of steel sheet. To reason for securing the hardenability is to make the microstructure of the weld metal a higher strength structure, to suppress the appearance of a structure where transformation starts at a high temperature and realizing a microstructure transforming at a lower temperature, etc. These effects are also preferable from the two viewpoints of securing the tensile strength and securing the fatigue strength, so in the present invention, it was thought to positively utilize them. The lower limit of the amount of addition of B was set at 0.001% as the lowest limit of the value at which the hardenability of the weld metal can be improved. The upper limit of the amount of addition of B was set as 0.03% because even if B is added in an amount over this, the effect obtained by addition of B does not increase. Mb, V, and Ti are all elements having the action of forming carbides and increasing the strength and can be expected to increase the strength by comparatively small amounts of addition. That is, in the present invention, these three elements are elements from which equal effects are expected. For this reason, in the present invention, the total amount of these elements is limited. The lower limit of 0.005% was set because if the amount of addition is less than this, not much of an increase in strength can be expected. On the other hand, if the amount of addition is over 0.3%, the strength of the weld metal becomes excessive and problems arise in the joint characteristics, so the upper limit was set as 0.3%. Note that Ti has the action of stabilizing the welding arc in addition to increasing the strength, so preferably the lower limit of the Ti content is set to 0.003%. An arc stabilizer is an element which is added for stabilizing the welding arc. The Na20 or K20 etc. described in (1) of the present invention also have the action of stabilizing the arc, so these may also be called "arc stabilizers". For this reason, in the present invention, it is not always necessary to add any additional arc stabilizers. Further, if adding these arc stabilizers, there is a risk that the first object of the present invention, that is, the reduction of the amount of the slag, can no longer be achieved. However, a compound of Na, Al, and F has the action of stabilizing the arc and, unlike Na20 or K20 etc., there are also cryolite (Na3AlF6) and other non-oxide types. If other than an oxide, even if added to the welding wire, it will not form a source of supply of oxygen, so will not produce slag formed by oxide. It is therefore possible to add these arc stabilizers to meet with demands for stabilizing the arc more. For this reason, setting a range of possible use of arc stabilizers other than oxide types in the present invention has significance. The lower limit of 0.05% of the arc stabilizer other than an oxide type was set as the lowest limit of the value by which addition gives an arc stabilization effect. On the other hand, the upper limit of 0.5% was set because oxide-based slag materials were already added as a binder and these elements also act as arc stabilizers, so even if added over that, the effect would not change. The above were the reasons for limitation of the ingredients of the graphite ingredient-based wire among the flux cored wire of the present invention. Next, the reasons for limitation of the ingredients of the SiC ingredient-based wire will be explained. The C other than graphite and other than SiC is the C added, for example, from the steel outer skin. The C in the weld metal has the same action no matter whether introduced from the steel outer skin, graphite, or SiC. However, the C in the steel outer skin is effective for preventing breakage during the drawing process while producing wire, so it is necessary to set a suitable range if even just for this. The C other than graphite and other than SiC was set with a lower limit of 0.01%, but the reason was that if the amount of C in the steel outer skin is less than this, the influence on wire breakage is great and the cost of the wire itself ends up greatly increasing and economic formation of a welded joint becomes impossible. On the other hand, if excessively adding C to the steel outer skin, this time the material ends up hardening during drawing, so the upper limit was set to 0.20%. Note that the C other than graphite and other than SiC may be C contained in the iron powder added to the flux. In this case, considering the hardening during wire drawing, it is desirable to set the C of the steel outer skin at 0.15% or less and make up for the balance by the C in the iron powder. Note that the lower limit of the C other than graphite and other than SiC is set higher compared with the graphite ingredient type. This is because the action of SiC as a lubricant is lower than graphite. The lower limit of the amount of addition of SiC is set as 0.6% and the upper limit as 1.2%. In SiC ingredient-based wire, the reduction of the transformation start temperature of the weld metal is mainly achieved by the C in the SiC, so these values were set to secure lowest limit of the amount of C. The lower limit was set as the lowest limit of the value at which an improvement of the fatigue strength can be expected. The upper limit of 1.2% was set because if SiC is added to the wire, as already shown in FIG. 3, not only C, but also Si is introduced into the weld metal and, with an amount of addition over this, when combined with the Si already added to the wire in addition to the SiC, the impact characteristics of the weld zone can no longer be secured and the problem of hardening of the weld metal or the problem of the austenite structure becoming greater and transformation no longer occurring and no improvement of the fatigue strength being able to be expected arise. The lower limit of the Si other than SiC and other than Si02 was set as 0.05% to obtain the lowest limit of the deoxidation effect during welding. Note that the amount of Si is preferably made at least 0.1% to improve the affinity of the molten pool and the steel sheets and obtain a good bead shape. On the other hand, excessive addition causes the weld metal to harden and is not preferable from the viewpoint of the joint characteristic, so the upper limit was set as 1.2%. Mn is an element required for securing strength. The lower limit of 0.2% of Mn was set since if below this, the weld metal strength becomes difficult to secure. On the other hand, if the amount of addition becomes excessively great, deterioration of the toughness of the weld metal is caused, so the upper limit was made 3.0%. P and S are unavoidable impurity elements. In the present invention, if these elements are contained in the weld metal in large amounts, the toughness deteriorates, so the upper limits of the contents of P and S were set at 0.03% and 0.02%, respectively. Si02, A1203, Na20, and K20 added to the flux are called "slag materials". The reasons for adding these are that the function as binders when granulating the flux ingredients before production of the metal-based flux cored wire, they function as lubricants reducing the resistance of the flux in the process from filling the inside of the steel skin to drawing to the predetermined wire diameter, etc. The lower limit of 0.05% was set because if below this, the above effect is no longer obtained and problems arise in the wire quality and production efficiency. The upper limit of 0.40% was set since with an amount of addition over this, the amount of production of slag after welding becomes greater and a problem arises in the oatability. In an SiC ingredient type as well, graphite may be added in accordance with need. Addition of graphite not only contributes to the reduction of the transformation start temperature of the weld metal, but also acts to suppress the resistance from the flux during wire drawing, so this is added to improve the efficiency of wire production. However, there is the problem of aerial dispersal during flux addition, so it is possible to select whether to add graphite to the flux after considering these problems. The lower limit of the amount of addition of graphite was set at 0.05%. This was set as the lowest limit of the value by which addition of graphite can be expected to have the effect of improving the fatigue strength of the welded joint. The upper limit of 0.4%, in an SiC ingredient type, was set since reduction of the transformation start temperature of the weld metal is mainly achieved by SiC, so with an amount of addition over this, the problem of hardening of the weld metal or the problem of the austenite structure becoming greater and transformation no longer occurring and no improvement of the fatigue strength being able to be expected arises. In the SiC ingredient-based wire in the present invention, in accordance with need, further, Ni, Cr, Mo, and Cu may be added. Ni is an element required for reducing the transformation start temperature of the weld metal and achieving an improvement in the joint fatigue strength. Further, it is an element improving the strength or toughness or other joint characteristics. The lower limit of 0.5% of Ni was set as the lowest amount by which an improvement of the fatigue strength can be expected. The upper limit of 5.0% was set since, in an SiC ingredient-based wire, the C already results in a considerable reduction of the transformation start temperature and with an amount of addition over this, the interaction with the already added amount of C may cause cooling to end while remaining in the austenite state without weld metal transformation and no improvement of the fatigue strength can be expected any longer. Cr and Mo, in the present invention, are elements added for reducing the transformation start temperature of the weld metal and raising the strength and hardenability. To improve the fatigue strength of the welded joint, it is necessary to make the structure martensite or another low transformation temperature structure. For this reason, the hardenability has to be secured. Cr and Mo are elements which are added to facilitate improvement of the strength and secure hardenability. This action is greater than with Ni. For this reason, the lower limits of these elements are 0.1%. This is set at the lowest limit of the value at which the effect of improving the strength and securing hardenability can be obtained. On the other hand, Cr and Mo can lower the transformation start temperature by addition in the same way as Ni, but unlike Ni, they are not as preferable as the addition of Ni from the viewpoint of the improvement of the toughness of the weld metal. Further, with the first ingredient type, since C already enables a considerable reduction in the transformation start temperature to be achieved, the upper limits of these elements were set at 2.0%. Cu, in the same way as Cr and Mo, is an element having the effect of reducing the transformation start temperature, improving the strength, and securing hardenability. However, if adding too much, there is a risk of causing Cu cracks in the weld metal, so the upper limit value has to be set lower than Cr or Mo. The upper limit of 0.5% was set to eliminate the risk of Cu cracks. On the other hand, Cu can be used for mainly plating wire to secure a current carrying ability. The lower limit of 0.1% of Cu was set as the lowest limit of the value required for the effect of improving the strength and improving the hardenability and for securing a current carrying ability. In SiC ingredient-based wire, the total amount of addition of Ni, Cr, Mo, and Cu is also limited. These elements have the effect of reducing the transformation start temperature, improving the strength, and securing hardenability and act in the same way. However, if these elements are overly added, the weld metal structure becomes an austenite structure, that is, there is no longer a transformation in the cooling process after welding, so the effect of improvement of the fatigue strength disappears. On the other hand, if the amount of addition is small, the transformation start temperature is no longer sufficiently reduced and no improvement of the fatigue strength can be expected. For this reason, the total of these elements also has to be limited. The lower limit of 0.5% was set since with an amount of addition below this, the effect of increase of fatigue strength can no longer be expected. The upper limit of 6.0% was set since C already enables a considerable reduction in the transformation start temperature to be achieved, so with an amount of addition over this, the weld metal becomes a structure mainly comprised of austenite, the transformation itself no longer occurs, and no improvement in the fatigue strength can be expected any longer. Note that the reason why the upper limit of the total amount of these elements is kept low compared with the graphite ingredient type is that the action of SiC as a lubricant is not as active as graphite, so the amount of addition is set larger than graphite and, as a result, the C in the weld metal is sufficiently secured. B is a hardenability element. For addition of B to secure the hardenability of the steel sheet, addition by wt% of 0.001% or so is sufficient, but in the case of weld metal, the oxygen content is higher than steel sheet. The B bonds with the oxygen to end up robbing it of that effect, so it has to be added in a greater amount compared with the case of steel sheet. The reasons for securing the hardenability are to make the microstructure of the weld metal a higher strength structure, to suppress the appearance of a structure where transformation starts at a high temperature and realizing a microstructure transforming at a lower temperature, etc. These effects are also preferable from the two viewpoints of securing the tensile strength and securing the fatigue strength, so in the present invention, it was thought to positively utilize them. The lower limit of the amount of addition of B was set at 0.001% as the lowest limit of the value at which the hardenability of the weld metal can be improved. The upper limit of the amount of addition of B was set as 0.020% because even if B is added in an amount over this, the effect obtained by addition of B does not increase. Nb, V, and Ti are all elements having the action of forming carbides and increasing the strength and can be expected to increase the strength by comparatively small amounts of addition. That is, in the present invention, these three elements are elements from which equal effects are expected. For this reason, in the present invention, the total amount of these elements is limited. The lower limit of 0.005% was set because if the amount of addition is less than this, not much of an increase in strength can be expected. On the other hand, if the amount of addition is over 0.3%, the strength of the weld metal becomes excessive and problems arise in the joint characteristics, so the upper limit was set as 0.3%. Note that Ti has the action of stabilizing the welding arc in addition to increasing the strength, so preferably the lower limit of the Ti content is set to 0.003%. An arc stabilizer is an element which is added to the flux for stabilizing the welding arc. The Na2o or K20 etc. described in (1) and (4) of the present invention also have the action of stabilizing the arc, so these may also be called "arc stabilizers". For this reason, in the present invention, it is not always necessary to add any additional arc stabilizers. Further, if adding these arc stabilizers, there is a risk that the first object of the present invention, that is, the reduction of the amount of the slag, can no longer be achieved. However, a compound of Na, Al, and F has the action of stabilizing the arc and, unlike Na20 or K20 etc., there are also cryolite (NasAlFe) and other non-oxide types. If other than an oxide, even if added to the welding wire, it will not form a source of supply of oxygen, so will not produce slag formed by oxides. It is therefore possible to add these arc stabilizers to meet with demands for stabilizing the arc more. For this reason, setting a range of possible use of arc stabilizers other than oxide types in the present invention has significance. The lower limit of 0.05% of the arc stabilizer other than an oxide type was set as the lowest limit of the value by which addition gives an arc stabilization effect. On the other hand, the upper limit of 0.5% was set because oxide-based slag materials were already added as a binder and these elements also act as arc stabilizers, so even if added over that, the effect would not change. The above were the reasons for limitation of the ingredients of the SiC ingredient-based wire. Next, the filling rate of the flux will be explained. In the present invention, the filling rate of the flux of the graphite ingredient-based wire is limited. The amount of production of slag after welding is reduced by limiting the ingredients of the flux filled in the wire, so the effect is obtained if in the range of the filling rate in ordinary metal-based flux cored wire. As shown in Japanese Unexamined Patent Publication No. 2001-179488, Japanese Unexamined Patent Publication No. 2001-287087, or Japanese Unexamined Patent Publication No. 2003-94196, there is no need to deliberately reduce the filling rate. Further, even if the flux filling rate is low, the effect can be sufficiently obtained if limiting the wire ingredients in the range of the present invention. For this reason, in the present invention, the range of the flux filling rate is limited when the amount of addition of alloy elements is large such as in a high fatigue strength weld material. For example, when trying to add Ni in wire limited in flux filling rate to 5% in an amount of 10% with respect to the wire as a whole, the Ni added to the flux is not sufficient alone. It is necessary to use the special steel outer skin to which Ni has been added. In this case, it is no longer possible to use an ordinary steel outer skin, so economic problems arise. If the filling rate is sufficiently high, this kind of problem does not arise. The lower limit of the flux filling rate was set as 10% as the range enabling sufficient freedom in wire ingredient design to be secured and a high strength weld material and high fatigue strength weld material to be achieved. The upper limit of 20% was set because if the filling rate is over this, the ratio of the steel outer skin in the wire becomes lower and the risk of breakage during wire production arises. Note that in the present invention, the filling rate of the SiC ingredient-based wire is not particularly limited. This is because SiC has a smaller action as a lubricant compared with graphite and therefore the amount of addition is set high and as a result the amounts of addition of the Ni and other alloy elements are set corresponding lower in the ingredient type and the lower limit does not have to be set to 10%. Further, the filling rate required for realizing SiC ingredient-based wire need only be in the range of the filling rate of ordinary flux cored wire, so the upper limit was also not set. Next, the shielding gas will be explained. In the gas shielded welding method, in general a shielding gas comprised of 100% CO2 or of Ar gas containing Co2 gas is used. Considering that the object of the present invention is to provide a method of formation of a high fatigue strength welded joint with a small amount of production of slag and that almost all of the slag is comprised of Sio2 or MnO or other oxides, it is desirable to select a shielding gas with a small oxygen content as well. For this reason, in the welding method in the present invention, it was decided to use shielding gas comprised of Ar + 3 to 25% C02 gas. Note that reducing the amount of C02 gas to 0% is not preferable in terms of the stability of the welding arc, so the Ar gas was stipulated as containing 3% or more of C02. With Ar gas containing over 25% of C02, the slag production becomes substantially the same as with 100% Co2 gas, so the upper limit was set as 25%. The 02 gas in the shielding gas is an impurity in the present invention. However, when Ar gas contains o2 gas, removal of the 02 gas costs money, so in general a shielding gas not containing 02 gas is more expensive than a shielding gas containing 02 gas. For this reason, the inventors considered that it would be meaningful to clarify the range of the allowable content of 02 gas and set the allowable range. When the amount of 02 gas is over 4%, an increase in the amount of production of slag cannot be avoided. For this reason, the upper limit of the 02 gas was set at 4%. Next, the thickness of the steel sheets and the steel sheet strength will be explained. First, the reason for limiting the thickness of the steel sheets will be explained. The present invention also has as its object the provision of a method of forming a high fatigue strength welded joint reduced in the amount of production of slag to the level of solid wire. In particular, regarding reduction of the amount of slag, even if not limiting the thickness of the steel sheets, the effect can be obtained if using metal-based flux cored wire in the range of the present invention. However, it is necessary to limit the steel sheet thickness from the viewpoint of forming a high fatigue strength welded joint. The principle behind the improvement of the fatigue strength in the present invention lies in the weld metal transforming during cooling and the expansion of volume of the weld metal at that time being utilized to reduce the residual welding stress at the parts where the fatigue becomes a problem. At this time, the expansion of volume of the weld metal has to be constrained by the steel sheet. That is, by having the expansion of volume constrained, a compressive stress is generated in reaction to this and the residual welding stress can be reduced. For this reason, it is necessary to set a lower limit of the thickness of the steel sheet. On the other hand, in the present invention, alloy elements are not added to the extent as shown in Japanese Unexamined Patent Publication No. 11-138290, so the transformation start temperature of the weld metal does not become as low as the weld material disclosed in Japanese Unexamined Patent Publication No. 11-138290. When the transformation of the weld metal ends, the weld metal shrinks in the cooling process after this, so even if introducing compressive stress, there is a possibility of a state of tensile stress being returned to. For this reason, in this art, the transformation start temperature tends to be set as low as possible. However, reducing the transformation start temperature means increasing the wire cost arid is therefore not preferable. Therefore, in the present invention, the upper limit of the thickness is limited so as to reduce the amounts of elements added as much as possible. This is because when the sheets are thin, compared with when they are thick, the welding heat ends up reaching the steel sheet back side at a comparatively early stage, so tensile stress becomes difficult to occur in the process of heat shrinkage after transformation of the weld metal ends. Regarding the lower limit of the thickness in the present invention, when the thickness is below 1 mm, even if using a metal-based flux cored wire in the range of the present invention to fabricate a joint, the depth of penetration with respect to the thickness becomes larger and even if the weld metal expands due to transformation, that expansion cannot be sufficiently constrained by the steel sheet, so the residual stress cannot be sufficiently reduced. That is, no improvement of the fatigue strength can be expected. For this reason, from the viewpoint of a high fatigue strength joint, the lower limit of the thickness was set to 1.0 mm. On the other hand, the industry in which the coatability of the welded joint becomes an issue is the automotive field. In the shipbuilding field etc., even if there is slag at the welding bead, no particularly large problem occurs. In general, in the automotive field, the thickness almost never exceeds 5 mm. The industry requiring such a thickness is the shipbuilding field. That is, it can be judged that there is no industrial merit. Further, if the thickness increases, as already explained, the welding heat has difficulty penetrating to the back side of the sheet, tensile stress ends up occurring in the process of heat shrinkage after the end of transformation of the weld metal, and no effect of improvement of the fatigue strength can be expected. For this reason, the upper limit of the thickness was set at 5.0 mm. Next, the reason for limiting the steel sheet strength will be explained. In the present invention, the reason which it is necessary to limit the steel sheet strength is also to improve the fatigue strength of the welded joint. It is not particularly required to limit it when the objective is to reduce the amount of the slag. When using the metal-based flux cored wire provided by the present invention to improve the fatigue strength of a welded joint, the means for realizing this is control of the residual stress of the weld zone utilizing the transformation expansion of the weld metal, that is, this method constrains the transformation expanding weld metal by the steel sheets to cause a reaction at both the weld metal and steel sheets. When the steel sheet strength is low, this reaction does not become sufficiently high and as a result the residual stress is not reduced. With regard to the weld metal, the alloy elements are already sufficiently added, so the problem of a low strength does not arise. For this reason, it is necessary to set a lower limit value for the steel sheet strength. The lower limit value of the steel sheet strength, 440 MPa, was set as the lowest limit of the value at which a sufficient reaction is obtained. On the other hand, the upper limit value of the steel sheet strength, 980 MPa, was set because with weld metal ingredients in the range of the present invention, the upper limit of strength of the weld metal ends up becoming 980 MPa or so and even if using steel sheets of higher strength, the joint strength would end up being defined by the weld metal, so it is judged that there would be no meaning to a higher value in practice. EXAMPLES Below, examples of the present invention will be explained. Example 1 Table 1 and Table 2 show the values of the ingredients in different metal-based flux cored wires. Table 1 shows the mass% of the ingredients added to each wire and the filling rate. The ingredients are shown by the mass% with respect to the total mass of each wire. Table 2 shows only the ingredients contained in the steel outer skin among the ingredients added to each wire. The ingredients in Table 2 are also shown by the mass% with respect to the total mass of each wire. That is, the ingredients added from the steel outer skin are only added in the amounts shown in Table 2. The remainders are added from the flux filled in each wire. Wire codes W01, 16, 17, and 19 are comparative examples, W01 is an example the same as Wll other than the graphite and having graphite outside the range of the present invention, and W16 and W17 have slag materials over the range of the present invention. W19 is the same as W14 in the wire ingredients other than the slag materials and has C included in the steel outer skin to make the amount of C the same as in W14. Further, the arc stabilizer in Table 1 is a compound of Na, Al, and F, that is, cryolite (Na3AlFe) . First, the inventors investigated the wire production efficiency. When producing the wires in Table 1, the wires other than W01 could be produced without any particular breakage during production. However, W01, one of the comparative examples, was the same as Wll other than the graphite, but broke in the middle of the process since no graphite was added and therefore could not be produced. W16 and W17 of the comparative examples are examples where no graphite is added, but had amounts of slag materials added at the level of conventional wire, so there was no particular problem in production. Next, as a mechanical characteristic, the inventors investigated the Charpy absorption energy - considered a problem when adding C. The Charpy absorption energy was investigated by preparing a 470 MPa class steel material having a thickness 3.2 mm, welding it by an I-groove, and obtaining a 2V notch Charpy test piece of a thickness of 2.5 mm from there. The thickness was set at 2.5 mm with a view to the fact that the main objective of the present invention is application to the automotive field. The notch position was designed to be at the center part of the weld metal for the purpose of investigating the characteristics of the weld material. The Charpy test was conducted at 0°C. Table 1 shows the results. The W01 wire was not subjected to a test since the wire broke during production. It is learned that the wires Wll to W18 and W20 have absorption energies over 20J and have sufficient joint characteristics. However, the W19 of the comparative example has C introduced from the steel outer skin, so the C of the weld metal becomes higher and the Charpy value is 11J or lower than the other wires. It is learned that the wire W19 has a total amount of C in the wire of 0.58% or the same amount of C as the wire W14 of the present invention. However, W14 wire has a Charpy value over 20J. It was learned that even with the same amount of C, the mechanical characteristics greatly differ between the case when added by graphite and the case when added from the steel outer skin. Table I (Table Removed) 1)Shows C other than graphite. 2) Shows Si other than SiO2. 3)"X" indicates breakage in the drawing process during wire production, while "O" shows no breakage. 4)Results of 2.5 mm thickness 2V notch Charpy test. The notch position is center part of weld metal. Table 2 (Table Removed) Next, the amount of production of slag was investigated. Among the wires in Table 1, except for W01 which broke during wire production, each of Wll to W18 was used for lap fillet welding. The weight of the slag produced at the bead surface was measured by the method of, first, measuring the weight of the test piece as a whole in the state with the slag present at the surface after the end of welding, then removing the slag, again measuring the weight of the test piece as a whole, and finding the difference of the two to thereby determine the weight of the slag. The test was run so that the length of the welding bead always became a constant 250 mm and so that the bead length was not affected. Table 3 shows the measurement results of the amount of the slag. From the results of Table 3, it is learned that the Wll, W12, W13, W14, W15, W18, and W20 with slag materials in the range of the present invention all had amounts of slag under 0.1 g, but when using the comparative examples of W16 and W17, the amount of the slag exceeds 0.3 g. Separate from the examples of Table 1, the amounts of the slag were measured by solid wire for the case of 100% C02 shielding gas and the case of Ar+20% C02 shielding gas, whereupon the amounts of production of slag were 0.09 g and 0.05 g, respectively. It was learned that the wires of the invention examples are suppressed in the amount of production of slag to the level of solid wire. From the above, it was learned that Wll, W12, W13, W14, W15, W18, and W20 suppressed in slag materials and in the range of the present invention had no problems in wire production efficiency, exhibited sufficient values of Charpy absorption energy, had sufficient mechanical characteristics of the joints, and had amounts of production of slag of the level of solid wire, that is, had amounts of slag sufficiently reduced. Table 3 also shows the results of the fatigue test. At this time, as the steel sheets, four types having a tensile strength of the 270 MPa, 470 MPa, 570 MPa, and 780 MPa classes were prepared. Table 3 shows combinations of wires and steel sheets. The test piece shape is one called a "lap fillet welded joint" shown in FIG. 5(a) and FIG. 5(b). First, the steel sheet 2 is superposed on a steel sheet 1 and is fillet welded to form a joint. Next, it is machined. The thicknesses 3 and 4 of the steel sheets 1 and 2 are as shown in FIG. 5(a) and FIG. 5(b). The hatching parts in FIG. 5(a) and FIG. 5(b) are the weld metal parts. The fatigue test was conducted by applying stress in the direction P of the arrow shown in FIG. 5(b). In this case, fatigue cracks occurred at the fillet weld toe. After this, they spread to the steel sheet 1 and finally ended with the steel sheet 1 breaking. That is, at this joint, the steel sheet with fatigue cracks indicates the steel sheet 1. Note that in these examples, the steel sheet 1 and steel sheet 2 do not necessarily have to be the same materials. Joints using different steel sheets may also be tested. Further, since the fatigue cracks occur at the steel sheet 1, the stress was measured by attaching a strain gauge at the weld toe, that is, in the vicinity of the welding bead of the steel sheet 1. Further, the fatigue limit was defined as the maximum stress at which no breakage occurred even after application of a repeated load for 2,000,000 cycles. Test No. 1 uses the wire Wll of Table 1 and has wire ingredients in the range of the invention examples, so is wire with a small amount of slag. However, the steel sheet 1 has a strength outside the range of the present invention, so the fatigue limit was 220 MPa or not a particularly high fatigue strength. On the other hand,' Test No. 2 has a strength of the steel sheet 1 where fatigue cracks occur of 470 MPa and realized a 2,000,000 cycle fatigue limit of 360 MPa, that is, had a high fatigue strength. On the other hand, Test Nos. 3 and 4 are cases where the thicknesses of the steel sheets were less than 1 mm and had 2,000,000 cycle fatigue limits of 250 and 260 MPa, that is, did not have high fatigue strengths. In each case, this is believed to be because the depth of penetration of the welding bead became relatively large compared with the thickness, the transformation expansion of the weld metal could not be sufficiently constrained, and the residual stress could not be reduced. As opposed to this, Test No. 5 had a sheet of a thickness of more than 1 mm and had a 2,000,000 cycle fatigue limit of 380 MPa, that is, had a high fatigue strength. Test No. 6 has wire ingredients in the range of the invention example and has an amount of the slag which is sufficiently reduced. Further, the strength of the steel sheet at which fatigue cracks occurs is high, so the 2,000,000 cycle fatigue limit is 360 MPa, that is, a high fatigue strength can be achieved. The same is true for Test No. 12. Test Nos. 7, 9, 10, and 13 have steel sheet thicknesses of 1 mm or more, strengths of 470 MPa or more, and wire ingredients all in the range of the present invention. The 2,000,000 cycle fatigue limits of the joints were all over 340 MPa. In particular, Test Nos. 7 and 13 using wires with larger total amounts of Ni, Cr, Mo, and Cu compared with Test No. 9 using the wire W13 have 2,000,000 cycle fatigue limits of over 400 MPa and have effects of improvement of the fatigue strength larger than Test No. 9. For this reason, it is learned that when aiming at a more reliable improvement of the fatigue strength, it is preferable to increase the amount of addition of the alloy element. On the other hand, Test No. 10 has wire ingredients substantially the same as in Test No. 9, but uses wire W14 containing B. In this case, the weld metal is improved in hardenability and the microstructure transforming at a low temperature can be increased over the case of Test No. 9, so the effect of improvement of the fatigue strength became larger than the case of Test No. 9. In this way, to more reliably realize a high fatigue strength, it is desirable to add B and/or to set the total amount of Ni, Cr, Mo, and Cu higher, etc. As opposed to this, Test Nos. 8 and 11 are cases where the wire ingredients are outside of the range of the present invention. The 2,000,000 cycle fatigue limits only reached 280 MPa, 260 MPa, and both of them are below 300 MPa. Test Nos. 14 to 21 are examples in the case where the strength of the steel sheet 1 is a comparatively high strength of 780 MPa. Test Nos. 14, 16, 17, 18, and 20 are cases where the thicknesses are over 1 mm, all of the wire ingredients are in the range of the present invention, the amounts of the slag are less than 0.1 g, and the fatigue strengths are all over 300 MPa, that is, high fatigue strengths are realized. As opposed to this, Test Nos. 15 and 19 are examples where the amount of the slag is not reduced and further are examples where the fatigue strength also does not become that high. Test No. 21 is an invention example using the wire W18 of Table 1 and therefore is an example where the amount of the slag is sufficiently reduced, but from the viewpoint of improving the fatigue strength, as shown in Table 2, the total amount of Cu, Ni, Cr, and Mo is 1.3% or lower than the wires Wll, W12, W13, W14, and W15 and the fatigue strength also does not become high. Similarly, Test No. 23 is a case the same as Test No. 5, that is, a case of using the wire W20 for a combination of the steel sheet 1 and 2. Test No. 23 is a case where both the steel sheet strength and wire ingredients are in the range of the present invention and therefore the amount of production of slag is small, but in the same way as Test No. 21, is a case where Cu, Ni, Cr, and Mo are not added, so is an example where the fatigue strength does not become higher. For this reason, it is learned that to reduce the amount of the slag while improving the fatigue strength, it is preferable to design the total of Cu, Ni, Cr, and Mo to be 2% or more. Test No. 22 is a result of the case of use of the comparative wire W19. In this case, the amount of C in the weld metal becomes higher, so the transformation temperature becomes lower and the fatigue strength can be sufficiently satisfied. However, as already explained, it is learned that the slag production and Charpy absorption energy are inferior to wire in the range of the present invention. Table 3 (Table Removed) As explained above, according to the present invention, the amount of production of slag of the metal- based flux cored wire can be suppressed to the same level as solid wire and the coatability of a welded joint fabricated using the metal-based flux cored wire can be greatly improved. Example 2 Table 4 shows the values of the ingredients of the metal-based flux cored wire. Table 4 shows the mass% and filling rate of the ingredients added to the wire. The ingredients are shown by mass with respect to the total mass of the wire. Table 4 (Table Removed) 1) Shows C other than SiC. 2) Shows Si other than SiC and SiO2. 3) Results of 2.5 mm thickness 2V notch Charpy test. The notch position is the center part of the weld metal. 4) Cracked, so no data could be obtained. Wires of the code in the 100 series in Table A are wires of invention examples corresponding to the SiC ingredient type in the present invention and their comparative examples. Further, in the #100 series wires, the #150 wire and higher are comparative wires for the SiC ingredient-based wires of the present invention. Further, wires of wire code of the #200 series are wires of the invention example comprised of wires containing graphite of the present invention plus small amounts of SiC and their comparative examples. Further, in the #200 series wires, wires of the #250 series are comparative wires for the wires of the present invention comprised of wires containing graphite plus small amounts of SiC. The C in the table shows the C other than SiC, while the Si in the table shows the Si of other than SiC and other than Si02. First, each wire shown in Table 4 was used to form a welded joint, then a test piece was taken from the weld zone and subjected to a Charpy test. The present invention has as its object the reduction of the amount of production of slag of the weld zone and consequent improvement of the coatability and improvement of the joint fatigue strength, but joint toughness is a basic characteristic of a welded joint, so the welded joint was tested in advance by a Charpy test to confirm the toughness of the weld zone of the two SiC ingredient-based wires of the present invention. The welded joint was formed and the Charpy test conducted by the following routine. First, two 780 MPa class steel sheets of thicknesses of 3.2 mm were prepared. The ends of these steel sheets were butt welded by an I-groove to form a welded joint, then a 2 mV-notch was formed by machining at the center part of the weld metal to form a Charpy test piece of a thickness of 2.5 mm. The sampling position of the test piece, as shown by the schematic view of FIG. 6, was made the position including the I-groove butt weld W, that is, the Charpy test sampling position S. This test piece was used at 0°C for a Charpy test. The absorption energy was measured. The results of the Charpy test are shown in Table 4. Note that the wire #152 in Table 4 suffered from solidification cracking at the welded joint, so could not be subjected to a Charpy test. The welded joint of the Sic ingredient-based wire in the range of the composition of ingredients prescribed by the present invention shown in Table 4 is of a level of joint toughness not posing any problem in practical use. Note that in Table 4, the wire of the comparative example of the wire code 151 has a Charpy absorption energy higher than the wire of the invention example, but as explained later, the ingredients for reducing the transformation temperature are outside the range of the present invention, so there was no effect of improvement of the joint fatigue strength. Next, the amount of production of slag of the weld zone and the joint fatigue strength were measured First, the method of measurement of the amount of production of slag was as follows: The amount of production of slag was measured by forming a welded joint for obtaining a fatigue test piece explained next, then measuring the weight of that joint. Next, the slag deposited on the surface was removed and the weight was again measured. Next, the difference between these weights was calculated and used as the amount of production of slag for that joint. Note that the welding joint for obtaining the fatigue test piece was formed so that the length of the welding bead always became a constant 250 mm, so wires were compared by comparing the amount of production of slag. The fatigue test piece was obtained from the joint after this.Next, the test piece taken from each welded joint was subjected to a fatigue test by the following routine to measure the joint fatigue strength. Two steel sheets were prepared, superposed, and fillet welded, the amount of production of slag was investigated, then a fatigue test piece shown in FIG. 5(a) and FIG. 5(b) was obtained from that joint. Further, a fatigue load was applied in the arrow direction of FIG. 5(b) as the fatigue load application direction P. The range of stress where no fatigue cracks occurred even with repeated application by 2,000,000 cycles was defined as the fatigue strength for that joint. This value was used for the comparisons. The stress ratio R was made R=0.1. The steel sheet 1 and the steel sheet 2 do not necessarily have to be the same steel sheets. Combinations of different strengths and thicknesses 3 and 4 were also selected. Note that the value of the stress applied to the test piece was measured by attaching a strain gauge near the welding bead on the surface of the steel sheet 1. First, each wire of the #100 series among the wires in Table A was used to form a welded joint under the conditions shown in Table 5, then the amount of production of slag of the weld zone and the joint fatigue strength were measured. The results are shown in Table 5. Test No. Al is an example having wire ingredients in the range of the present invention, so having an amount of production of slag of 0.07 g, that is, less than 0.1 g, and thereby achieving slag reduction. Further, to achieve a high fatigue strength, it is sufficient to raise the steel sheet strength as shown in Test No. A2. Test Nos. A3 and A10 are also examples having wire ingredients in the range of the present invention, so having amounts of production of slag of less than 0.1 g and thereby achieving slag reduction. Further, to achieve a high fatigue strength, as in Test No. A4 or A2, it is sufficient to adjust the thickness of the steel sheet. The comparative example of Test No. All has a steel sheet strength and steel sheet thickness both in the range of the present invention and has an SiC content in the wire ingredients in the range of the present invention, so improvement of the fatigue strength is achieved, but the total content of the slag materials in the wire ingredients was higher than the range of the present invention, so reduction of the amount of the slag could not be achieved. The comparative examples of Test Nos. A12 and A13 have SiC contents in the wire ingredients and total contents of the slag materials outside of the range of the present invention, so neither an improvement of the fatigue strength nor reduction of the amount of production of slag could be achieved. Further, the wire of the comparative example of the wire no. 152 shown in Table 4 ended up cracking at the time of the Charpy test, so the fatigue test could not be conducted. On the other hand, in the invention examples of Test Nos. A2 and A4 to A9, the fatigue strengths are all over 300 MPa and amounts of production of slag of less than 0.1 g can be achieved. By the above, in all of the invention examples of Al to A10 of Table 5, amounts of production of slag of less than 0.1 g can be achieved, while in the comparative examples of All to A13, the amounts of the slag are all 0.5 g or more. It is learned that by using the present invention, slag reduction can be achieved. Further, if adjusting the steel sheet strength and thickness, as shown in Test Nos. A2 and A4 to A9, the fatigue strength can be made 300 MPa or more. Table 5 (Table Removed) Next, each wire of the #200 series in the wires of Table 4 was used to form a welded joint under the conditions shown in Table 6, then the amount of production of slag of the weld zone and the joint fatigue strength were measured. The results are shown in Table 6. The amounts of production of slag and fatigue strengths of Table 6 are for examples of the cases of wires containing graphite of the present invention plus small amounts of SiC. The method of investigation of the fatigue strength and the procedure for investigation of the amount of production of slag are as already explained. Test No. B3 is an example having wire ingredients in the range of the present invention and having an amount of the slag of 0.09 g, or less than 0.1 g, that is, achieving slag reduction. However, steel sheet has a small thickness and the transformation expansion of the weld metal could not be sufficiently constrained, so the fatigue strength was not improved. To improve the fatigue strength for Test No. B3, it is sufficient to adjust the thickness of the steel thickness as with Test No. B4. Test No. B9 is an example where the wire ingredients are in the range of the present invention and the amount of the slag is 0.05 g, or less than 0.1 g, that is, slag reduction could be achieved. However, the steel sheet has a large thickness, so the welding heat could not sufficiently reach the back side of the sheet, stress again changed to tensile stress after the transformation of the weld metal ended, and the fatigue strength was not improved. To improve the fatigue strength for Test No. B9 as well, it is sufficient to adjust the thickness of the steel sheet as in Test No. B8. Further, the comparative example of Test No. BIO has a steel sheet strength and a steel sheet thickness in the range of the present invention and has a SiC content and total content of Cu, Ni, Cr, and Mo in the wire ingredients in the range of the present invention, so the fatigue strength could be improved, but the total content of the slag materials in the wire ingredients was a high one outside the range of the present invention, so a reduction of the amount of the slag could not be achieved. The comparative example of Test No. fill has a low total content of Cu, Ni, Cr, and Mo in the wire ingredients outside the range of the present invention, so the reduction of the transformation start temperature of the weld metal was insufficient and therefore the fatigue strength could not be improved. Test Nos. B4 to B8 are examples where the wire ingredients and steel sheet strength and thickness are in the range of the present invention. In these examples, slag reduction resulting in an amounts of production of slag of less than 0.1 g in all cases can be achieved and high fatigue strength joints having fatigue strengths all over 300 MPa can be achieved. From the above, it is learned that slag reduction is achieved only when the wire ingredients are in the range of the present invention and that a higher fatigue strength is achieved only when the steel sheet strength and thickness are in the range of the present invention. Table 6 (Table Removed) INDUSTRIAL APPLICABILITY According to the present invention, it becomes possible to keep the amount of the slag produced when arc welding by metal-based flux cored wire to the level of solid wire. Since the amount of production of slag is small, the welded joint formed can be coated without going through a slag removal process and it becomes possible to maintain the efficiency of current automobile production processes using solid wire as it is. Further, it becomes possible to limit the ingredients in the wire and thereby greatly improve the fatigue strength of the welded joint, so the method of formation of a welded joint provided by the present invention can form a high fatigue strength welded joint while maintaining a high automobile production efficiency. 'We claim, 1. Metal-based flux wire for gas shield arc welding; comprising a steel outer skin in which a flux is filled, characterized in that: said metal-based flux wire contains, by mass percentage of the steel outer skin and metal-based flux filling in total. C other than graphite and other than Sic: 0.00 1 to 0.20%, graphite: 0.10 to 0.7%, Si other than Sic and other than Si02: 0.05 to 1.2%, and Mn: 0.2 to 3.0%, restricted to P: 0.03% or less and S: 0.02% or less, and containing one or more of Si02, AI2O3, Na20, and K20 in a total of 0.05 to C.O% and the balance of iron and unavoidable impurities, said graphite and one or more of Si02, A1203, Na20, and KO contained as at least said flux. 2.The metal-based flux wire as claimed in claim I, wherein said metal-based flux wire further contains, by mass% of the steel outer skin and metal-based flux filling in total, Sic; 0,05 to 0.5%. 3. 'The metal-bascd flux wire as claimed in claimed in I or 2. v,iherein said metalbased flux wire has a flux filling rate of 10 to 20%. 4. The metal-based flux wire as claimed in any one of claims I to 3, wherein said metal-based flux wire further contains, by mass% of the steel outer skin and metalbased flux filling in total, one or more of Ni: 0.5 to 12.0%, Cr: 0.1 to 3.0%, Mc: 0.1 to 3.0%, atid Cu: 0.1 to 0.5% in a total amount 01'0.2 to 12.5%. 5.The metal-based flux wire as claimed in any one of claims 1 to 4, wherein said metal-based flux wire further contains, by mass% of the steel outer skin and metalbased flux filling in total, B: 0.001 to 0.03%. 6. The metal-based flux wire as claimed in any one of claims I to 5, wherein said metal-based flux wire further contains, by mass% of the steel outer skin and metalbased flux filling in total, one or more of Nb, V, and Ti in a total amount of 0.005 to 0.3%. 7. The metal-based flux wire as claimed in any one of claims 1 to 6, wherein said metal-based flux wire further contains an arc stabilizer other than an oxide system, by mass% of the steel outer skin and metal-based flux filling in total, in an amount of 0.05 to 0.5% as said flux. 8. A gas shield arc welding method for formation of high fatigue strength welded joint producing a small amount of slag, co~nprisingm etal-based flux wire as claimed in any one of claims 1 to 7 for welding steel sheets and wherein the thickness of said steel sheet is 1.0 to 5.0 ~nman d thc tensile strength is 440 to 980 MPa.

Documents:

7534-delnp-2007-Abstract-(19-02-2014).pdf

7534-delnp-2007-Abstract-(19-12-2012).pdf

7534-delnp-2007-abstract.pdf

7534-delnp-2007-Claims-(19-02-2014).pdf

7534-delnp-2007-Claims-(19-12-2012).pdf

7534-delnp-2007-claims.pdf

7534-delnp-2007-Correspondence Others-(19-02-2014).pdf

7534-delnp-2007-Correspondence Others-(19-12-2012).pdf

7534-delnp-2007-Correspondence Others-(27-11-2013).pdf

7534-delnp-2007-Correspondence-others (04-02-2008).pdf

7534-delnp-2007-correspondence-others 1.pdf

7534-delnp-2007-correspondesce-others.pdf

7534-delnp-2007-Description (Complete)-(19-12-2012).pdf

7534-delnp-2007-description (complete).pdf

7534-delnp-2007-Drawings-(19-12-2012).pdf

7534-delnp-2007-drawings.pdf

7534-delnp-2007-form-1.pdf

7534-delnp-2007-Form-13 (04-02-2008).pdf

7534-delnp-2007-form-18.pdf

7534-delnp-2007-Form-2-(19-02-2014).pdf

7534-delnp-2007-Form-2-(19-12-2012).pdf

7534-delnp-2007-form-2.pdf

7534-delnp-2007-form-26.pdf

7534-delnp-2007-Form-3-(19-12-2012).pdf

7534-delnp-2007-form-3.pdf

7534-delnp-2007-form-5.pdf

7534-delnp-2007-GPA-(19-12-2012).pdf

7534-delnp-2007-pct-210.pdf

7534-delnp-2007-pct-304.pdf

7534-delnp-2007-pct-308.pdf