Visualization and Analysis of Groove Residual Magnetism for Narrow Gap Arc Welding

Abstract

Residual magnetism obviously exists in welding groove of high strength steel and thus influences narrow gap arc welding quality. The present work investigates the characteristics of residual magnetism for U-shape welding groove, and reveals its formation mechanism by physically modeling the residual magnetism. An experimental data based visualization approach is then proposed to characterize dominant distributions of residual magnetism. It is shown that residual magnetism is much stronger in groove width and at groove bottom respectively than in groove length and at groove top, as well as of bar magnet features for each groove sidewall. Two-dimensional visualization of residual magnetism vector is realized by plane and curved surface cloud maps of equivalent color bars, while three-dimensional visualization is expressed by assembling six outermost plane cloud maps of residual magnetism in groove measuring space. The built models and digital visualizations well contribute to understand the phenomena of residual magnetism.

1. Introduction

Narrow gap welding is an attracting process of high efficiency and high quality, and has been used in thick-wall steel structure manufacturing.¹⁾ To further improve its groove sidewall penetration and process adaptability in gas metal arc (GMA) welding, several single-wire weaving arc approaches appeared to replace traditional rotation arcs,^2,3) typically such as electromagnetic oscillation arc⁴⁾ and mechanical swing arc.^5,6,7) Nevertheless, the narrow gap GMA processes are subject to welding groove conditions as well as arc movements.

In practice, ferromagnetic base material is easily magnetized during storage, transportation, machining and welding^8,9) particularly when containing high contents of Ni and Cr etc. alloying elements, because of its unavoidable exposure to various magnetic fields, which causes obvious residual magnetism to occur in narrow gap welding groove.^9,10) This residual magnetism deflects the arc by Lorentz force, and finally influences welding quality due to arc blow. Therefore, it is considerably significant for reducing this harm to reveal characteristics of residual magnetism.

Badheka et al. measured the residual magnetic amplitudes of narrow gap square groove with a metallic backing strip after various welding runs, and investigated the effect of residual magnetism on sidewall fusion for narrow gap GMA welding.⁹⁾ Sun et al. simply analyzed the reason of residual magnetism in a high strength steel of ~800 MPa grade, and used a one-dimensional gauss meter to detect residual magnetism in square groove with ultra narrow gap of 5 mm which was directly machined from a whole thick steel plate.¹⁰⁾ However, scare systematic research is addressed to the residual magnetism of narrow gap arc welding groove, especially for U-shape groove which is another type frequently utilized.

For a deep understanding of residual magnetism phenomena and further reducing appropriately this harm, this work aims to investigate in detail two-dimensional vector distributions of residual magnetism in U-shape grooves of high strength steel, as well as to analyze its forming mechanism by physically modeling residual magnetism, and finally visualizes the dominant vector of residual magnetism by two- and three-dimensional cloud maps on Matlab software platform.

2. Experimental Procedure

A high strength alloy steel S690QL of 690 MPa grade, commonly applied in shipbuilding and offshore engineering etc, was chosen as experimental base material. As indicated in Table 1, this metal composes of more Ni, Cr, Si and Mo etc alloying elements, which are considered as primary factor inducing residual magnetism. To assemble a U-shape groove of 290 mm length and 1 mm root gap, two groove sidewalls of 50 mm thickness and 100 mm width were cut from the same steel plate, and then machined according to groove bottom dimensions as given in Fig. 1. Furthermore, this groove was paced on an electrical insulating epoxy resin board 24 hours before measuring, so as to form a stable residual magnetic field.

Table 1. Main chemical composition of experimental material.

Elements	C	Si	Mn	Cr	Ni	Mo	Nb
Mass%	0.16	0.23	1.4	0.23	2.0	0.7	0.06

Fig. 1.

Coordinate system of narrow gap arc welding U-shape groove. (Online version in color.)

To conveniently describe spatial positions to be measured before welding, a three-dimensional rectangular coordinate system is established in Fig. 1, where the origin O of coordinates is located at diameter line of groove arc surface and 2.5 mm away from left sidewall on end plane of groove, and where the axes of x, y and z were set respectively in the directions of groove width, length and depth. In actual measuring, a Cartesian coordinate manipulator of three freedoms precisely moved a Hall probe of two-dimensional digital gauss meter to shortly reside at measuring point and to simultaneously detect at 0.0001 mT resolution residual magnetic vectors that point to x and y axial directions, in which the two components are accordingly defined as B_x and B_y.

3. Characteristics and Modeling of Residual Magnetism

3.1. Characteristics

Figures 2 and 3 show the distributions of residual magnetic intensity at groove center (i.e. coordinate x=4 mm) along length y and depth z directions of two different U-shape grooves, where 30 detecting points were planned along groove length at measuring interval of 10 mm respectively for four various groove depths, and where the grooves U₁ and U₂ were assembled in opposite magnetic poles (see also Fig. 4(a)). In the legends, the indexing number of the letter Z indicates a z-directional position coordinate value of measured point, for example, Z₋₅ denotes the coordinate value z=−5 mm, and so on. Moreover, the positive and negative values of the components B_x and B_y mean that the vectors respectively point to the forward and opposite directions of corresponding axes.

Fig. 2.

Residual magnetism at different lengths of groove U₁ (x=4 mm). (a) B_x; (b) B_y. (Online version in color.)

Fig. 3.

Residual magnetism at different lengths of groove U₂ (x=4 mm). (a) B_x; (b) B_y. (Online version in color.)

Fig. 4.

Distribution model of residual magnetic flux lines for U-shape groove assembled in opposite magnetic poles. (a) Pattern A along groove length; (b) Pattern B along groove depth.

Both B_x and B_y components of residual magnetism are always large at both ends and small in the middle along groove length, and the directions of which are reversed near the middle of groove length regardless of groove depth. At both ends of groove, the two components slightly decrease due to outside escape of magnetic field lines, as compared to their own neighboring inside regions. Along groove depth, the residual magnetism is large at the bottom of groove subject to low magnetic reluctance from narrow air gap, and becomes obviously small near the top of groove because of both long magnetic path and partial escape of magnetic flux lines. Particularly, the amplitude of B_x is much greater than that of B_y, in which the former is about ten times the latter, reaching to a maximum value of ~9 mT. It is also shown that the overall changing tendencies of residual magnetism are similar although its amplitudes differ for different grooves.

3.2. Modeling

From the above features of residual magnetism, each groove sidewall can be regarded as a bar magnet, thus a physical distribution model of residual magnetic flux lines is accordingly reasoned for the U-shape groove assembled in opposite magnetic poles, as illustrated in Figs. 4(a) and 4(b). This model covers two primary patterns, that is, patterns A and B. In the pattern A, as shown in Fig. 4(a), the field lines traverse through narrow groove gap from N to S magnetic poles by the shortest paths and become gradually dense from the middle toward both ends of groove length, which accordingly generates the components of B_x and B_y. The pattern B, as given in Fig. 4(b), plots a distribution diagram of residual magnetism along groove depth and primarily leads to the occurrences of B_x and z-directional magnetic vector B_z at different depths, where the field lines get obviously dense near groove root because of narrowing air gap and appear to escape from both ends of groove.

It can be seen from patterns A and B that the values of components B_y and B_z are considerably small relative to that of vector B_x, which has been proved by experimental results. Moreover, two sidewalls may be combined in similar magnetic poles to form a narrow gap welding groove. Consequently, another physical model of residual magnetic flux line distribution is shown in Fig. 5, that is, pattern C. In this case, the field lines return mainly from the N to S magnetic poles of each sidewall, and finally groove residual magnetism entirely becomes weaker than it does in opposite magnetic polarity combination because of longer magnetic paths. Noticeably, the shape and distribution of field lines are usually asymmetrical, since the magnetized degree of two groove sidewalls always differs. The models including three patterns distinctly express distribution characteristics of magnetic poles and flux lines, and thus further aid to understand the phenomena of residual magnetism in the two assembled types of grooves.

Fig. 5.

Distribution model (Pattern C) of residual magnetic flux lines for U-shape groove assembled in similar magnetic poles.

4. Visualization of Residual Magnetism

4.1. Visualization Approach

General scheme is to visualize residual magnetism based on actual experimental data as well as on Matlab software platform. In measuring, the width, length and depth of groove are divided into 4, 29 and 9 equal segments respectively in positive directions of the above x, y and z coordinate axes, except that groove depth is added by another equal in negative direction of z axis for the measurements in the lengthwise central plane of groove. Accordingly for any one of x-y, x-z and y-z measuring planes and those which parallel them, a measuring point matrix P forms below   

$$P=\left[\begin{array}{cccc}{p}_{11}& p_{12}& \cdots & p_{1j}\\ p_{21}& p_{22}& \cdots & p_{2j}\\ ⋮& ⋮& ⋮& ⋮\\ p_{i1}& p_{i2}& \cdots & p_{ij}\end{array}\right]$$(1)

where (i-1) and (j-1) respectively correspond to the divided numbers of the shorter and longer axes of measuring plane. According to such a plan of measuring points, residual magnetism is sequentially detected by the Hall probe in the row direction of the matrix P, and a measured data matrix M that contains i×j values is finally built as   

$$M=\left[\begin{array}{cccc}{m}_{11}& m_{12}& \cdots & m_{1j}\\ m_{21}& m_{22}& \cdots & m_{2j}\\ ⋮& ⋮& ⋮& ⋮\\ m_{i1}& m_{i2}& \cdots & m_{ij}\end{array}\right]$$(2)

In visualizing, the data matrix M is loaded on Matlab software platform, and the functions Surf and Colormap are successively called to realize the two-dimensional visualization of residual magnetism by plane and curved surface cloud maps of some equivalent color bars, while the three-dimensional visualization is expressed by assembling six outermost plane cloud maps of the residual magnetism in groove measuring space.

4.2. Two-dimensional Visualization

4.2.1. Visualization on Groove Width-length Plane

The following visualizations are focused on the component B_x of residual magnetism for the grooves assembled in opposite magnetic poles, due to the above fact that B_x is always much larger than B_y and B_z, and that B_x is remarkably greater in this assembling type than in the other one. Figures 6 and 7 give distributions of B_x at the lengths and widths for grooves U₁ and U₂, where groove depth coordinate z=0 and the indexing numbers i and j in the matrixes P and M equal respectively to 5 and 30, and where concurrently gives curved surface and its downward projection maps for better observation. The plane and curved surface cloud maps were estimated from 150 measured values, while the displayed values of residual magnetism were divided into 15 equals by different color segments.

Fig. 6.

Distribution of B_x at the lengths and widths of groove U₁ (z=0). (a) Curved surface cloud map; (b) Plane cloud map.

Fig. 7.

Distribution of B_x at the lengths and widths of groove U₂ (z=0). (a) Curved surface cloud map; (b) Plane cloud map.

Clearly, the color and width of the segment goes deep and wide toward both sides of groove middle, which interprets that the residual magnetism becomes serious while its variation rate tends to be small. Furthermore, the residual magnetism of two grooves distributes asymmetrically along groove length, where the magnetic neutral area that corresponds to the lightest color occurs respectively at y=~142 mm and ~126 mm. The boundary line of equivalent color bar gradually bends from the middle to both ends of groove, which suggests that the x-axial difference in residual magnetism accordingly increases in width direction of groove. However, this difference is actually very small, merely being several gausses, so B_x distributes approximately in an equal-intensity state along groove width. The similar varying tendencies exhibit in both plane and curved surface cloud maps of residual magnetism, although their values and distribution uniformities differ always in different grooves, for example, the groove U₁ obviously differing from the groove U₂.

4.2.2. Visualization on Groove Length-depth Plane

Figures 8 and 9 show the distributions of B_x in lengthwise central plane for grooves U₁ and U₂, where the groove width coordinate x=4 mm and the negative value of z coordinate represents the bottom position of U-shape groove (see Fig. 1). The plane and curved surface cloud maps were inferred from 300 measured values, since i=10 and j=30 for the matrixes P and M. To distinctly demonstrate the features of residual magnetic distribution in groove bottom, 28 color segments of equivalent values were created to separate different intensities. Note that light color blocks have to be used in expressing the large absolute amplitudes for the deeper groove places which correspond to some negatives of z coordinate, although many attempts were made to distinguish them from the light color bars of the lowest intensities in groove middle. Fortunately, the intensity of residual magnetism can be additionally identified from the curved surface map.

Fig. 8.

Distribution of B_x at the depths and lengths of groove U₁ (x=4 mm). (a) Curved surface cloud map; (b) Plane cloud map.

Fig. 9.

Distribution of B_x at the depths and lengths of groove U₂ (x=4 mm). (a) Curved surface cloud map; (b) Plane cloud map.

From the top to bottom of groove, as well as toward both sides from the magnetic neutral area that corresponds to the lightest color in groove middle, the color of equivalent bar gradually goes deep beyond the zero line of z coordinate, and accordingly the residual magnetism grows. Below the zero line of z coordinate, that is, toward the bottom of U-shape groove, the color block gets close and its color becomes light, which, together with the curved surface maps, thus indicates a remarkable rise of residual magnetic intensity due to narrowing gap. Consequently, the obvious convex and concave occur respectively at both lengthwise ends of groove bottom in the curved surface map. Because of the utilization of much more colors in the two cases than in Figs. 6 and 7, the reverse turns of boundary line of color bar clearly exhibit near both lengthwise ends of groove, which exactly demonstrates that the magnetic intensity a little falls subject to partial escape of magnetic flux lines.

Probably resulting from uneven magnetization, moreover, the inner boundaries of some color bars appear to bulge toward groove middle beyond zero line of z axis for both grooves U₁ and U₂, which is particularly obvious at great groove lengths and suggests that the residual magnetism does not simply get serious toward lower parts of groove. These inward bulges in varying degrees, together with the other magnetic differences between the grooves, further demonstrate that the distributing details of residual magnetism in various grooves always differ in spite of similar general laws.

4.2.3. Visualization on Groove Width-depth Plane

Figures 10 and 11 indicate the distributions of B_x at the depths and widths for grooves U₁ and U₂, where the values of residual magnetism are all positive since they are observed on cross sections at groove length coordinate y=30 mm. Each set of maps were actually suggested from 45 measured values at i=5 and j=9 in Eq. (2), and separately contain 23 color segments which respectively represent an incremental value of ~0.067 mT.

Fig. 10.

Distribution of B_x at the depths and widths of groove U₁ (y=30 mm). (a) Curved surface cloud map; (b) Plane cloud map.

Fig. 11.

Distribution of B_x at the depths and widths of groove U₂ (y=30 mm). (a) Curved surface cloud map; (b) Plane cloud map.

From the bottom to the top of groove, that is, with an increase in groove depth coordinate z, the residual magnetism goes weak. Toward the upper part of groove from z>~35 mm, the escape of magnetic flux lines significantly increases, which accordingly causes the intensity of residual magnetism to rapidly fall (obviously in Figs. 10(a) and 11(a)), and thus the color bar becomes considerably dense. From z=~10 to 0 mm, the color segment gets narrow again, which shows that a rise of residual magnetism is slightly accelerated owing to a centralizing tendency of magnetic flux lines to groove bottom. Furthermore, these dense color bars of curved boundary lines actually suggest greater x-axial fluctuations of residual magnetism at lower regions of groove.

4.3. Three-dimensional Visualization

To more intuitively demonstrate and comprehensively understand distribution laws of residual magnetism in narrow gap welding groove, Fig. 12 plots three-dimensional distributions of B_x for the grooves U₁ and U₂ from a spatial perspective. Here, two-dimensional cloud maps were generated respectively in the width-length, length-depth and width-depth planes of groove space from 150, 270 and 45 actual measured data, and were then assembled spatially in the same color mark system containing 20 bands.

Fig. 12.

Three-dimensional distribution of B_x. (a) Groove U₁; (b) Groove U₂.

Obviously, toward both ends of groove from the magnetic neutral area which corresponds to the lightest color mark, the color bar becomes gradually dark and wide, and accordingly the residual magnetism gets serious while its growing rate slows down. Although the observed faces differ from the above ones and the displayed values vary as well, the distributing laws of residual magnetism in the width-length, length-depth and width-depth planes of groove are respectively similar to those in Figs. 6 and 7, Figs. 8 and 9, and Figs. 10 and 11. Particularly, these color bands are considerably well joined among different detected planes, thus confirming the accuracy of detected results and the effectiveness of visualization approach from another side.

5. Conclusions

(1) Each separated sidewall of narrow gap welding groove can be regarded as a bar magnet of nonuniform magnetic field. For U-shape grooves where two sidewalls are assembled respectively in opposite and similar magnetic poles, the proposed physical models of residual magnetism can well interpret the distribution features of residual magnetic vectors.

(2) For U-shape groove assembled in opposite magnetic poles, the components of residual magnetism in groove width and length directions rise from the middle to both ends of groove and occur in opposite direction at the both ends, the former component of which increases from the top to bottom of groove but appears nearly uniform along groove width.

(3) An experimental data based visualization approach is presented to characterize two- and three-dimensional distributions of dominant residual magnetic vector. The two-dimensional visualization is realized by plane and curved surface cloud maps of equivalent color bars, while the three-dimensional visualization is expressed by assembling six outermost plane cloud maps of residual magnetism in groove measuring space.

Acknowledgements

This work is supported by the National Natural Science Foundation of China (Grant No. 51475218), and the Specialized Research Fund for the Doctoral Program of Higher Education of China (Grant No. 20133220110001).

References

• 1)   M.  Murayama,  D.  Oazamoto and  K.  Ooe: JFE Tech. Rep., 20 (2015), 147.
• 2)   J. Y.  Wang,  Y. S.  Ren,  F.  Yang and  H. B.  Guo: Sci. Technol. Weld. Join., 12 (2007), 505.
• 3)   N.  Guo,  M. R.  Wang,  W.  Guo,  J. B.  Yu and  J. C.  Feng: Int. J. Adv. Manuf. Technol., 75 (2014), 15.
• 4)   Y. H.  Kang and  S. J.  Na: Weld. J., 82 (2003), 93s.
• 5)   J.  Wang,  J.  Zhu,  P.  Fu,  R.  Su,  W.  Han and  F.  Yang: ISIJ Int., 52 (2012), 110.
• 6)   J.  Wang,  J.  Zhu,  C.  Zhang,  N.  Wang,  R.  Su and  F.  Yang: ISIJ Int., 55 (2015), 1076.
• 7)   J.  Wang,  J.  Zhu,  C.  Zhang,  G.  Xu and  W.  Li: ISIJ Int., 56 (2016), 844.
• 8)   M.  Farman,  A. U.  Haq,  Q. R.  Ali and  A. Q.  Khan: Mater. Des., 9 (1988), 263.
• 9)   V. J.  Badheka and  S. K.  Agrawal: Weld. Cutt., 7 (2008), 280.
• 10)   F.  Sun,  G. D.  Zhang,  L. Q.  Ma and  F. J.  Zhang: Electr. Weld. Mach., 35 (2004), 56 (in Chinese).