John F. Kiefner
Kiefner & Associates Inc.
Worthington, Ohio
Understanding the manufacture and unique problems of electric-resistance welded (ERW) line pipe can aid operators in safely maintaining and using existing ERW pipe and in sharpening key requirements when they specify new ERW materials.
Evidence that inferior ERW pipe has been manufactured and installed, especial before 1970, abounds not only in pipeline accident statistics1-2 but also in the investigations of individual pipeline failures which seldom appear in open literature.
Abandoning or replacing large quantities of ERW line pipe is impractical and undesirable.
This first of two articles briefly reviews what constitutes ERW pipe, the weaknesses and deficiencies that can occur in ERW materials, and their possible effects on pipeline safety.
The concluding article offers some suggestions for continuing to operate such pipelines with a minimum of risk to the public.
ERW SEAMS
First and foremost, it must be made clear that when the objective of ERW pipe manufacturing is met, namely, when the microstructural appearance and the mechanical properties of the bond line region are, in fact, virtually indistinguishable from or superior to those of the parent skelp, the quality of the product is at least equal to and can be superior to pipe made by any other method.
The accompanying box reviews the ERW manufacturing process and provides definitions of key terms.
HF WELDS
Fig. 1 shows a polished and etched section across a typical high-frequency (HF) ERW seam.
The first things to note are that the top surface represents the ID surface of the pipe; the bottom surface, the OD surface. The skelp material in this case is a conventional carbon-manganese X-60 Grade steel with a banded pearlite-ferrite microstructure.
Characteristic of HF welding, the weld heat-affected zone (HAZ) is hourglass shaped. This shape arises because of how the radio-frequency current tends to concentrate near the surfaces.
The bond line is highlighted as a white line because just prior to bonding the edges were hot enough to have undergone some decarburization. There is nothing unusual or degrading about this bond line.
This bond line (oriented horizontally in Fig. 2) is much like the surrounding microstructure. In Fig. 1, the post-weld heat treatment can be seen to have transformed the microstructure on both sides of the bond line over a width equal to about twice the W.T. and that it has penetrated the entire wall rather uniformly.
That this heat treatment was successful is further explained by Fig. 2 which shows that the HAZ consists of fine equiaxed ferrite grains.
The remaining features to note in this typical HF ERW seam weld (Fig. 1) are the degree of upset at the inside surface, the flash trim at both surfaces, the contact marks, and the upturned fiber pattern of the microstructure.
First, a thickening of the material in two locations just beyond the edges of the HAZ is noticeable. Actually,
before the flash was trimmed, the entire weld zone was thicker as the intentional result of upsetting.
The extra upsetting is done intentionally for several possible reasons. One reason can be to compensate for a possible loss of strength from post-weld heat treating. Another is to increase the chances that oxides and scale on the edges of the skelp will be forced out of the bond line region.
The ID surface trim is usually arc shaped (Fig. 1). At the OD surface, however, the trim is flush so that nothing can interfere with the inspection of the weld.
The dark areas of the microstructure at the intersection of the edges of the HAZ and the OD surface are contact marks. These correspond to the path of the sliding contacts and are points of very high current.
Often these zones are transformed to austenite and may upon cooling have a somewhat different microstructure from that of the original skelp. ln some cases, these zones may exhibit excessive hardness not unlike arc burns.
The final feature of note in Fig. 1 is the upturning of the ferrite-pearlite banded structure of the skelp. Because in the skelp the banding is parallel to the surfaces of the skelp, it ends up being parallel to the bond line adjacent to it.
LOW FREQUENCY
A typical older ERW weld is shown in Fig. 3. This photograph illustrates a direct current welded (DC) ERW seam but the appearance of a typical low-frequency (LF) ERW seam is virtually the same. Note the wide weld HAZ which results from the deeper penetration of direct or low-frequency current.
This particular weld, although sound in terms of the absence of defects, exhibits some of the undesirable features of many older ERW seams. The bond line itself is rather wide and surrounded by grain-coarsened base metal, a sign of excessive heat. Also, the contact marks are essentially contact burns.
Like many older ERW seams, the ID flash has been trimmed but not to the extent found in HF ERW seams. A typical flash-welded (FW) seam, like the LF ERW weld, exhibits a distinct bond line and a wide HAZ. Characteristic features of an FW seam are flash trim and a wide, square flash at both the ID and OD surfaces. Some ERW manufacturers were known to have left a significant OD flash as well as an ID flash but none is as wide as that of a true FW seam.
SEAM DEFECTS
Because several confusing terms have arisen to refer to imperfections in ERW welds, the present. discussion will use the terminology found in API Bulletin 5TI (9th Ed.), "Definitions of Imperfections and Defects Occurring in Electric Resistance Welds."3
For example, Fig. 4 illustrates a hydrostatic test break which originated at a series of defects in an LF ERW seam.
The tell-tale black oxide patches visible on the fracture surface indicate one of the typical defects in ERW seams, namely, "cold-weld" zones.
According to API Bulletin 5TI, a cold weld is "a lack of adequate bonding" and is caused by application of "insufficient heat and/or pressure." It "may or may not have separation in the weld line."
More examples of cold welds will be presented later.
Another term, "misalignment," does not refer to a defect but to a condition which may result in defects or conditions which will later produce defects.
That's why misalignment is not by itself a defect. Many types of misalignment arise in the making of ERW welds.
The edges of the skelp may be misaligned vertically, creating a kind of high-low condition not actually defined in API Bulletin 5TI or in API Specification 5L.
More common, however, is the misalignment of various steps in the process resulting from cambered skelp and the tendency of the skelp to twist as it moves through the forming and welding stands. One possible result is the misalignment of the post-weld heat treatment (Fig. 5).
In this case, the heat treatment was offset from the bond line probably because the skelp was twisting through the mill. One way to reduce these kinds of problems is to institute automatic feedback controls. This involves transducers and limit switches monitored by a computer which senses irregularities as the skelp moves through the mill and makes appropriate corrections.4
Another way that some mills ensure that the seam gets normalized is to have multiple offset normalizers.
COLD WELDS
Colds welds, already mentioned, can be a significant problem in older ERW materials. Fortunately, the incidences of cold welds have diminished significantly in materials manufactured in the last 20 years, especially since the advent of HF and high-frequency induction (HFI) welding.
Nevertheless, the condition may still occur even in an HF or HFI material. Examples of cold welds are shown in Fig. 6.
Fig. 6a is a metallographic section through a cold weld in an LF ERW seam. The smooth surfaces, except for two small "tabs," suggest that no bonding occurred.
The typical appearances of fracture surfaces involving cold welds are shown in Figs. 6b and 6c.
Fig. 6b shows a series of black oxide patches where no bond existed. The repetitive pattern accompanying these patches is "stitching," a phenomenon unique to LF welded seams. It is a "variation in properties of the weld occurring at short, regular intervals."3
Stitching arises from the pipe being welded too rapidly so that 60-cycle power fluctuations were being translated into nonuniform heating of the seam. A stitched region may or may not be interspersed with cold welds, but it certainly results in less-than-optimum bond line toughness.
Note the absence of stitching on the continuous disbonded region of the DC-welded seam shown in Fig. 6c. The "chevrons" in Fig. 6c denote brittle fracture and highlight the unbonded defect which caused this hydrostatic test rupture.
Cold welds indicate that either the welding heat or the pressure or both were not conducive to proper welding. They may also indicate the presence of dirt or contamination on the edges of the skelp.
The problem was much more likely to occur with LF welding than with HF welding because the former was a much more sensitive to contact resistance. In fact, when HF welding was introduced, it was found to be unnecessary to pickle or sandblast the edges of the skelp to achieve good contact.5
Furthermore, in the period before widespread use of HF welding, the ability of mill-inspection techniques to detect cold welds was much poorer than it is today. In particular, neither the ring flattening nor the weld tensile tests were reliable means of detecting poorly bonded seams.
Cold welds are the worst kind of defect associated with ERW seams. When a poorly bonded seam exists, it is virtually certain that what material remains intact is of very low toughness.
A series of full-scale tests of one particular lot of LF ERW pipe showed that the fracture toughness of the bond line region equalled 2 ft-lb of absorbed energy in a Charpy V-notch impact specimen (full size).6 In comparison, one often finds that the skelp of an X-52 line-pipe material will exhibit a toughness equivalent to 15 ft-lb or more.
As will be shown subsequently, poorly bonded ERW materials are unable to tolerate very large defects. Fortunately, the improvements in recent years in the manufacturing of ERW pipe have led to seams with toughnesses more like that of the parent skelp.7
MARTENSITE; HOOK CRACKS
A rare condition associated with some older ERW materials or unnormalized HF ERW materials is the existence of untempered martensite in the weld zone as the result of inadequate post-weld heat treatment.
This condition may cause a problem wherein, even if the material contains no defects initially, it is susceptible to developing hydrogen cracking in service if, for example, it is exposed to hydrogen charging in the soil environment from cathodic protection. An example of hydrogen cracking in a martensitic weld zone is shown in Fig. 7.
API Bulletin 5t1 describes hook cracks as "metal separations resulting from imperfections at the edge of the plate or skelp, parallel to the surface which turn toward the ID or OD pipe surface when the edges are upset during welding."
Examples of hook cracks are shown in Figs. 8 and 9.
Fig. 8, a metallographic section, shows the presence of a large hook crack near the ID surface of the pipe material. Note the crack is not in the bond line but curves outward from the plane of the plate in the upset material near the bond line.
In Fig. 9, the horizontal lines on the fracture surfaces correspond to hook cracks. Note the contrast between this "woody" appearance and the vertical orientation of stitches and bond line imperfections in the photograph of Fig. 6b or the smooth weld of Fig. 6c.
Whether hook cracks occur in association with LF or HF seams, they are the same because their source is the skelp not the welding process. They arise from "dirty" (high sulfur) steels because of the tendency of such materials to contain elongated nonmetallic inclusions.
The inclusions are formed when the ingot is cast and are elongated parallel to the surfaces of the skelp during hot rolling of the strip. As long as they remain in that orientation, they have little effect on the integrity of the skelp (although they tend adversely to affect the ductile-fracture resistance of the material).
When the edges of the skelp are upset during welding, however, these nonmetallic layers become oriented perpendicularly to the maximum principal stress direction. In this orientation, they reduce the pressure-carrying capacity of the pipe.
Hook cracks can be a severe problem in the retesting of pipelines comprised of older ERW materials. They often cause hydrostatic-test failures. And frequently, the test itself enlarges those which do not fail.
Hook cracks can be practically eliminated from any new pipe materials by use of ultra low-sulfur steels or sulfide shape-control practices wherein those sulfides which do form are not rolled out into elongated layers.
SELECTIVE CORROSION
Selective corrosion can occur in service in some ERW and FW seam materials. In particular, the high-sulfur materials seem to be more of a problems.8 The problem is often worse if the material is improperly normalized.
Selective corrosion of an ERW seam may occur in conjunction with corrosion of the body of the pipe. The difference is that because certain ERW weld zones are more susceptible to corrosion than the surrounding skelp, the bond line of weld and some of the HAZ are preferentially corroded at a higher rate than the surrounding material.
The result is often the formation of a V-shaped groove centered on the bond line as shown in Fig. 10. The selective seam corrosion occasionally affects a very narrow region right at the seam, as shown by the knife-like linear slot in Fig. 11.
Selective corrosion represents a potentially severe defect. It tends to form a relatively sharp notch in a material which is usually much weaker than the parent skelp.
Pipe with such defects should be removed if the grooving is anything but superficial. In particular, no attempt should be made to evaluate the remaining strength of the pipe utilizing the ASME B31G criterion.
ACKNOWLEDGMENT
The author is grateful for suggestions and technical advice provided by Ted Bruno, Metallurgical Consultants Inc., Ted Clark, Columbia Gas Transmission Corp., and Jim Cox, Colonial Pipeline Co.
REFERENCES
- "Electric Resistance Weld Pipe Failures on Hazardous Liquid and Gas Transmission Pipelines," Technical Report OPS 89-1, Office of Pipeline Safety, Research and Special Programs Administration, U,S. Department of Transportation, August 1989.
- Fields, R. J., Pugh, I. N., Read, D. T., and Smith, J. H., "An Assessment of the Performance and Reliability of Older ERW Pipelines," U.S. Department of Commerce, National Institute of Standards and Technology, July 1989.
- API Bulletin on Imperfection Terminology (5Tl), 9th Ed. May 31, 1988.
- Minamiya, S., Watarabe, S., Sugie, Y., and Shibagat, M., "The Automatic Heat Control System for the Production of High Grade Materials," Mechanical Working and Steel Processing XXI, AIME.
- Heald, S. T., "Radio-Frequency Resistance Welding of Carbon Steel Pipe and Tubing," Mechanical Working of Steel 1, AIME, 1963.
- Beavers, J, A., Davis, C. O., Gundaker, W. F., Kiefner, J. F., Morin, C. R., and Shoemaker, A. K., "An Operational Reliability Analysis of the Williams Pipe Line Company No. 2 8-Inch Diameter Pipeline," submitted to the Department of Transportation Research and Special Programs Administration, August 1988.
- Jones, D, G., and Raine, G. A., "Inspection and Performance of Modem ERW Pipe," the Welding Institute Third International Conference on Welding Performance of Pipelines, Nov. 18-20, 1986.
- Kato, C., Otogura, Y., Kado, S,, and Hisamatsu, Y., "Grooving Corrosion in Electric Resistance Welded Steel Pipe in Sea Water," Corrosion Science, Vol. 18 (1978), pp. 61-74.
Copyright 1992 Oil & Gas Journal. All Rights Reserved.