|
| |
|
Return to
Failure Analysis Case Histories
|
Incomplete Weld in a Carbon Steel Gas Main Pipe
|
|
ENVIRONMENT:
|
Underground |
|
EQUIPMENT:
|
Gas Main Pipe |
|
MATERIAL: |
Carbon Steel |
|
SERVICE TIME:
|
approximately fifty years |
|
FAILURE:
|
Incomplete Weld |
|
|
|
Summary
|
|
A section of 8-inch diameter
steel lap-welded gas pipeline, that leaked, was evaluated. No gross defects or
fractures were found, however the plane of the lap weld contained numerous areas
of lack-of-fusion defects as well as excessive amounts of non-metallic
inclusions. These lack-of fusion defects were believed to be interconnected and
to be the primary cause of the gas leakage. Impact fracture tests made on this
pipe indicated that the pipe will fracture in a brittle fashion at room
temperature or below.
|
| |
Background
|
|
A routine leak detection survey
discovered a natural gas leak in an underground pipeline. After excavation of the leak area, the leak site was pinpointed
using soap and water bubble indication. A three-foot section of the pipe
containing the leak was cut out for evaluation, as were three other nearby
sections, for comparison. The bitumastic coating was removed from the pipe
sections in the field. The four pipe sections were brought to Corrosion Testing
Laboratories for analysis and determination of the cause of the leak.
|
| |
Description of Pipe
|
|
The as-received pipe section samples were numbered 1
through 4. Each pipe section was about 3 feet long. The location of the leak
was marked on pipe section #2. Pipe section #’s 2 and 3 were removed from the
same pipe joint. Pipe section #’s 1 and 4 were removed from, respectively, the
pipe joints located to the west and to the east of the leaking joint. This
piping was steel lap welded and 8-inches in diameter. Most of the bitumastic
coating had been removed from sample #2. Less of the coating had been removed
from the other samples. The exterior surfaces of the pipe did not appear to be
significantly corroded.
|
|
The steel pipe was purchased as API 5L and was lap welded
and was believed to have been in the ground for nearly fifty years. It was
originally coated, and in recent years had an impressed current cathodic
protection system. The operating internal gas pressure was 380 psi. |
|
|
|
Findings |
Visual & Macro-Examination
|
|
The pipe sections are similar in surface
appearance. The pipes are superficially rusty on the outer surface and have
considerable coating remaining on the surface. The interior surfaces of the
pipe sections are rusty, but not excessively so. The lap welded pipe and the
location of the lap weld is indicated by a line of cross hatching marks made by
the pressure rolls applied during the lap welding process. There are
intermittent longitudinal grooves along each edge of the lap weld about 1 inch
apart, see Figure 1. |
|
|
 |
|
Figure 1. Close-up
of leak area on pipe section #2. Note
also cross-hatching marks from pressure rolls during welding. |
| |
|
The area on pipe sample #2
indicated to be the leak zone, Figure 1, was scrutinized under the
binocular microscope at magnifications up to 40X but no visible crack or metal
discontinuity was observed in the leak zone. This area was further subjected to
dye penetrant testing, but no cracks or discontinuities were found using this
technique. The indicated leak zone of pipe
sample #2 was cleaned using ultrasound and citric acid and when it was reexamined
under the binocular microscope, no cracks or discontinuities were observed.
|
Metallography
|
|
A section from pipe #2 containing the leak area was cut
out and macro etched to reveal the location and profile of the lap weld. Then
this piece was mounted and polished for microscopic examination. The unetched
weld cross section was observed, Figure 2. Several areas of
lack-of-complete-weld-fusion were visible along the lap weld plane. The entire
lap weld cross section was seen. Approximately 20% of the lap weld was not
fused. Figure 3 showed the same area after etching to reveal the metallographic
structure. In addition to the voids along the weld plane resulting from lack of
weld fusion, there was an almost continuous line of heavy sulfide inclusions all
along the weld fusion line. |
|
|
 |
 |
| Figure 2. Pipe
section #2. Unetched transverse cross section through weld showing
lack of fusion and sulfide inclusions. (25X Original Magnification) |
Figure 3. Pipe
section #2: Transverse cross-section through weld. Microstructure
revealed by etching with 2% nital. Note string of gray sulfide inclusions
along fusion line. (100X Original Magnification) |
| |
|
Metallographic structure of the same
pipe sample from the side opposite the weld. The unwelded side of the pipe
shows a typical ferritic microstructure with patches of pearlite. This is the
normal microstructure of annealed low carbon steel. Heating to the welding
temperature along the weld lap allows some decarburization of the pipe material
in the weld zone, and thus there is less pearlite present in the microstructure
and the ferritic grain size is somewhat larger.
|
|
Microscopic examination of the
as-polished longitudinal section of the pipe permits an assessment of the
quantity of non-metallic inclusions present in the steel. This is fairly dirty
steel, with lots of manganese sulfides and alumina inclusions and some oxides.
However it is probably typical for the period in which this steel was
manufactured.
|
Chemical Analysis
|
|
Samples from the pipe sections were chemically
analyzed and were determined to be in compliance with API 5L requirements.
|
Mechanical Properties
|
|
Longitudinal samples were cut
from pipe sections #1, #2 and #4 at the lap weld line and opposite the weld
line. The purpose of testing both the weld zone and the non-welded zone was to
define any property differences in the two areas. Standard tensile tests were
made in accordance with ASTM A370, from which the ultimate tensile strength, the
0.2% yield strength, the reduction in area, and elongation in 2 inches were
measured.
|
|
API 5L standard for pipe has
two levels of strength, grades A and B. This pipe barely meets the requirements
of the lower strength A grade, and in fact the #2 pipe, which leaked, has a
yield strength slightly below the specified minimum. Otherwise the mechanical
properties look normal for this grade of pipe.
|
Notched Impact Tests
|
|
Samples were taken from the lap
weld area and from the side opposite the lap weld. These slotted samples were
fractured by impact at varying test temperatures. The fracture appearance of
the broken pieces was recorded in terms of the percent of the fracture area
exhibiting ductile fracture characteristics.
|
|
The test results show that all three pipe samples exhibit
predominately brittle fracture characteristics under impact loading at room
temperature and below. Above room temperature the fractures become increasingly
ductile and at 150F are predominantly ductile. Figure 4 illustrates the
separation of the lap weld in the impact sample from the lap weld area of pipe. |
|
|
|
 |
|
Figure 4. Impact
fracture of pipe section #2 at the lap weld line. Note weld
separation starting at OD. surface and extending part of the way
through the pipe wall. (4X Original Magnification) |
| |
Bend Test
|
|
An extra impact specimen prepared from pipe section #2
was left un-notched and subjected to a bend test in which it was slowly bent
through a 45° angle. Examination of
the bent specimen showed nearly complete separation of the edges of the plate in
the weld zone, indicating lack of fusion in the weld. Figure 5 shows the bend
test specimen after bending. |
|
|
|
 |
|
Figure 5. Bend
specimen taken from lap weld on pipe section #2. Bending of specimen
caused separation along pipe edges in weld zone (arrows), indicating lack
of fusion. Pipe OD. is at bottom. Left arrow points to transverse
cross-sectional face. (2X Original Magnification) |
| |
Discussion
|
|
Visual and macroscopic
examination failed to reveal any gross defect or fracture in pipe section #2
that might account for the leakage. It was also noted that corrosion, either
external or internal, did not appear to be a factor.
|
|
The metallography of the lap
weld cross-sections showed that the weld fusion in the lap weld zone was
incomplete, and further that the lap weld plane was infused with an excess of
non-metallic inclusions. This condition, while not leading to fracture of the
pipe, probably provided minute passages for gas leakage.
|
|
Continuous lap welded pipe is a
process which involves rolling the edges of the pipe scalp to a taper, then
rounding up the cylinder, superheating the thinned edges to the welding
temperature and applying pressure to fuse the tapered edges together. It is a
form of forge welding no longer used in pipe production.
|
|
