J Fail. Anal. and Preven. (2011) 11:227–232 DOI 10.1007/s11668-011-9447-y
CASE HISTORY—PEER-REVIEWED
Pipeline Failure Results from Lightning Strike: Act of Mother Nature? G. T. Quickel • J. A. Beavers
Submitted: 6 January 2011 / Published online: 24 March 2011 ASM International 2011
Abstract There are more than 2.5 million miles of oil and gas pipelines in the United States. Approximately 900 failures occurred on hazardous liquid pipelines from 2002 to 2003, and 9% of these failures were attributed to damages due to natural force, which included lightning strikes, among other naturally occurring events. This paper provides a case history in which failure analysis was applied to determine the metallurgical cause of a failure involving a polyethylene-coated hydrocarbon pipeline that leaked as a result of a lightning strike.
location of failure was 2760 kPa, which corresponds to 20% of SMYS. The pipeline was installed in 1977 and was externally coated with extruded polyethylene. The pipeline has an impressed current cathodic protection (CP) system. The objectives of the analysis were to determine the metallurgical cause of the leak and identify the contributing factors, if any.
Pipeline Lightning Failure Polyethylene
The following steps were performed for the analysis. The pipe section was visually inspected and photographed. A portion of the pipe that contained the feature at the leak location was removed from the pipe section. Magnetic particle inspection (MPI) was performed on the external pipe surface at the feature. The feature was cleaned in a solvent, examined with a stereo light microscope at low magnifications, and examined in a scanning electron microscope (SEM) at high magnifications. The portion that contained the feature was sectioned, mounted in epoxy, polished, and etched. Light photomicrographs were taken to document the feature morphology and steel microstructure. Hardness testing was performed on the mount to document the hardness. Chemical analysis was performed on a sample removed from the base metal of the pipe section. Mechanical testing (duplicate tensiles and full Charpy curve) was performed on samples removed from the base and seam weld of the pipe section.
Keywords Coating
Background A metallurgical analysis was performed on a section of a pipeline that leaked while transporting hydrocarbons. The portion of the pipeline containing the failure was composed of API 5L X46 line pipe steel, 21.91-cm diameter by 0.478cm wall thickness, which contained an electric resistance welded (ERW) longitudinal seam. The maximum operating pressure (MOP) was 9370 kPa, which corresponds to 68% of the specified minimum yield strength (SMYS). The normal operating pressure was 5170 kPa, which corresponds to 37% of SMYS. The pressure at the time and
Approach
G. T. Quickel (&) J. A. Beavers DNV Columbus, Inc., Dublin, OH, USA e-mail:
[email protected]
Results
J. A. Beavers e-mail:
[email protected]
This section of the paper discusses the results of the metallographic analysis that was performed.
123
228
Optical Examination Figure 1 is a photograph of the pipe section, as received. A seam weld was present 2.3 in. clockwise (looking downstream (D/S)) of the feature described elsewhere, at the 2:00 O’clock orientation. Coating was present on approximately 75% of the pipe surface. The remainder had been removed from the pipe section by pipeline personnel. The coating adjacent to the feature was adhered and intact. A Plidco sleeve was located over the feature and removed by pipeline personnel before arrival. The feature was through-wall, located at the 1:00 orientation. There was no evidence of significant external corrosion on the pipe section. Figures 2 and 3 are photographs and stereo light photomicrographs, respectively, of the external surface of the pipe section. The figures show the feature, which had a semi-hemispherical shape. The through-wall portion of the feature is indicated in both figures. The feature on the external surface was approximately 1.1 cm in axial length by 1.2 cm in circumferential length. It was 0.40 cm (83% of wall thickness) in depth (measured using a Thorpe pipe pit depth gage) with a small through-wall region in the center. A majority of the external surfaces of the feature appeared to be smooth and had a shiny appearance. Figure 4 is a photograph of the internal pipe surface at the feature after performing the MPI of the external pipe surface. The through-wall region is visible. Also evident is the fluid from the MPI that leached through the perforation. There was no evidence of significant internal corrosion. The circumference and wall thickness of the pipe section were measured. The circumference of the pipe section was 69.5 cm (calculated diameter of 22.1 cm), which is consistent with a diameter of 21.91 cm for nominal 20.32-cmdiameter pipe. The wall thickness was measured at the U/S end and D/S end; see Table 1. The average wall-thickness value was 0.486 cm. The wall-thickness values of the line pipe steel are consistent with a nominal wall-thickness of 0.478 cm.
J Fail. Anal. and Preven. (2011) 11:227–232
image of the external surface of the feature. Dendrites are visible in the feature. There was no evidence of machining marks on the feature surfaces. Metallographic Examination Figure 7 is a stereo light photomicrograph of the mounted cross section (Mount A) that was removed from the feature. The through-wall location is indicated in the figure. The
Fig. 2 Photograph of the external pipe surface at the feature, as-received; area indicated in Fig. 1
Scanning Electron Microscopy Figure 5 is an SEM image of the external surface of the feature. The figure shows that the surface of the feature is relatively smooth compared with the surrounding external pipe surface. Figure 6 is a higher magnification SEM
Fig. 3 Stereo light photomicrograph of the external pipe surface at the feature
Fig. 1 Photograph of the pipe section, as received. The leak location is in the black box
123
J Fail. Anal. and Preven. (2011) 11:227–232
229
Fig. 4 Photograph of the internal pipe surface at the feature Table 1 Results of wall-thickness measurements performed at the U/S and D/S ends of the pipe section
Fig. 6 Close-up SEM image of the external pipe surface at the feature; area indicated in Fig. 5
Wall thickness, cm O’clock orientations
U/S end
D/S end
12:00
0.470
0.478
3:00
0.467
0.518
6:00
0.465
0.551
9:00 Average
0.467 0.467
0.472 0.505
Fig. 7 Stereo light photomicrograph of the mounted cross section (Mount A) that was removed from the feature (4% Nital Etchant); hardness location D is indicated in the figure
Fig. 5 SEM image of the external pipe surface at the feature
microstructure of the steel is obviously different at the feature relative to away from it. Figure 8 is a light photomicrograph near the outer diameter (OD) surface. Region 3 is the unaffected base metal. Region 2 is similar to a heat affected zone (HAZ) of a poor weld with improper cooling. The HAZ is similar to that one may see if an arc weld was performed on the external surface of a pipe; although, the width of the HAZ would likely be wider in arc welding. Region 1 has an unusual appearance for line pipe steels,
Fig. 8 Light photomicrograph showing the mirror image of the area indicated in Fig. 7 (4% Nital Etchant); hardness locations A, B, and C in Mount A are indicated in the figure
123
230
so that it did not etch with Nital. Figure 9 is a light photomicrograph of the interface between Regions 1 and 2. The whiter region (Region 1 on the right) is likely eutectic iron, based on its location, appearance, and hardness; see later. Eutectic iron contains a high carbon content and, in low carbon steels, results from diffusion of carbon from an external carbon source into the metal at temperatures above 1130 C. The likely source of the carbon was the extruded polyethylene and/or hydrocarbon product. Also present in the white region are cracks, consistent with the low ductility of eutectic iron. Region 2 consists of coarse and finegrained HAZs. The microstructure in Region 2 near the white region has features consistent with martensite or bainite. The presence of martensite and/or bainite indicates heating of the metal followed by rapid cooling. Figure 10 is a light photomicrograph of the typical unaffected base metal microstructure. The microstructure
J Fail. Anal. and Preven. (2011) 11:227–232
consisted of pearlite (dark areas consisting of lamellae) and ferrite (white areas) and is typical for this vintage of line pipe steel. Hardness Testing Hardness testing was performed on Mount A at the locations shown in Figs. 7 and 8 (at hardness locations A–D); see Table 2 for results of the testing. The hardness in the white portion of Mount A (Region 1) was *800 HK, which is consistent with the high hardness of eutectic iron. Region 1 must have been carburized from the coating or hydrocarbon product. The hardness in Region 2 (hardness locations B and C) ranged from 433 to 463 HK (1370–1480 MPa equivalent ultimate tensile strength (UTS)). The hardness is consistent with martensite for the carbon content of the pipe section. The hardness of the base metal is consistent with the UTS for this grade of line pipe steel. Mechanical Testing The results of the tensile testing for the pipe section are shown in Table 3. The tensile properties for the base metal were determined to be 405 MPa (yield strength (YS)) and 530 MPa (UTS), respectively. The base metal tensile properties meet the minimum YS and UTS specifications for API 5L Grade X46 line pipe steels of 317 and 434 MPa, respectively. Table 2 Results of hardness testing performed on the mounted cross section
Fig. 9 Light photomicrograph of the microstructure of the white and dark regions shown in Fig. 8 (4% Nital Etchant)
Hardness locations
Hardness Mount Aa (Knoop)
Hardness Mount Aa (Rockwell)
Equivalent UTS, MPa
A (eutectic iron)
846
65 HRC
[2420
B (coarse HAZ)
463
45 HRC
1480
C (fine HAZ)
433
43 HRC
1370
D (base metal)
195
89 HRB
607
See Figs. 7 and 8 for hardness locations A, B, C, and D a
Average of three measurements
Table 3 Results of tensile testing performed on transverse samples compared with tensile specifications for API 5L X46 line pipe steel Base metal Across seam API 5L X46 samples weld samples (minimum values)a
Fig. 10 Light photomicrograph of the microstructure of unaffected base metal (4% Nital Etchant)
123
Yield strength, MPa
405
…
317
Tensile strength, MPa
530
578
434
Elongation in 5 cm, %
27
…
23.2
Reduction in area, %
47
…
…
a
API Spec 5LX, March 1975, 20th edition
J Fail. Anal. and Preven. (2011) 11:227–232
231
The UTS of the samples removed from across the seam weld was determined to be 578 MPa. The UTS value across the weld exceeded the minimum specified value for API 5L Grade X46 line pipe steel of 434 MPa. Charpy V-notch (CVN) impact testing was performed on 15 base and 10 seam weld samples. An analysis of the base metal data indicates that the 85% fracture appearance transition temperature (FATT) is 44 C and the upper shelf Charpy energy is 47 N m, full size. An analysis of the seam weld data indicates that the 85% FATT is 1.2 C, and the upper shelf Charpy energy is 28 N m, full size. There is no API 5L standard for this vintage of line pipe steel; these are considered as good values for this vintage and grade of line pipe steel. The CVN test results can be adjusted to account for material constraint effects by applying temperature shifts to the data [1]. The modified transition temperature (brittleto-ductile fracture initiation temperature) for the base metal was estimated as 41 C based on a pipe wall thickness of 0.478 cm; see Table 4. The modified transition temperature for the seam weld was estimated as 2.7 C for this pipe section, based on a pipe wall thickness of 0.478 cm. Based on this analysis, the tested materials are expected to exhibit ductile-fracture behavior above their modified transition temperatures. Chemical Analysis The results of the chemical analysis performed on a sample removed from the pipe section are shown in Table 5. The results of the analysis are consistent with composition
Table 4 Results of analysis of the Charpy V-notch impact energy and percent shear plots for samples removed from the base metal and across the seam weld Base metal
specifications for API 5L X46 line pipe steel at the time of manufacture.
Discussion and Conclusions The results of the analysis indicate that the cause of the perforation was a high-voltage arc-discharging energy from the pipe wall to ground and attributable to energy sources such as lightening strikes or ground faults from high-tension power lines. Supporting evidence for this conclusion includes the morphology of the through-wall feature, which is consistent with resolidification of molten metal and the presence of hardened microstructures near the feature, including eutectic iron. Various methods can be employed to protect belowground pipelines from high-voltage lightning damage. Three devices/methods for protecting insulated joints include 1) standard lightning arresters, 2) grounding cells with zinc anodes, and 3) isolation surge protectors (ISP)s [2]. The appropriate voltage rating needs to be determined for standard lightning arresters to provide sufficient protection. Grounding cells can be used for protection by combining voltage surge protection with CP on the unprotected side of an insulated joint. ISPs are electronic mechanisms that block DC current flow and allow AC current to flow. ISP’s are more expensive than grounding cells, but can provide exceptional protection. For facilities and other above ground structures, the probability of a direct lightning strike can be reduced with the installation of lightning rods [3]. Lightning rods can be designed to either attract or discourage direct strikes by the shape of the rod tip. The following is a summary of our observations and conclusions: •
Seam weld samples
• Upper shelf impact energy, N m
47
28
85% FATT, C
44
1.2
Maxey adjusted 85% FATT, C
41
2.7
Table 5 Results of chemical analysis performed on a sample removed from the base metal compared to composition specifications for API 5L X46 line pipe steel Composition, wt.%
API 5L X46, wt.%a
C (carbon)
0.163
0.28 (max)
Mn (manganese)
0.920
1.25 (max)
P (phosphorus)
0.015
0.04 (max)
S (sulfur)
0.008
0.05 (max)
Element
a
API Spec 5LX, March 1975, 20th edition, welded, cold expanded
•
•
•
A through-wall hemispherical feature was identified on the external surface of the pipe section. The feature on the external surface was smooth, had a shiny appearance, and contained dendritic grains, which are consistent with resolidified molten metal. Narrow heat-affected zones (HAZs) were present adjacent to the feature surfaces. The width of the zone indicates localized heating. Eutectic iron was likely present at the feature, which would suggest that an external source of carbon was present during the local heating of the pipe at the perforation, resulting in local carburization. The hardness of the region is consistent with a carburized region. The likely source of the carbon was the coating or hydrocarbons. The hardness and microstructure of the heat-affected zone (HAZ) are consistent with martensite and/or bainite. The presence of martensite/bainite indicates heating of the metal followed by rapid cooling.
123
232
•
• •
•
J Fail. Anal. and Preven. (2011) 11:227–232
The hardness and microstructure of the base metal for away from the perforation are consistent with this vintage and grade of line pipe steel. There was no evidence of external or internal corrosion. The pipe steel chemistry and tensile properties meet the composition and tensile specifications for API 5L X46 line pipe steel at the time of manufacture. The microstructure of the base metal was consistent with the vintage and grade of line pipe steel.
123
References 1. Rosenfeld, M.J.: A simple procedure for synthesizing Charpy impact energy transition curves from limited test data. In: International Pipeline Conference ASME, vol. 1, p. 216 (1996) 2. Peabody, A.W.: Peabody’s Control of Pipeline Corrosion, 2nd edn., pp. 252–254. NACE International (2001) 3. Kaiser, B.A.: Lightning protection for pipeline compressor stations and other facilities. Pipeline Gas J. 232(11), 64 (2005)