Corrosion Failure Analysis of a Tee in a Natural Gas Field
3. College of Power and Energy Engineering, Harbin Engineering University, Harbin 150001, China
Incident involving failure occurred for a 40 months old upright type tee of gas filed gathering and transmission pipeline in a west natural gas field. The tee consisted of a main pipe of 16Mn steel and branch pipes of bimetal composite of 316L stainless steel+L416 steel. Examinations were performed to identify the cause of the failure, which include visual physical inspection, pipe thickness measurement and acquiring the galvanic corrosion behavior of 16Mn steel and 316L stainless steel in an oxygen-free medium containing chloride ion concentration 150 g/L with CO2 partial pressure 0.1 MPa at 65 ℃ using electrochemical methods, as well as Computational Fluid Dynamics (CFD) analysis of the tee with assumed angles as 90°, 45°and 30°respectively between the branch pipe and main pipe along the direction of inlet using the FLUENT software. Investigations revealed that for the present tee design, there existed galvanic couple of the metallic materials and the straight cutting structure of the tee caused the fluid changes radically. Gas field water gradually adhered to the inner side of the pipe due to the low velocity that caused by the vortex which was formed nearby the pipe explosion site. Corrosion then ensued as CO2 dissolved in the water. The area around the explosion site suffered from large wall shear stress caused by the fluid scouring action. Because of the synergy of galvanic corrosion and fluid flow erosion the pipe gradually thinned, and then perforated.
There are abundant natural gas resources in the western part of China. In the course of natural gas exploitation, CO2 and other gas and formation water, thousands of meters underground, are being mined together with natural gas . The gas and formation water are collected by the gas gathering pipeline of the gas field, and then transported to the treatment station for gas and water separation to obtain high purity natural gas. The tee components of gas transmission pipeline become weak parts of pipeline due to the welding of dissimilar materials and high temperature and high pressure gas convergence. In the gas transmission pipeline, the high speed flowing medium will scour the pipe. The change of the flow velocity of the medium will change the flow state of the medium, thus it has an important influence on the corrosion of the pipeline.
The water content of natural gas produced in a gas field in the western region is about 7.4%, the CO2 partial pressure is about 0.1 MPa, the concentration of chloride ion in the formation water is about 150 g/L, and the medium produced is strong corrosivity. In order to prevent corrosion and consider the economic benefits comprehensively, the tee branch pipe is made of 316L stainless steel and 16Mn steel is used as the main material. The tee structure is directly inserted, that is, the branch pipe and the main pipe are connected at 90 degree angles along the entrance direction. The inside diameter is 422 mm, the wall thickness is 26 mm, and the inner diameter of the branch pipe is 144.3 mm. The design flow rate of the branch inlet is 5.3 m/s, the medium temperature is 95 C, the designed inlet flow rate is 5.6 m/s, and the medium temperature is 61 C. The flow rate of the transmission medium in the pipe is 6.14 m/s and the temperature is 65 degrees. However, in the process of production, all tee channels of the collection pipeline have obvious local thinning phenomenon, and the pipeline thinning speed of some positions is about 8 mm/a, and even after 40 months of use, the pipe burst accident occurred.
Fig.1 Three views of the tee joint
It is not clear whether such a rapid local thinning is due to corrosion or fluid erosion. Therefore, in this study, the reasons for the failure of the tees are analyzed mainly through the observation of the morphology of the tee pipe fittings, the wall thickness measurement, the simulated electrochemical experiment and the computer simulation flow.
2 experimental method
The experimental material is a tee failure of a gas field in Western China. The material is 16Mn steel, and its chemical composition (mass fraction,%) is C 0.144, Si 0.313, Mn 1.548, P 0.014, S 0.012, Ti 0.002, Nb 0.002, Al 0.028, V 0.005, and Fe. The material of the branch is 316L stainless steel, and its chemical composition (mass fraction,%) is C 0.020, Si 0.504, Mn 1.059, P 0.036, S < 0.001, Ni 11.90, Cr 17.19, Mo 2.093, and Fe.
Macroscopical inspection of failure tee links is carried out and macroscopic appearance is observed. A square sample of 30 mm x 30 mm was intercepted from a failed tee conductor by means of electron microscope. A spiral micrometer is used to measure the wall thickness of the failure tee head, and the three views of the invalid tee pipe fittings, as shown in Figure 1, measure the line segment AA ‘, BB’, CC ‘, DD’ and EE ‘at the starting point of A, B, C, D, E (close to the weld of the tee angle), and analyze the thinning law of the tee pipe.
The 16Mn steel and 316L stainless steel were processed into electrodes with a size of 10 mm x 10 mm (effective working area 1 cm2), which were encapsulated with epoxy resin and then grinded to 2000# by water paper. Galvanic corrosion tests were carried out using CST500 electrochemical noise and galvanic corrosion detector. The experiment adopts tee electrode system, two working electrodes are 16Mn steel and 316L stainless steel (WE1 and WE2) respectively, and the reference electrode is Ag/AgCl electrode. The galvanic corrosion current density of two working electrode faces 0, 10, 20 and 30 cm respectively, the data acquisition frequency is 0.1 Hz, and the experimental time is 36 h. The corrosion medium is used to simulate the formation water of a gas field in Western China, and 247.18 g/L NaCl is used as electrolyte solution. Nitrogen was used to deoxygenate 0.5 h before the experiment. The electrode system and solution temperature were controlled by constant temperature water bath, and maintained at 65 C and 0.1 MPa CO2.
The internal flow field of tee links is simulated with FLUENT software. According to the characteristics of the tee pipe fittings, taking the tee pipe as the center and taking the pipeline of 5 m in the upper and lower reaches as the research object, the physical model of the straight plug type tee is established by using the Gambit. According to table 1, the boundary conditions are set up and the grid is divided into [2,3]. 356231 grid volume units are set up, as shown in Figure 2A. In the calculation process, the right pipe mouth and the branch port are defined as the entrance, and the left port of the main pipe is the outlet. Taking the intersection of the supervisor and the axis of the branch as the Descartes coordinate origin; taking the axis of the supervisor as the X axis, the downstream of the director as the negative direction of the X axis, the upstream of the X axis, the axis of the branch of the branch as the Y axis, the direction of the fluid flow in the branch pipe as the positive direction of the Y axis, and the plane of the axis of charge and the axis of the branch pipe as the XOY plane, that is, the Z plane. On the basis of the direct tee physical model, around the Descartes coordinate Z axis, the branch pipe is rotated counter clockwise 45 and 60 degrees respectively, and the tee physical model of the branch tube is 45 and 30 degrees in the direction of the inlet. According to table 1, the boundary conditions are set up and meshed respectively, as shown in figures 2B and C. The velocity cloud images, turbulence intensity maps, temperature cloud images and wall shear force nephogram of tee section specific section fluid with different structures are calculated by iteration.
Fig.2 Grids of three tee joints with different structures: (a) upright type tee, (b) 45° lateral tee, (c) 30° lateral tee
Fig.3 Field inspection of tube explosion (a), the tee joint with failure (b), inner wall occurred thickness reduction of the tee joint (c) and inner wall directly faced by the mouth of branch pipe (d)
3 experimental results
3.1 field observation results
The scene of the 3.1.1 accident scene is shown as shown in Figure 3a. The bursting position is located downstream of the head and along the tee fillet weld (branch to head connection) at 8 o’clock direction (Fig. 3b). The downstream head of the tee fillet weld was thinning down from 7 to 11 o’clock, and as the distance increased, the thinning area was expanded in a hyperbolic type (Figure 3C); the tube was not thinned, and the direct pair with the branch pipe was thinner but not serious (Figure 3D).
The morphology of the 3.1.2 failure tee is observed by the naked eye. The pipeline corrosion in the hyperbolic thinning area is serious, and with the characteristics of the fluid scour, the topography of the local area is flaky, and the edge of the region is scoured out by the fluid (Figure 4a). Obvious pitting is observed in the inner wall of the upstream pipeline, which is symmetrical with the hyperbolic thinning area (Fig. 4b), which is a common corrosion pattern of carbon steel in Cl- medium.
Fig.4 Macro morphologies of the failure tee: the hyperbolic-type thinned area (a), inner wall of the tee upstream (b)
Fig.5 Micro morphologies of the failure tee:inner wall of the tee upstream (a), the hyperbolic-type thinned area with gully (b), wash zone (c) and pitting (d)
Fig.6 Thickness (a) and thinning rate (b) of the main pipe corresponding to the position of AA′~EE′ in Fig.1
As shown in Figure 5a, the micromorphology of the inner wall of the upstream pipe of the hyperbolic thinning zone can be found, and the pitting pit of larger size can be found, and the pitting pit is fused with the continuous expansion of the pitting corrosion. The pipeline in the hyperbolic thinning area presents a ravine morphology (Fig. 5b) and shows a river like scour on the surface (Fig. 5C). In addition, there are pitting pits in the fluid erosion area (Fig. 5C, d).
The thickness of the thinnest part of the pipe wall (blasting site) is only 1.68 mm, and the distance from the tee fillet welds is 100 mm, and the distance from the branch axis is about 175 mm. 3.1.3 failure tee pipe supervisor’s thinning rule. The inner wall hyperbolic thinning area is about 400 mm in length, and no obvious thinning in other locations. The wall thickness of the line AA 1, BB ‘, CC’, DD ‘, EE’ corresponding position is shown in Figure 6A. According to the design of the wall thickness (26 mm) of the tee head, the thinning speed of the wall thickness at the corresponding position of each line section in Figure 1, as shown in Figure 6B, has been calculated for 40 months after the tee connection is put into use. As can be seen from Fig. 6, with the location of the measuring point far away from the tee fillet welds, the thinning amount of the head decreases and the thinning speed decreases.
Fig.7 Open circuit potential of 16Mn steel and 316L stainless steel
Simulation results of 3.2 galvanic corrosion
The open potential of 16Mn steel and 316L stainless steel is shown in Figure 7. The potential of 16Mn steel is negative, and the potential of 316L stainless steel is more positive. The electric couple will be formed when the two is connected, and galvanic corrosion occurs. Among them, 16Mn steel is anode and 316L stainless steel is cathode. According to Faraday’s law, the relationship between corrosion current density and corrosion rate is as follows: :
|Distance of galvanic couple / cm||Current density μAcm–2||Corrosion rate mma–1|
(1) CR=K1IEw/ p
In the formula, the value of K1 is 3.27 * 10-3, mmg/ (Acma); I is the current density, – A/cm; P = density fand g/cm.
According to formula (1), the galvanic corrosion current density of different galvanic couples is converted to corrosion rate, and the results are shown in Table 2. From table 2, it is known that the galvanic corrosion rate of 16Mn steel and 316L stainless steel decreases with the increase of the opposite distance. Correspondingly, the thinning rate and the thinning rate of 16Mn steel decrease with the increase of the distance between the couple faces.
Fig.8 Contours of wall shear stress of the upright type tee joint
Fig.9 Contours of velocity magnitude on sectionsof Z=0 (a), X=-80 mm (b), X=-200 mm (c), X=-400 mm (d) of the upright type tee
3.3 flow simulation results
Figure 8 is a shear force cloud of a straight insert tee inner fluid to the wall of a pipe. It is seen from the figure that the fluid has a larger shear force on the wall of the downstream head located in the direction of the tee – to – corner weld from 7 to 11 o’clock, and the region is roughly hyperbolic extension. Among them, the shear force along the tee fillet weld is the largest in the 8 o’clock and 10 o’clock directions, that is, the pipe wall in the downstream direction.
Figure 9A is a velocity cloud map of a straight tee Z=0 mm cross section fluid. It can be seen that the low velocity eddy current zone is formed at the lower reaches of the pipe near the branch when the branch is confluence with the two strands of high speed fluid in the pipe. The velocity cloud of the tee X=-80 mm section (Figure 9b), X=-200 mm section (Figure 9C) and X=-400 mm section (Fig. 9D) fluid is known as the center of the eddy region at the 6 o’clock direction (9 o’clock direction of the tee welds), containing 5 to the head (tee welds 8 to the direction), and the low speed eddy current area is about 5 00 mm. With the increase of the absolute value of X, the eddy current affected area of the downstream mother tube increases first and then decreases. The tee X=-200 mm cross section is the central area of the eddy.
The turbulent intensity cloud map of the tee Z=0 mm cross section (Fig. 10a) shows that the turbulence intensity of the downstream main wall near the branch is larger. The turbulence intensity distributions of the tee X=-80mm section (Figure 10b), the X=-200 mm section (Figure 10c) and the X=-400 mm section (Figure 10d) fluid show that the turbulent intensity distribution of the fluid is associated with the velocity distribution, the turbulent intensity of the fluid in the center of the eddy current is the largest, and the turbulence has a significant influence on the wall surface of the 5 to 7 o’clock direction, which is in shadow with the eddy current. The ringing area is consistent.
Fig. 11a is a temperature cloud map of the straight through tee Z=0 mm section fluid. It can be seen that there is a high temperature influence area near the downstream pipe of the branch pipe. In the tee X=-80 mm cross section (Figure 11b), X=-200 mm cross section (11c) and X=-400 mm cross section (Figure 11d) fluid temperature cloud map, it is found that the high temperature influence zone is related to the eddy current zone, and the temperature gradient of the fluid located in the direction of the supervisor from 5 to 7 o’clock is intense.
Fig.10 Contours of turbulent kinetic energy on sectionsctions of Z=0 (a), X=-80 mm (b), X=-200 mm (c), X=-400 mm (d) of the upright type tee
If the angle of entry of the tee is changed to 45 degrees, the wall shear force in the tee links will have great changes compared with the direct insertion tee (Fig. 12a). The shear force of the fluid on the pipe wall is significantly reduced, and the area with strong shear effect is obviously reduced. If the remittance angle continues to be reduced to 30 degrees, the area of the fluid with stronger shear effect on the pipe will continue to decrease and the shear action will continue to weaken (Figure 12b).
Fig.11 Contours of static temperature on sections of Z=0 (a), X=-80 mm (b), X=-200 mm (c), X=-400 mm (d) of the upright type tee
Fig.12 Contours of wall shear stress of the 45° lateral tee (a) and 30° lateral tee (b)
As shown in Figure 13 of the velocity of fluid on different sections of the 45 degree tee, it can be seen that the low velocity eddy current intensity of the wall fluid in the lower reaches of the branch is remarkably weakened after the convergence of the two strands of high speed fluid in the tee, and the influence range of the eddy current area is obviously reduced. As shown in FIG. 14, we can see that the low-speed eddy current phenomenon of the 30 speed tee fluid flow at different sections of the tee section is disappearing when the two high-speed fluids in the tee links converge.
Fig.13 Contours of velocity magnitude on sections of Z=0 (a), X=-80 mm (b), X=-200 mm (c), X=-400 mm (d) of 45° lateral tee
Figures 15 and 16 are the turbulent intensity clouds of different sections of 45 and 30 degree tees. It is shown that the turbulent intensity of the fluid in the pipe is basically consistent, and there is no greater turbulence intensity in the downstream wall of the pipe near the branch.
The failure tee material is 16Mn steel and the material of the branch pipe is 316L stainless steel. Because the self corrosion potential of 16Mn steel and 316L stainless steel is not equal, there is a risk [6,7] of galvanic corrosion in the design of tee materials. The simulation results of the straight tee flow state show that after the branch is confluence with the two strands of high speed fluid in the head, the flow velocity of the downstream supervisor is about 500 mm in the direction from 5 to 7 o’clock, and the flow velocity decreases and the temperature is reduced. First, due to the emergence of turbulence and the decrease of flow velocity, the probability of collision between water vapor particles increases suddenly and converges to integrate larger water droplets and condenses. Secondly, relative to the branch pipe, the temperature decreases after the gas enters the gathering pipeline, and the saturated vapor pressure of water vapor in the natural gas is also reduced, resulting in the water vapor condensing more easily. Finally, the turbulence makes the acid solution stay in the region for a long time, and has long enough action with the wall of the area. Under the joint action of these factors, the condensation of water vapor adheres to the inner wall of the pipe, and the corrosive gases such as CO2 melt into the condensate, thus forming a typical acid corrosion environment, which provides the necessary conditions for the galvanic corrosion of 16Mn steel and 316L stainless steel.
Fig.14 Contours of velocity magnitude on sections of Z=0 (a), X=-80 mm (b), X=-200 mm (c), X=-400 mm (d) of 30° lateral tee
Fig.15 Contours of turbulent kinetic energy on sec-tions of Z=0 (a), X=-80 mm (b), X=-200 mm (c), X=-400 mm (d) of 45° lateral tee
In addition, compared with the environment on the other side of the pipeline, the temperature in the low speed eddy current region is relatively high, and the higher temperature is combined with the acid medium, which is more likely to accelerate the acid corrosion. It is observed that the thinning area of the head is about 400 mm, and the thinning area is located in the direction from 7 to 11 o’clock in the tee weld, which is from 5 to 7 o’clock in the downstream, which is in good agreement with the simulation results. The simulation test results of the galvanic corrosion in the laboratory also show that the thinning rate of 16Mn steel decreases with the increase of the positive opposite distance of the galvanic, which is the same as the thinning law of the field observed. The results show that galvanic corrosion is one of the reasons that can not be neglected in the reduction of tee links.
Fig.16 Contours of turbulent kinetic energy on sections of Z=0 (a), X=-80 mm (b), X=-200 mm (c), X=-400 mm (d) of 30° lateral tee
But comparing the results of field observation and galvanic corrosion test, it is found that there is a big discrepancy between the two. The failure rate of the tee supervisor is greater than that of the 16Mn steel, as shown in Figure 17, which indicates that there is another reason for the failure of the tee. According to the wall shear force cloud diagram of the tee inner fluid, we can see (Figure 8) that the branch pipe and the two high speed fluid in the pipe together have a strong shear effect on the wall of the hyperbolic area after the tee connection. Under the continuous cutting of high speed fluid, the head in the hyperbolic region will be thinned by the mechanical scour, and the hyperbolic area is in good agreement with the thinning area of the failure tee head observed in the field. The microscopic morphology of the pipeline in the hyperbolic region also indicates the existence of mechanical scour. Corrosion leads to the depression and roughness of the surface of the tube, which leads to a greater change in the flow pattern. After the change of the flow, the water vapor is easier to condense, the wall shear force of the air flow leads to the corrosion products to be blown away, which makes the corrosion products unable to prevent corrosion, and the fresh metal is exposed again and causes corrosion to accelerate. By comparing the upper and lower inner wall surfaces of the tee heads, it is found that the gully morphology of the pipes in the hyperbolic region is formed by a long time erosion of the local corrosion pits. In addition, the acid solution will gather at the 7 o’clock direction of the downstream supervisor in the direction of the downstream head because of the gravity action, that is, the tee fillet weld is at 8 o’clock direction. This leads to the most serious erosion and corrosion of the wall surface of the tee X=-200 mm section. The results of the field observation strongly confirm this analysis. The position of the tee tube detonator is located in the lower reaches of the supervisor, along the 8 o’clock direction of the tee weld, that is, the head is in the 7 o’clock direction and the axis of the branch is 175 mm. The synergistic effect between galvanic corrosion and fluid erosion is the fundamental cause of the tee failure.
Fig.17 Thinning rate of the main pipe corresponding to theposition of CC′ in Fig.1 and corrosion rate of 16Mn steel
Comparing the flow simulation results of the tee through, 45 degree tee and 30 degree tee connections, it can be seen that the straight tee structure leads to the violent changes in the flow state after the junction of the branch pipe and the two high speed fluid in the head, and the low velocity eddy current region is formed on the wall of the downstream head near the branch, and the turbulence intensity and the pipe of the region are in the opposite direction. The wall shear force of the road is large, which leads to galvanic corrosion and mechanical erosion of the tee links. With the gradual decrease of the angle between the branch pipe and the inlet direction, the low velocity eddy current disappears, the shear force of the fluid to the wall of the pipe gradually decreases, and the synergistic effect of the tee galvanic corrosion scour can be effectively controlled.
5 conclusions and suggestions
(1) there is a galvanic in the failure tee material design, and the direct plug structure leads to the violent change of the flow state after the joint of the branch pipe and the two high speed fluid in the head, which provides the condition for the galvanic corrosion and the mechanical scour of the tee pipe.
(2) the thinning area of the failed tee supervisor is the low-speed eddy current area in the pipeline. The associated gas CO2 and water adhere to the region, forming an acid corrosion solution, which causes the galvanic corrosion of the 16Mn steel and the 316L stainless steel, and eventually leads to the thinning of the 16Mn steel.
(3) the shear force and fluid turbulence intensity of the fluid in the pipe wall are the largest. The inner wall of the pipeline is strongly scoured. In addition, the acid solution formed by the eddy current is aggregated by the gravitational force at the detonator, which makes the acid corrosion of the pipe in the region the most serious, and eventually leads to the maximum thinning of the pipe in the tube.
(4) under the synergistic effect of galvanic corrosion and fluid erosion, the final 7 o’clock direction of the downstream supervisor, namely the 8 o’clock direction of the tee fillet weld, and the tube accident at 175 mm distance from the axis of the branch pipe.
(5) in order to eliminate galvanic corrosion, it is necessary to avoid the use of dissimilar metal or coating on the inner wall of the tee to avoid direct contact between the electrolyte solution and the metal when the tee and the head are selected. At the same time, the structure design of the tee can be optimized by reducing the angle between the branch pipe and the direction of the inlet, thus avoiding the formation of the low speed eddy current in the pipe and effectively weakening the scouring effect of the fluid on the pipe.
Source: Network Arrangement – China Tee Manufacturer – Yaang Pipe Industry (www.metallicsteel.com)
(Yaang Pipe Industry is a leading manufacturer and supplier of nickel alloy and stainless steel products, including Super Duplex Stainless Steel Flanges, Stainless Steel Flanges, Stainless Steel Pipe Fittings, Stainless Steel Pipe. Yaang products are widely used in Shipbuilding, Nuclear power, Marine engineering, Petroleum, Chemical, Mining, Sewage treatment, Natural gas and Pressure vessels and other industries.)
If you want to have more information about the article or you want to share your opinion with us, contact us at firstname.lastname@example.org
Please notice that you might be interested in the other technical articles we’ve published:
• WHERE TO BUY HIGH QUALITY STEEL PIPES
• WHERE TO BUY HIGH QUALITY FLANGES
• WHERE TO BUY HIGH QUALITY PIPE FITTINGS
• ADVANTAGES OF DUPLEX STAINLESS STEEL PIPE AND SELECTION
• WHERE TO BUY HIGH QUALITY BELLOW EXPANSION JOINTS
• Difference Between Pipe Elbow And Pipe Bend
• Where to get high quality alloy steel pipes
• Distinguish Inferior Steel Pipes
• WHERE TO GET HIGH QUALITY HEAT EXCHANGER TUBES
• The Heat Treatment Process of Steel Pipes
• WHERE TO GET HIGH QUALITY STUB ENDS
• WHERE TO GET HIGH QUALITY CARBON STEEL PIPES
• Exports are blocked and there is excess capacity, where is the road for china’s stainless steel industry
• After a hundred years, why is 304 still the absolute mainstay of the global stainless steel market
• Analysis and Countermeasures of corrosion failure of 316L stainless steel flange
• Stress Corrosion Cracking Behavior of Nuclear Grade 316LN Stainless Steel Bend in High Temperature and Pressure Water
 yuan Geng. Oil and gas pipeline corrosion prediction and [D]. Dalian: Dalian University of Technology, 2011
 Yu Yong. FLUENT introductory and advanced tutorial [M]. Beijing: Beijing Institute of Technology press, 2008: 24
 Majid Z A, Mohsin R, Yaacob Z. Failure analysis of, Failure 2010, 17 (4): 818
 Yang Lietai. Corrosion monitoring technology [M]. Beijing: Chemical Industry Press, 2012: 51
 ASTM G102, Annual Book of ASTM Standards -Standard Practice for Calculation
 Li Xiaogang. Corrosion and protection of materials [M]. Changsha: Central South University press, 2009: 69
 Wang Zhaohui, Al, Geng hiroteru. Anticorrosion design of galvanic corrosion and protection effect can not be ignored, [J]. 2001, 8 (1): 33