STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

Pipeline Corrosion
Corrosion Basics: Metals are normally found in nature in one their lowest energy states - usually as oxides, sulfides, chlorides, etc. In reducing and refining metals to produce useful alloys (such as the carbon and low alloy steels used in gas and oil transmission pipelines), significant amounts of energy are consumed and stored within the reduced metallic structures. Subsequent corrosion of steel pipelines thus represents the natural tendency of the iron in the pipe to return to a preferred, lower energy state (usually as an oxide, carbonate or sulfide). Corrosion of steel - at the relatively low temperatures (less than 200 degrees F.) normally encountered in pipeline operations takes place by an electrochemical process. This process, in turn, requires the presence of anodic and cathodic areas on the surface of the pipe and the presence of a suitable, conductive aqueous environment that contacts both the anodic and cathodic areas. For buried pipe, the external corrosion environment will usually consist of moist, relatively high conductivity soil. Internal corrosion can occur if water exists within the line and is allowed to accumulate at low spots in the line. Significant internal corrosion also usually requires the presence of a significant partial pressure of carbon dioxide and/or oxygen within the line. The consumption of the steel pipe occurs at the anodic areas on its surface by oxidation of the iron of the pipe wall. The anodic portion of the corrosion process can thus be represented by equation (1): Fe Fe++ + 2e(1)

The cathodic portion of the electrochemical corrosion process may reportedly occur by one of several reactions, depending upon the conditions of the environment: O2 + 2H2O + 4e- 4OHO2 + 4+ + 4e- 2H2O 2H2O + 2e- 2OH- + H2 H+ + e- _ H2 (2) (3) (4) (5)

The ultimate fate of the Fe++ ion from equation (1) also depends upon the environment. The Fe may stay in solution as the ion or it may be precipitated as Fe(OH)2 or as FeCO3. For external corrosion in moist soils, the ultimate corrosion product is usually Fe(OH)2, while internal corrosion involving carbon dioxide often results in FeCO3 as a corrosion product. Page 1 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

The kinetics of the electrochemical process can be shown schematically using the diagram in Figure 1. The open circuit potentials of the local cathodes and anodes, c and A , are shown on the diagram, along with the polarization paths for the cathodes and anodes that result as increasing amounts of current are produced by the local electrodes. The over-all (average) corrosion potential for a surface covered with small, adjacent local anodes and cathodes (in a solution with moderate to high conductivity) thus occurs where the polarization curves for the electrodes approach one another, as shown in Figure 1. A similar diagram, as shown in Figure 2, can be used to illustrate the basic characteristics of corrosion prevention using cathodic polarization. The diagram in Figure 2 shows the continuation of the cathodic polarization curve that occurs as increasing amounts of positive current are forced onto the initially freely corroding sample surface (line c e f). Consider the situation at point e. At this point, the total current being supplied to the surface, Ie , consists of the sum of the current being supplied from local anodes, Ib , and the current being supplied from an outside voltage source, Ie Ib . As the cathodic polarization of the sample surface is increased to point f, all current from the local anodes has been shut off and all of the current flowing to the sample surface is coming from the external applied voltage source. It should be noted that at point e, the sample surface is experiencing only partial protection from the applied current, I e Ib , that is being forced onto its surface. Some corrosion (as indicated by the anodic current, Ib ) is still occurring on the sample surface. The sample becomes fully protected only after the polarized potential of the sample has dropped to A and the anodic contribution to the total sample current has dropped to zero. It should also be noted, however, that continued polarization of the sample surface, to potentials more negative than A, has no additional beneficial effects in preventing corrosion and may, instead, cause difficulties due to hydrogen-induced disbonding of coatings and hydrogen induced cracking of the steel of the pipeline.

Corrosion Prevention: External corrosion The principal methods used to prevent external corrosion of pipelines are coatings and cathodic protection (CP) of the lines. In recent installations, coatings and CP have normally been used together in a complimentary fashion, since high quality coatings substantially reduce the CP current requirements and the application of a functioning CP system allows some relaxation in the requirement for 100% holiday (defect) free coatings.

Page 2 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

Coatings: The NACE Standard RP0169-96 [1] lists most of the desirable characteristics of a pipeline coating. These include the following: 1. The coating should have a high electrical resistance and high dielectric strength. 2. The coating should be an effective moisture barrier. 3. The coating should be reasonably easy to apply and the application process should not change the properties of the pipe. 4. The coating should exhibit good adhesion to the pipe. 5. The coating should be resistant to chemical and physical damage/degradation during installation and service. 6. The coating should be reasonably easy to repair in the field. 7. The use of the coating should not present any environmental or health risks. Pipeline coatings have been used for more than 70 years and numerous systems have been developed. The coating systems that are currently being applied include the following: 1. Coal tar enamels containing embedded glass fiber mats. 2. Mill-applied tape systems. 3. Extruded polyethylene and polypropylene coatings. 4. Fusion bonded epoxy (FBE) coatings. 5. Multi-layer, FBE under extruded polyethylene or polypropylene. The last three coating systems listed above are reportedly currently experiencing increasing acceptance by consumers and their future use should therefore expand. Cathodic Protection: The electrochemical basis for cathodic protection systems was presented briefly above in the Corrosion Basics section (see Figure 2). The current used to cathodically polarize the sample to be protected can typically come from an impressed current system using an external, D.C. power supply that supplies current to the pipe by way of a remote anode ground bed. Alternatively, the protective current can come from a reactive, galvanic anode or group of

Page 3 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

anodes. Galvanic anodes are typically located within ten to twenty feet of the spot on the pipe to be protected. A schematic representation of a typical impressed current, cathodic protection system is shown in Figure 3. The anodes in the ground bed are usually made of graphite or high alloy cast iron rods. The rectifier that serves as the source of the polarizing current may have a voltage range of 10 to 100 volts and an available D.C. current range of 5 to 200 amperes. Since positive current flows from the positive to the negative terminal of the power supply in an external circuit that is connected to a D.C. power supply, it is critical that the pipeline to be protected be connected to the negative terminal of the rectifier. Connection of the pipeline to the positive terminal of the rectifier would result in greatly accelerated corrosion of the line (instead of the planned reduction/elimination of corrosion). The kinetics of the cathodic protection process when using a sacrificial, ganvanic anode are illustrated in Figure 4. The sacrificial or galvanic anodes are typically fabricated of relatively pure zinc or magnesium or alloys of these reactive metals. The polarized potential of typical zinc anodes is approximately 1.1 volts (as measured using a saturated copper copper sulfate reference electrode - CSE). The polarized potential of a typical magnesium alloy anode is, on the other hand, approximately 1.50 to 1.55 volts vs. a CSE. The available driving potentials from the sacrificial anodes for polarizing steel structures are, therefore, relatively limited, and the length of pipe that can be protected using sacrificial anodes is relatively small. The three primary inspection criteria currently used to assess if appropriate levels of cathodic protection (CP) are being supplied to protected piping by a CP system are also described in NACE Standard RP0169-96. These criteria are: 1. A piping potential of 850 mV vs. a CSE, measured with the CP system in operation. 2. A polarized piping potential of 850 mV vs. a CSE, as measured within approximately 1/2 to 1 second after (simultaneously) turning off all sources of direct current to the piping. 3. 100 mV of polarization with respect to the native corrosion potential of the pipe. The polarized potential used in this evaluation criterion is the same instant off polarization used in criterion 2. In using criterion #1, it is recognized by NACE that there are IR drop errors in the potential measurements that must somehow be estimated and evaluated in applying this criterion. There are no firm guidelines presented, however, on how this estimation and evaluation should be performed. In making the measurements involved in criteria #2 and #3, the IR drop errors caused by the flow of D.C. current to the protected structure are eliminated by measuring the polarized potential of the structure or piping within a half to one second after simultaneously shutting off all D.C.

Page 4 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

current sources to the structure or piping. There may, of course, be considerable difficulty and expense in finding and arranging for the simultaneous interruption of all D.C. currents to the piping and failure to eliminate these sources of current will result in errors in the measurements. In using criterion #3, the measurement or estimation of the native corrosion potential of the existing pipe or structure may also present some difficulties. For new piping that has not been previously protected by a CP system, it is only necessary to measure the initial native corrosion potential and then turn the CP system on and wait for the potential of the piping to drop to a stable value. At this point, switching off the source of all D.C. currents allows the measurement of the instant off polarized potential of the pipe and the shift in potential with respect to the original native potential. For existing piping that is currently under the protection of a CP system, shutting off all D.C. currents will allow the measurement of the instant off polarized potential of the pipe. A considerable waiting period (and some significant opportunity for error) may, on the other hand, be encountered in obtaining an estimate of the native corrosion potential in this case. Unfortunately, previously used piping systems that have been under the influence of a CP system for some extensive period are typically the objects of a CP system evaluation. Care must be taken during the installation and/or adjustment of CP systems to insure that the applied CP voltage is neither too low nor too high. Applied voltages that are too low could, of course, result in some corrosion to the piping. Also, there is some evidence that the high pH, stress corrosion cracking that is sometimes seen on the external surfaces of pipelines occurs in the range of lower polarized potentials (from approximately 0.50 and 0.85 volts vs. a CSE). In addition, elevated temperatures in the pipe are known to promote corrosion of the pipe. For piping or piping areas that operate at temperatures significantly above the surrounding earth temperature, an operating CP potential of 0.95 volts or more should be considered. The presence of bacteria in the soil may also promote the presence of microbiologically induced corrosion (MIC) on the outside surface of pipelines. In areas where MIC is suspected or confirmed, a CP potential of -0.95 volts or more should be considered. On the other hand, CP voltages that are larger than approximately 1.05 to 1.10 volts are thought to cause hydrogen induced cracking of some pipeline steels (particularly older steels containing higher levels of sulfur and phosphorus). This hydrogen induced cracking appears to be greatest in hard spots produced in the pipe during manufacture and in the heat affected zone of welds where small, localized hard areas may be present. Finally, elevated CP voltages may cause hydrogen-induced damage of coatings. It is generally recommended that CP voltages larger than approximately 1.10 volts be avoided in order to minimize the possibility of coating damage due to evolution of hydrogen.

Page 5 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

Internal Corrosion Internal corrosion in a pipeline requires the presence of liquid water within the line. In gas transmission lines (the only pipelines that will be discussed in this document), internal corrosion also usually signals the presence of significant partial pressures of carbon dioxide and/or hydrogen sulfide in the line. It is also known, however, that on a weight percentage or weight fraction basis, dissolved oxygen is more corrosive to ordinary steels than either carbon dioxide or hydrogen sulfide. Although the probability of having appreciable concentrations of oxygen inside a gas transmission line is apparently quite low, it should be remembered that even small partial pressures of oxygen can produce surprisingly high internal corrosion rates in steel pipes that also contain liquid water. One method to reduce the danger of internal corrosion by the acid gases, carbon dioxide and hydrogen sulfide, is to reduce the concentration of the acid gases in the gas transmission stream by a process known as gas sweetening. Many gas sweetening processes have been developed and used. These include, for example: 1. Solid bed absorption (using iron sponge, mole sieves or zinc oxide), 2. Chemical solvents (such as mono ethanol amine, di ethanol amine, potassium carbonate, etc), 3. Proprietary physical solvents, 4. Conversion of hydrogen sulfide to sulfur, 5. Distillation. Water can form in a pipeline if there has been no attempt to dehydrate the gas prior to its introduction into the line or if the gas dehydration process that was used did not produce water contents in the gas that were low enough to prevent condensation of liquid water in the line. If the gas temperature drops below its water dew point, liquid water will probably form. Liquid water that is produced in the line will, of course, tend to accumulate in the low points in the line. Here, the water will equilibrate with carbon dioxide and/or hydrogen sulfide in the gas and can produce local areas of high internal corrosion rates. A second effective method used to prevent internal corrosion of gas transmission pipelines is thus dehydration of the gas prior to its introduction into the line. The aim of the dehydration process is to reduce the water content of the gas to a low enough level that the water will not condense in the line under the lowest pressure and temperature that the gas will experience in the line. By far the most common dehydration process for natural gas involves contacting the gas with a hygroscopic liquid such as a glycol. The most common glycol used for gas dehydration is

Page 6 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

triethylene glycol. The dehydration process takes place in a multi-tray column known as a glycol contactor. The glycol is regenerated before recycling to the contactor by heating to drive out the absorbed water. Glycol dehydration can usually easily reduce the dew point of the gas to the level required to prevent water condensation during transmission. The use of gas sweetening in conjunction with gas dehydration will, of course, minimize the chance of problems with internal corrosion in gas pipelines. Early work by de Waard and coworkers at Shell [2,3,4,5] resulted in what has come to be known as the Shell model for predicting the corrosion rates of steel by carbon dioxide. For example, the Nomogram for CO2 Corrosion, shown in Figure 5, allows easy estimation of the predicted corrosion rate of steel at various temperatures and carbon dioxide partial pressures. The combined effects of temperature and carbon dioxide partial pressure on the anticipated corrosion rates are shown in Figure 6. It should be pointed out that the Shell model is generally felt to be moderately to substantially conservative. For example, the model was developed for clean systems (containing no oil or other liquid hydrocarbons) and the presence of condensed hydrocarbons may substantially reduce the observed corrosion rates. In contrast to the weight loss corrosion problems produced by carbon dioxide, hydrogen sulfide (at the relatively low temperatures encountered in gas pipeline operations) generally causes environmental cracking (sulfide stress cracking, SCC) problems rather than weight loss corrosion. Guidelines for the selection of candidate materials for use in hydrogen sulfide environments (sour environments) are given in NACE Standard MR0175-2000 [6]. The concentrations of hydrogen sulfide above which the gas stream should be considered sour (and the threshold concentrations which will thus probably cause SCC) are also defined in NACE MR0175-2000 (see Figure 7 below).

Corrosion Monitoring: External Corrosion Survey methods that are commonly used to evaluate the external corrosion conditions of pipelines include: 1. Pipe-to-soil potential measurements, 2. Soil resistivity measurements, 3. Measurements of D.C. currents flowing along the pipeline, 4. Bellhole examinations of the pipe.

Page 7 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

Pipe-to-soil potential measurements are typically made using a saturated copper-copper sulfate (CSE) reference electrode that is placed in contact with the soil directly over the line. The potential measurements are made with a high input impedance voltmeter. Hooking the negative terminal of the voltmeter to the CSE electrode and the positive terminal to the pipeline gives readings with the normally used sign convention (e.g., the native corrosion potential of bare steel in moist soil will normally read between 0.1 and 0.5 volts). In pipe-to-soil potential surveys of pipe that is not under cathodic protection (and that has been allowed to reach its native corrosion potential prior to starting the measurements), the points on the line with the largest negative potential values will normally be the areas with the highest corrosion rates. Newly installed pipe (and pipe sections) will, however, usually have pipe-to-soil potentials that are substantially more negative than older sections of line and the pipe-to-soil potentials of new pipelines (without CP) will usually tend to decrease in magnitude (become less negative) with the passage of time. In applying pipe-to-soil potential measurements to pipelines under CP, one of the three primary acceptance criteria given in NACE Standard RP-01-69 (and discussed above) can be used. An example of actual pipe-to-soil potential measurements taken from the literature [7] is given in Figure 8. As shown by the upper curve in Figure 8, the section of pipeline represented in the figure would have satisfied criterion # 1 (-0.850 V vs. CSE with the CP system on). The pipeline would not, however, have satisfied criterion # 2 (a -0.850 V, instant off polarized potential). This criterion is represented by the intermediate curve in Figure 8. As can be seen, the section of the pipe between 0 and approximately 150 meters in the plot had a polarized potential that was smaller (less negative) than the required 0.850 V vs. CSE. By subtracting the bottom curve (the native corrosion potential curve) from the intermediate curve (the instant off polarized potential curve), it can be seen that most of the pipeline also failed to meet criterion # 3. The calculated differences between the intermediate curve and the lower curve in the figure appear to generally be smaller than the 100 mV required by criterion # 3. Soil resistivity measurements can be made using either two terminal or four terminal meters. Either an A.C. or D.C. power supply can be used in conjunction with an instrument that accurately measures the current and potential between the test electrodes. Four terminal instruments are usually used when larger soil areas are examined or when resistivities at a greater depths are desired. Corrosion rates of buried pipes are generally higher in lower resistivity (higher conductivity) soils. Guidelines correlating observed corrosion rates with soil resistivities have been developed. These guidelines are documented in Table 1. Because of the possibility of errors caused by voltage drops in the soil due to the flow of CP currents, it is recommended that soil resistivity measurements be made with CP systems shut off. Line current measurements are typically made using test stations that are installed at the time the pipe was laid. Electrical leads are connected to both ends of the pipe test span and these leads are subsequently used to measure the voltage drop across the test span. The electrical resistance

Page 8 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

of the test span is then either estimated or measured and the net electrical current in the test span is calculated using Ohms law. The sign of the voltage drop indicates the direction of the current flow through the test span. In order to eliminate the effects of any active CP system, line current measurements should be made with those systems shut off. The currents detected in line current measurements are long-line currents that are typically caused by widely separated macro electrodes (e.g., different soil conditions along the line) or by interferences from foreign D.C. fields in the earth (such as those caused by an adjacent, unconnected CP system). Long-line currents are not caused by the local anodes and cathodes that produce the corrosion normally observed on the line. However, at the location(s) where long line currents leave the pipe, the resulting corrosion rates can be very high. For example calculation shows that, if only 10 milliamps of D.C. current leaves a pipe over an area of 1 square inch on the pipe surface, a corrosion penetration rate of approximately 700 mils (or about 0.7 inches) per year would be observed at that location.

Internal Corrosion Successful monitoring of internal corrosion of pipelines is apparently significantly more difficult than monitoring of external corrosion, as discussed above. One method that may yield valuable information concerning the general internal condition of a line is to periodically run scraper pigs through the lines. Evaluating the quantity and composition of material that is removed from the line by the scraper pig may be useful in evaluating whether or not significant internal corrosion has been occurring in the line. The development and use of smart pigs may soon allow the successful simultaneous detection and monitoring of both external and internal corrosion/damage in pipelines. Measurement techniques that have been considered and/or used in previous smart pig development efforts include: 1. Multi-finger, mechanical calipers that detect and record the effective internal radius of the pipe, 2. Magnetic flux-leakage tools that may be configured to respond to both longitudinal and circumferential defects in the pipe. These tools may also include high frequency eddy current sensors that can differentiate between internal and external damage, 3. Ultrasonic tools that couple directly to the pipe wall through a surrounding liquid and that may measure either the internal radius or the wall thickness of the pipe, 4. Ultrasonic tools that use electromagnetic acoustic transducers (EMATS) to evaluate the condition of the pipe wall. These transducers use electromagnetic signals to generate

Page 9 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

ultrasonic signals in the pipe wall. Future use of EMAT technology may eliminate many of the difficulties and short comings with direct coupling ultrasonic tools.

Corrosion Economics: A recent review [8] of the economic effects of corrosion upon the U. S. economy has been published. The results of this review indicate that corrosion of metals and alloys costs U. S. companies (and consumers) a total of approximately $300 billion per year. The authors of this review (scientists at Battelle Institute and the National Institute of Standards and Technology) also concluded that approximately one third of these total costs (approximately $100 billion per year) could be significantly reduced or eliminated by the use of current best available corrosion prevention techniques and materials. In the review, it was estimated that the pipeline industry accounted for something less than 1 percent of the total industry-wide corrosion costs. This would thus probably put the total costs for corrosion in the pipeline industry somewhere in the range of $2 billion to $3 billion per year. It also thus seems possible that the use of improved materials and corrosion prevention techniques in the pipeline industry might reduce the total costs of corrosion in this industry by as much as $600 million to $900 million (by ~ 30%). In the case of the pipeline industry, as in several other industry segments, the authors of the review felt that, although the need for corrosion-related repairs and re-coating had apparently gone down in the recent past, the savings due to the drop in repairs had been essentially balanced by the use of more expensive original materials of construction. In our opinion, the development of more sensitive and more accurate inspection techniques (such as improved smart pigs) and the possible regulatory requirement for the use of these more sensitive inspection techniques could substantially increase the repair costs associated with the future operation of aging gas transmission pipelines.

Page 10 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

References 1. NACE RP0169-96 Control of External Corrosion on Underground or Submerged Metallic Piping Systems. 2. C. de Waard and D.E. Milliams, Carbonic Acid Cirrosion of Steel, Corrosion, Vol. 31, 1975. 3. C. de Waard, U. Lotz and D.E. Milliams, Predictive Model For CO2 Corrosion Engineering in Wet Natural Gas Pipelines, Corrosion, Vol. 47, 1991. 4. C. de Waard and U. Lotz Prediction of CO2 Corrosion of Carbon Steel, Corrosion 93, Paper 69, 1993. 5. C. de Waard, U. Lotz and A. Dugstad Influence of Liquid Flow Velocity on CO2 Corrosion, Corrosion 95, 1995. 6. NACE MR0175-2000 Sulfide Stress Cracking Resistant Materials for Oilfield Equipment. 7. Peabodys Control of Pipeline Corrosion, NACE, 2001. 8. Economic Effects of Metallic Corrosion in the United States: a 1995 Update, Battelle Institute, 1996.

Page 11 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

TABLE I

Corrosion of Steel in Soil

Area

Corrosion (mpy)

Corrosion Severity

Resistivity (- cm)

Ave. of Several Soils Tidal Marsh Clay

61

Moderately Corrosive Corrosive Very Corrosive Mildly Corrosive Noncorrosive

1000 to 2000

100 137

500 to 1000 Less than 500

Sandy Loam

21

2000 to 10000

Desert Sand

Above

10000

Page 12 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

Page 13 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

Page 14 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

Page 15 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

Page 16 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

Page 17 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

Page 18 of 19

STRESS ENGINEERING SERVICES, INC.

Houston Cincinnati New Orleans Denver Atlanta Chicago

Page 19 of 19