EVALUATING CORRODED PIPE-1 METHOD FOR EVALUATING CORRODED PIPE ADDRESSES VARIETY OF PATTERNS

Thomas J. O'Grady II
ARCO Alaska Inc.

Daniel T. Hisey
ARCO Transportation Alaska Inc.
Anchorage

John F. Kiefner
Kiefner & Associates Inc.
Columbus, Ohio

ARCO Alaska Inc. has developed a method for evaluating corroded pipelines which includes and expands upon criteria embodied in the ASME/ANSI B31G "Guide for the Evaluation of the Remaining Strength of Corroded Pipe."

This method permits evaluation of large areas of metal loss, corrosion in some welds, groups of noncontinuous pits and, to a limited extent, the effect of circumferential corrosion.

The B31G approach, on the other hand, deals with the effect of single arrays of pitting on the pressure-carrying capacity of the pipe.

The method includes the "Modified B31G Criterion" developed by Battelle Columbus for the AGA Pipeline Research Committee. An objective of Battelle's work was to address the built-in conservatism of B31G.

This first part of two articles sets forth the criteria and procedures for evaluating metal loss; the conclusion presents a method for determining the safe operating pressure for corroded pipe.

INSPECTION

Inspection of corroded pipe involves four major concerns: safety, surface preparation, critical dimensions, and equipment.

If the extent of the corrosion is completely unknown, one must assume that it could be severe enough to cause failure at any time. With certain exceptions, the pressure level in the pipe should be reduced before inspection, the extent depending on the pipe's service and its last known condition.

Low-pressure liquid lines require no change. A 25% lowering, however, is recommended for high-pressure liquids and low-pressure gas. A 50% lowering is recommended for high-pressure gas service.

Exceptions to these guidelines apply if the nature and extent of the corrosion are already roughly known and appear insufficient to cause failure or if the current pressure (P) is not the highest the pipe has experienced within the past year and the previous high within that year was at least 125% of P.

If aggressive corrosion is present, the period since the last high pressure should be shorter than 1 year.

Although great care should be taken when heavy equipment works near any pipeline, care is especially necessary around a corroded area. Moving the pipeline while it remains in service, for example, could cause it to fail even if it is lifted carefully enough not to damage an uncorroded pipeline.

It is prudent to establish the general nature and severity of any metal loss before any attempt is made to move the pipeline.

All insulation, coating material, and corrosion products in the area of metal loss must be removed. A final light sand blast which just removes the residual nonmetallic material is recommended.

Exceptions and qualifications notwithstanding, the critical dimensions of metal loss, whether external or internal, are shown in Fig. 1.

The axial or longitudinal extent (L) and the maximum depth (d) must be determined for an evaluation of the remaining pressure-carrying capacity of a corroded pipe.

A is the area of the defect in the longitudinal plane through the wall, and t is the wall thickness. Much of what follows will discuss how to determine these dimensions, especially when the extent of corrosion is less obvious than that shown in Fig. 1.

Needed to inspect the pipe and to characterize the metal loss are an ultrasonic thickness tester, straight edge, measuring tape, wire (contour) gauge, pit gauge, and small ruler or scale.

CLASSIFICATION

Classifying corrosion according to several criteria is necessary to make its measurement and evaluation easier. The classifying parameters and considerations for evaluating the effects of metal loss include the following:

External vs. internal corrosion
Localized vs. general corrosion
Interaction of closely spaced areas
Circumferential extent
Bends, fittings, welds.

Although the effect of metal loss is the same whether on the inside or outside of the pipe, opinions vary on how to determine the extent and severity of the corrosion and on what type of repair is to be used.

With external corrosion, measurement with simple tools (that is, rulers, pit gauges, etc.) is possible. Once the pipe's surface has been cleaned, the extent of external corrosion will be obvious.

Determining the extent of internal corrosion, however, involves mapping the area with an ultrasonic device.

Another difference between evaluating external vs. internal corrosion involves consideration of a "corrosion allowance" for internal corrosion.

When an area of external corrosion is cleaned, evaluated, and recoated, it may be assumed that the corrosion will not continue. For internal corrosion, although mitigating steps may be taken, it is generally safe to assume that the corrosion may continue.

In such cases, a specified amount (the corrosion allowance) may be added to the measured depth of the corrosion to render the evaluation valid for some time. The amount added should be consistent with the corrosion rate and the time anticipated before the next inspection.

Localized pitting consists of clearly defined, relatively isolated regions of metal loss. The axial and circumferential extents of such regions are easily measurable as are the depths of the pits.

Often a pit gauge will suffice to measure depth because the pipe's original surface can usually be used as a plane of reference.

In contrast, widespread or general corrosion can make measurement difficult. Sometimes, small islands of original pipe surface can serve as reference points.

Two such points on the same axial line can be linked with a straight edge, and a ruler or scale can then be used to measure pit depths along the straight edge.

In the absence of such islands or when the pipe is not straight, the remaining wall thickness must be measured with an ultrasonic device. This approach often requires grinding flat spots.

When grinding is required, it should be done carefully because any metal removal potentially lowers the failure pressure of the pipe. It is a good idea to start such grinding on high spots to obtain an indication of the remaining thickness.

Such spots can become reference surfaces for the straight-edge-and-ruler technique.

If none of the pits within an area of general corrosion exceeds 20% of the wall thickness, no repair is required regardless of the length of the corrosion. 1 However, the evaluator should determine by ultrasonic measurements that the remaining thickness at such a location is at least 80% of the minimum required wall thickness for the design stress level.

REGIONAL INTERACTION

Interaction of separate regions of metal loss must be considered if the separate regions are near enough to each other to have a combined effect on the strength of the pipe. Fig. 2 offers a guide to the types of interaction.

Type 1 groups are separated in the circumferential direction but overlap in the longitudinal direction.
When the circumferential separation between individual pits is less than six times the wall thickness (t), the pits should be considered to act together and the length should be taken as the overall length (L), shown in Fig. 2.
When the circumferential separation is 6t or greater, the pits should be analyzed separately.
Type 2 groups lie on the same axial line but are separated by an island of full wall thickness.
When the island of full wall thickness between the pits is less than 1 in., the pits should be treated as a continuous flaw the length of which is the overall length (L1 + L2 + L3).
When the separation distance exceeds 1 in., the pits can be considered separately (one with length L1; the other with length L2, and so forth).
Type 3 groups are one or more deep pits within longer, shallow corroded areas. This type of corrosion can be handled in one of three ways: by the use of standard formulas, as a shorter, deeper flaw in a pipe of reduced wall thickness (tr), or by iterative analytical methods (Table 1).
In an evaluation of a corroded area where the depth of the corrosion is greater than 50% of the original t and the circumferential extent is greater than 1/12 (8.33%) of the circumference, the circumferential extent should be measured and recorded.

Corrosion in or affecting welds, fittings, and bends may require special consideration. Rules of thumb for disposing of these situations are as follows:

Submerged-arc seam welds. Corrosion in a submerged-arc seam weld should be treated as if it were located in the parent pipe.
ERW (electric-resistance welded) seam welds. Occasionally, one finds that the centerline region of an ERW weld becomes preferentially or "selectively" corroded.
Such corrosion can be significantly more severe than the methods presented earlier would indicate. Therefore, evaluation of selective corrosion in an ERW seam by those methods is not recommended.
Girth welds. In some types of pipelines, girth welds are subject to preferential internal erosion-corrosion which leads to circumferential groove-like metal loss.
Preferential groove-like external corrosion is also common where tapewrap has "bridged" at a girth weld and water has migrated into this trap.
Corrosion in bends. Corrosion in field bends of 1.5D (D = OD) or greater (not elbows) can be evaluated by the methods used for straight pipe.
Fittings. Corrosion in or near the intersection of a branch and run pipe or on the intrados of a factory elbow cannot be evaluated by the methods presented earlier.

Such cases will require a detailed stress analysis of the particular condition.

PIT-INTERACTION PARAMETERS

If only one continuous area of metal loss exists, the overall length (L) and maximum depth (d) are used to evaluate its effect on the strength of the pipe. If one or more areas of metal loss exist in close proximity, their possible interaction needs consideration (Table 2).

The "worst case" is a composite of all of the profiles within a given metal-loss area. Combining profiles in this manner is necessary because it has been observed in tests of corroded pipe that the origin of failure encompasses the thinnest regions, as one might logically expect.

Obviously, there are limits beyond which widely separated profiles need not be combined if they cannot possibly interact. Concepts to be used in establishing whether interaction is to be taken into account are illustrated in Fig. 3.

Combining profiles requires some judgment based upon experience. When experienced personnel are unavailable, the most conservative procedure is to combine the worst-case profile from all of the individual profiles regardless of their circumferential separation.

ESTABLISHING DIMENSIONS

Key dimensions related to the effect of circumferential metal loss on the integrity of a pipeline are shown in Fig. 4.

In the following discussion of concepts for determining key dimensions, note that the measurement and evaluation of the longitudinal extent of corrosion is always of greatest importance and should always be done first.

Discussions of circumferential corrosion are based on an assumption of maximum longitudinal stresses at the top and bottom of the pipe and a neutral axis at the sides.

(Pipelines considered during the development of this method are those in place in Prudhoe Bay; they are therefore predominantly aboveground with longitudinal stresses in the top and bottom quadrants. The principles, however, apply equally to buried pipelines as well whose longitudinal stresses are predominantly in the lateral quadrants.)

If there are significant lateral loads, the sides cannot be ignored and must be analyzed in the same manner as the top and bottom.

If d 50%t and the circumferential extent of the corrosion is greater than 1/12 (8.33%) of the circumference, the maximum circumferential extent (C) of the corrosion shall be determined and indicated on a cross sectional drawing of the pipe, as shown in Fig. 4.

The area or areas of metal loss shall be positioned on the drawing relative to the top of the pipe. The critical dimensions for the top or bottom quadrant of the pipe are C and d, where: C = the maximum circumferential extent within the top or bottom quadrant and d = the maximum depth within the top or bottom quadrant.

In the example (Fig. 4), C1 = pi D/4 and represents the circumferential extent in the top quadrant of the pipe and d, is the maximum depth within that quadrant.

In the bottom quadrant of the pipe, the circumferential extent is C2 + C3 for two separate areas. The maximum depth is d3 because it is greater than d2.

Note that the corroded areas, both top and bottom, extend outside the top and bottom quadrants. Unless the location of the corrosion is near an expansion loop (or otherwise laterally loaded), however, only the metal loss within the top and bottom quadrants is relevant to the analysis.

Note that C and d may not necessarily occur at the same cross section. Once d has been located, the inspector needs to examine the pipe for a distance (D) on both sides of the location of d. The maximum C within D of the location of d should be used with d to define the effect of the circumferential extent.

Metal loss in the top quadrant and in the bottom quadrant should be considered separately. Based on the analysis presented in the concluding article of this series, it is reasonable to permit specific amounts of metal loss within either the top or the bottom quadrant.

Because the effects of missing metal differ in the two different quadrants, interaction would not be expected. Therefore, the maximum allowable metal loss may exist simultaneously in both quadrants of a given segment.

The allowable circumferential metal loss may be as much as half the wall thickness (50%t) with no restriction on length. If, after an analysis for the remaining hoop strength (that is, longitudinal defects) the pipe is not derated, one may use the following criteria to establish the effect of circumferential corrosion.

When the depth of penetration exceeds 50%t, the circumferential extent shall be limited as follows:

d < 0.5t, no limit on C
d 0.5t < 0.6t, C shall not exceed 1/6 of the circumference
d 0.6t but < 0.8t, C shall not exceed 1/12 of the circumference
d 0.8t, remove or repair

DETERMINING MAXIMUM PRESSURE

Once L, d (or t-d), and C have been determined, the design pressure (or current operating pressure) needs to be compared to a calculated "safe" pressure and the circumferential extent evaluated for acceptability.

Two levels of evaluation of corroded pipe are available; both are detailed in the concluding article of this series.

The first method involves determining the safe operating pressure on the basis of the overall axial length and maximum depth of the corroded area (ASME B31G analysis).

The second involves use of a more rigorous "iterative" method to determine the effective length of a corroded area and its corresponding safe operating pressure on the basis of a detailed map of pit depths or remaining wall thicknesses.

The latter is used only when the B31G analysis indicates the need for a reduction in pressure or a repair. Both levels entail consideration of possible interaction of groups of pits and consideration of the circumferential extent of the corrosion.

Whenever evaluation of an area of metal loss indicates that the safe operating pressure exceeds the maximum operating or design pressure, the affected piping segment may be recoated and returned to service.

If evaluation of an area of metal loss results in that area not meeting these requirements, the piping segment requires further evaluation. This evaluation may result in replacement or repair of the affected segment of pipe.

If it is determined that an area of corrosion requires a more detailed analysis, then detailed measurements of the metal-loss area must be made.

The analysis may then be carried out with the iterative procedure described by Kiefner and Vieth 1 and embodied in RSTRENG, a PC program available from the AGA and developed under project PR 3-805 sponsored by the AGA's Pipeline Research Committee.

The detailed analysis should be based upon pit-depth measurements taken at a maximum spacing of 1 in. generally along the axis of the pipe for a minimum distance of D on both sides of the deepest pit.

Several profiles of the metal loss are measured parallel (or nearly parallel) to the axis of the pipe. It may be preferable to select profiles that are slightly at an angle to the axis of the pipe to reduce the number of profiles taken.

The deviation from the axial direction, however, should not exceed 5; otherwise, the profile will be significantly affected by the curvature of the pipe. The main idea in selecting profiles is to record the deepest areas of metal loss within the boundaries of the corroded area.

STEP-BY-STEP EVALUATION

Fig. 5 presents a flow chart which provides a condensed guide for the step-by-step evaluation of an area of metal loss.

First, the evaluator needs to confirm the outside diameter (D), wall thickness (t), grade, seam type, product, and design pressure for the segment which is to be evaluated.

Next, calculate the maximum allowable operating pressure (MAOP) corresponding to the applicable design code and design factors.

The evaluator must determine the overall axial and circumferential extent of each separate area of metal loss.

The maximum depth of corrosion pitting (d) or the least remaining wall thickness (t-d) of each separate area is determined.

The overall axial and circumferential extent and the deepest penetration or least remaining wall thickness are recorded along with the pipe data of the affected segment.

The interaction rules presented previously apply.

If any one continuous area extends 100 in. or more along the axis of the pipe, one may consider for the purposes of analysis that the effective length is twice the diameter of the pipe provided that all of the metal loss has been examined and the deepest penetration or minimum remaining wall thickness has been identified and quantified to a high degree of certainty.

If the deepest pit (d) is determined to be less than or equal to 20% of the specified (nominal) wall thickness (t) and that the least remaining wall thickness (t-d) is at least 80% of the original (nominal) wall thickness, the segment containing the area is acceptable for continued service.

If the deepest penetration (d) is determined to be greater than 80% of the specified (nominal) wall thickness, or that the least remaining wall thickness is less than 20% of the specified (nominal) wall thickness, the segment containing the area of metal loss should be replaced or repaired.

Using d or t-d and L, one can establish with the procedure presented in Part 2, the "safe" operating pressure for the corroded area. If the safe pressure equals or exceeds the original MAOP of the pipe, the segment containing the metal loss can be considered acceptable for continued service.

If the safe pressure is less than the original MAOP but greater than or equal to the inspection pressure, one should then either set a new safe working pressure or specify an appropriate repair consistent with the design pressure.

If the safe pressure is less than the inspection pressure, necessary steps to reestablish safe conditions should be taken.

Using the limits described here, one can next establish the effect of circumferential corrosion on the segment of pipe.

REFERENCE

Kiefner, J. F., and Vieth, P. H., "A Modified Criterion for Evaluating the Remaining Strength of Corroded Pipe," Project PR 3-805: Pipeline Search Committee, American Gas Association (Dec. 22, 1989).