STORAGE tanks used for petroleum and related products need regular monitoring to ensure that they remain in good operating order. The American Petroleum Institute recently announced a new recommended practice for service fitness evaluations of aboveground storage tanks. A review of the recommended practice was presented in September 2000 in Austin, Texas, at API's fifth annual Storage Tank Management & Technology Conference.
Tanks degrade while in service, and deficiencies due to degradation or from original fabrication, may be found during inspections and assessments. API's Recommended Practice 579 provides guidance for conducting these fitness-for-service (FFS) assessments using methodologies that are specifically tailored to the storage equipment used by the refining and chemical process industries.
API 579 can be used to make run-repair-replace decisions to help ensure that pressurized equipment containing flaws can continue to operate safely. Methods in API 579 can be used for assessments based on "generally recognized good engineering practices" until references from inspection codes are provided.
The first fitness-for-service (FFS) assessment goal is run-repair-replace: running the equipment and monitoring flaws, such as corrosion or crack growth; altering the process for products, temperatures, rates, or equipment; repair the equipment; monitor while running as-is equipment; rerate, such as lowering the maximum fill height (MFH); or replace or retire the equipment. An FFS assessment also can provide guidance with inspection intervals. Another FFS goal is to present integrity and projected remaining life interval.
API 579 is applicable to pressurized components in tankage, pressure vessels, and piping. Specific codes of original construction are specified in API 579 for:
- tankage: API 620/650
- vessels: ASME BP&V section VIII division 1 and 2
- piping: ASME B31.1/B31.3.
API 579 principles also can be applied to other pressure-containing equipment constructed to recognized standards for public domain or internal corporate standards. "If you use API 579 with equipment not listed under codes of construction, then you should have sufficient knowledge for the allowable stress or other construction requirements in the equipment's original code so that you are able to factor or adjust some of the API 579 acceptance criteria," said Joel Andreani, M & M Engineering. "For example, if the original code of construction has slightly higher allowables, you have to scale API 579 acceptance criteria in proportion."
Section 1 of API 579 provides an introduction. Section 2 defines FFS engineering evaluation procedure. Sections 3 through 11 are assessments of the following: equipment for brittle fracture; general metal loss; localized metal loss; pitting corrosion; blisters and laminations; weld misalignments and shell distortions; crack-like flaws; equipment operating in the creep regime; and fire damage.
Appendix A provides thickness, maximum allowable working pressure (MAWP), and membrane stress equations for an FFS assessment. Appendix B covers a stress analysis overview for an FFS assessment. Appendices C and D are compendiums of stress intensity factor solutions and reference stress solutions. Residual stresses for an FFS assessment are found in Appendix E. Material properties for FFS assessments are listed in Appendix F. Deterioration and failure modes are located in Appendix G.
The 1,000-page document is organized into modules. Each section is based on a type of flaw or damage, such as crack-like flaws. "The document is primarily aimed at the petrochemical industry," Andreani said. "So the types of damage listed in API 579 are the ones you would see in petrochemical applications, even though they are present in other industries.
"Although the document essentially is stand alone, you do need some information from the original code of construction, such as allowable stress. API 579 makes liberal use of tables, charts, and mathematical data."
Each section is divided into subsections: applicability and limits of assessment; type of inspection data required to perform the assessment; mathematics used to determine if the flaw is acceptable; calculations for assessing remaining life; remediation techniques; and future monitoring. The end of each section provides reference information as to how the assessment technique was developed.
Levels of Assessment API 579 uses a multi-level assessment method. Assessments of higher levels become less conservative but require more detailed analysis and data. Level 1 is the least detailed method with a minimum amount of inspection or component information. The assessment can be performed by plant inspection or engineering personnel.
The information requirement for Level 2 is similar to Level 1, but the analyses are more detailed. Inspections typically are performed by plant engineers or engineering specialists with FFS experience. Level 3 requires the most detailed inspection information. Analysis includes some type of numerical technique, such as finite element analysis (FEA). Inspection must be performed by engineering specialists. An inspector is defined as someone who has API 510, 570, or 653 certification. An engineer is defined as someone with a degree in engineering and more than two years of experience.
Assessment methods generally use one or more of three acceptance criteria: allowable stress; remaining strength factor (RSF); and failure assessment diagram (FAD). However, the majority of API 579 is based on RSF and FAD acceptance criteria.
Allowable stress, such as fraction of yield, tensile, or rupture, has limited use in FFS because of the difficulty in establishing stress classifications for components with flaws. The most frequently used criteria for most types of flaw assessments is RSF, which is the ratio of limit or plastic collapse load for a damaged and undamaged component.
FAD is used for crack-like flaws. Failure is evaluated by two phenomena: unstable fracture (brittle) and limit load (ductile). "By using an x-y axis, you can judge the possibility of plastic collapse on one axis and the possibility of brittle fracture on the other axis at the same time," Andreani said.
FFS Engineering Assessment Section 2 of API 579 provides general background information for determining flaw-type assessments. It explains how to identify damage, collect data, and determine general assessment techniques, remaining life, and remediation. The main acceptance criteria for noncrack-like flaws is the RSF, which equals the limit or plastic collapse load of damaged component divided by the limit or plastic collapse load of undamaged component. An RSF subscript a (allowable) is needed to compare to the RSF for various flaws. A conservative value for RSF subscript a is 0.90. The RSF subscript a may be reduced for conditions such as short-term loads (wind) and/or based on consequence of failure (stores water).
"As part of run-repair-replace, you might decide to rerate the equipment capability," Andreani said. "For a storage tank, this would be the maximum fill height (MFH). The same ratio is applied to maximum allowable working pressure for piping and pressure vessels."
API 579 Section 3 provides criteria for the assessment of existing equipment for brittle fracture. Prior technology for assessment of brittle fracture was API 653 Section 3, which uses a decision flow chart and exemption curve method. Other technology includes traditional fracture mechanics.
API 579 Section 3 applies to API 620 and 650 nonrefrigerated tanks that are welded or riveted. The critical exposure temperature (CET) for an API 650 tank is defined as the lower of the lowest one-day mean temperature (LODMT) plus 15 F or the hydrostatic test temperature. CET is used in other sections of API 579.
Level 1 assessment requires that the material meet toughness requirements in API 650 seventh edition or later, or other recognized code. "For example, if the tank is built to an international standard and you meet the toughness requirement of that standard, then you would be at the Level 1 assessment for brittle fracture," Andreani said.
The first step of a Level 2 assessment is to evaluate conditions using Figure 3.2 flow chart, which is similar in content to API 653 Section 3. Answer the following questions to eliminate the possibility of brittle fracture:
- Has there been a previous hydrotest?
- Is the tank thickness less than 0.5 inch?
- Has the temperature exceeded 60 F?
- Is the maximum stress less than 8 ksi?
- Is the material exempt from impact testing?
- Has the tank been operated successfully at the lowest one-day mean temperature (LODMT)?
Level 3 assessment uses fracture mechanics. Determining critical flaw size is based on actual stress and actual toughness of the material. This assessment can be used when Level 1 or 2 has not been met and for a storage tank with refrigerated product. An aboveground storage tank that meets Level 1 or 2 can be used for continued service. Level 3 remaining life is associated with the time it takes a flaw to grow to its critical size.
For remediation of brittle fracture, a hydrotest can be used to enhance damage tolerance to crack-like flaws. It is recommended to perform at a temperature that will permit plastic flow without the possibility of brittle fracture. Typical temperature would be 30 F above minimum allowable temperature (MAT), which is a function of thickness. Changing factors, such as lowering fill height and specific gravity, can limit stress.
Assessment of General Metal Loss General metal loss results from corrosion and/or erosion on the inside or outside surface of the tank shell. Level 1 and 2 procedures apply when:
- Tanks are designed to a recognized code such as API 650 or 620.
- The area of metal loss has to be smooth and not notched.
- The tank is not subject to cyclic service. Appendix B defines a cycle as 20% more than the stress of the original design basis.
- No crack-like flaws are in the area of general loss.
- A fill height-thickness design relationship of the tank shell exists. This would not apply to a bottom-to-shell region of the tank or nozzles.
- Supplemental loads are negligible.
Level 2 procedures can be used on most supplemental loads, components without fill height, or equipment, such as a nozzle that has a thickness design equation. Two options are available for thickness data - point thickness readings (PTR) and critical thickness profile (CTP). PTR can be used when metal loss is general. A minimum of 15 data points is recommended. A sample data sheet for recording measurements is given in Table 4.2. Calculate coefficient of variation (COV) of thickness readings minus the future corrosion allowance (FCA). If the COV is greater than 10% and the metal loss is local, a CTP profile should be considered. A CTP is the measurements of thickness on a prescribed grid in the area of thinning.
There are numerous figures and tables in API 579 that provide guidance through the inspection and data requirements. For example, Figure 4.5 shows the inspection planes required for atmospheric storage tanks. The vertical lines M1 through M4 are longitudinal.
For Level 1 assessment, determine t subscript min, which is the minimum required thickness according to the code of construction. For storage tanks, the code of construction is determined by API 653 Section 2, t subscript min is for design conditions plus supplemental loads, and t subscript nom is furnished or nominal thickness. From the inspection grid, determine the minimum measured thickness, t subscript mm. Then determine L, which is the length for thickness averaging and is different than the L used in API 653. Compute the remaining thickness ratio (R subscript t), which is the minimum value less future corrosion divided by the required minimum from API 653.
"Using typical tank dimensions and reasonable remaining thickness ratios, you'll get about one to two feet for the value of L for tanks with a 40- to 120-ft diameter,"Andreani said. "If a tank with general thinning fails Level 1 assessment, you could reduce the tank's MFH based on the RSF ratio, use repair-replace-retire, or proceed to Level 2 or 3. However, Level 2 assessment of general thinning is no different for storage tanks; so you would have to skip to Level 3. Level 2 assessment differs only for piping and vessels."
Level 3 assessment uses numerical stress techniques in Appendix B, typically in terms of finite element analysis. Remaining life can be determined by calculating the anticipated corrosion rate (C subscript rate).
Remediation for general thinning is accomplished by reducing the C subscript rate. This can be done by decreasing the process temperature (but not past MAT); adding coatings/linings; or chemical injections.
Assessment of Local Metal Loss Local metal loss can be caused by local corrosion and/or erosion inside or outside the surface of the shell. Examples include the blend grinding of repair areas or mechanical damage, such as a gouge. Types of metal loss are locally thin areas (LTAs), grooves, and gouges. The length times the width of loss are the same magnitude for an LTA. A groove is an elongated area caused by corrosion or erosion. A gouge is an elongated area caused by some mechanical removal or relocation. The limits of use for Level 1 and 2 assessments for Section 5 essentially are the same as for Section 4. Additional data, however, will be needed for grooves and gouges.
For Level 1 assessment, a calculation is made for the shelf parameter (l) in addition to the values for t subscript mm, t subscript min, and R subscript t.
Then check the limiting flaw size criteria using the following mathematical values:
L subscript msd is the nearest distance from flaw to structural discontinuity.
To check for acceptability of a Level 1 assessment, R subscript t and l are plotted on an x-y graph. If the point lies under the curve, the local thinning is considered unacceptable. Local metal loss is considered acceptable if the point is located above the curve.
If a storage tank with an LTA fails Level 1 assessment, reduce the tank's MFH based on the RSF ratio, use repair-replace-retire, or proceed to Level 2 or 3 assessment.
"If you decide to rerate the storage tank instead of selecting a higher assessment level, you must determine the RSF ratio," Andreani said. "The RSF/RSF subscript a ratio would be applied to your maximum fill height (MFH) to determine the new reduced fill height of the storage tank."
Level 2 assessment involves additional mathematics to calculate a better estimate of RSF for rerating. Procedures for a Level 2 assessment ensure that the weakest ligament is evaluated:
- l#5.0
- Determine RSF as before but in a multi-step and fairly laborious procedure using equations 8.13 to 8.15
- Rank thickness in ascending order based on metal loss
- Start at point of maximum metal loss (t subscript mm)
- Subdivide the CTP and compute RSF superscript i for each subdivision
- Repeat for next ranked thickness - Use the minimum of all RSF superscript i and compare to RSF subscript a.
"The idea is to take progressively larger and larger areas of the flaw, evaluate the RSFs, and use the minimum of the RSFs in your rerating calculation," Andreani said. If a tank with an LTA fails Level 2 assessment, reduce the tank's MFH based on the RSF ratio (as before), use repair-replace-retire, or proceed to Level 3, which is the same as Section 4. Level 3 assessment uses numerical technique, such as finite element analysis.
Remaining life for local thinning (Section 5 type flaws) is more difficult to determine than that of general thinning. Two values can change: the actual corrosion rate of the metal and the size or growth of the region of local metal loss. It is an iterative process, which makes the monitoring process more important. Remediation is the same as Section 4, except to coat or line deep areas of metal loss, which may first require an application of a filler.
Assessment of Pitting Corrosion Section 6 of API 579 defines four types of pitting: widely scattered pitting over a significant region; LTA in a region of widely scattered pitting; localized regions of pitting; and pitting confined to within an LTA. Limits of use are essentially the same as Section 5. Level 2 assessment can be used for pitting on both ID and OD surface. Level 3 assessment is required for areas of structural discontinuity and pitting located where there is bending stress.
Pit-couple refers to adjacent pits. A couple is defined by dimensions of pits and distance between centers. Section 6 recommends that at least 10 unique pits be included in the inspection group. The following five dimensions help to define the pit-couple.
A significant amount of math is involved in the calculations involving pitting damage. The ultimate goal is to arrive at the RSF and compare it to the RSF subscript a, which is 0.90.
Widespread pitting is acceptable if the RSF is greater or equal to the RSF subscript a. For local pitting, use this RSF to get an equivalent LTA for Section 5 (t subscript eq=RSF x t subscript min). For local metal loss within widespread pitting, two separate calculations are necessary: determine the RSF for pitting and the RSF for the thinning. Then multiply them together: RSF subscript comb= RSF subscript pitRSF subscript lta. This is the value that is compared to the RSF subscript a (0.90). The metal loss is acceptable if RSF subscript comb RSF subscript a.
To complete a Level 1 assessment, certain pit diameter and depth ratios must be checked. A remaining thickness ratio of 20% applies to pits. If the diameter limit for a large pit fails, then treat it as a Section 5 LTA. If the RSF is greater than or equal to the RSF subscript a and the dimension limits are met, the tank is acceptable. If a tank with pitting fails Level 1 assessment, reduce the tank's MFH based on RSF ratio (as before), use repair-replace-retire, or proceed to Level 2 or 3.
Level 2 assessment is the only available method for asssessing pitting on both sides. It also introduces orientation of the pit with respect to principal stresses. Additional inspection data is required for a Level 2 assessment The ultimate goal, however, is the calculation of the RSF.
If a tank with pitting fails Level 2 assessment, reduce the tank's MFH based on the RSF ratio (as before), use repair-replace-retire, or proceed to Level 3 assessment, which typically is a limit-load technique found in Appendix B of API 579.
Remaining life is more difficult to determine than local thinning because of the possibility of pitting growing in size, density, and region. Thus monitoring becomes a critical issue if the decision is made to leave the pits in the tank.
Assessment of Blisters and Laminations Blisters and laminations are not a common type of damage in storage tanks. Prior technology is API 653 Section 2. Other technology, such as traditional fracture mechanics, would come into use if the blister or lamination resulted in a surface crack. Assessments of blisters and laminations are very qualitative. Level 1 assessment of blisters is acceptable if the various dimensional criteria are met; t subscript mm is greater than or equal to 50% of t subscript nom; there are no periphery cracks directed towards the surface; and L subscript w (weld) and L subscript msd (discontinuity) spacing criteria are met. Level 2 assessment also is qualitative. It generally is treated as an LTA per Section 5. Level 3 is a combination of nonlinear analyses found in Appendix B and fracture mechanics found in Section 9. Remaining life is not practical because it is difficult to determine growth.
API 579 Section 8 covers the assessment of weld misalignment and shell distortions. Weld misalignment can be offset, angular misalignment, or both. Shell distortion comes in the following forms: general; out-of-roundness (OOR); bulge; or dent. General distortion or OOR can affect the entire height of a shell by its deviation from a circular shape. Bulge is a local inward or outward deviation from an ideal circular geometry characterized by a local radius. Dent is an inward or outward deviation from an ideal circular geometry characterized by a sharper local radius than a bulge.
Regardless of the type of flaw, measurement of the flaw is key. Level 1 assessment is qualitative and makes use of fabrication tolerances. "I don't know of many problems that would meet this assessment," Andreani said. "The tank satisfies the assessment if all tolerances are met and service is not cyclic. But I would not recommend using repair-replace-retire for a tank based solely on failing Level 1 assessment."
Level 2 assessment is organized according to the type of damage. Weld misalignment and OOR are characterized as one type of damage. Calculate ratios of induced bending stress (R subscript b) due to misalignment /OOR to applied membrane stress. Select the proper set factors for OOR and weld misalignment. Finally, as before, the flaw is acceptable if RSF RSF subscript a, or 90%. Level 2 assessments of bulges also are based on local stress, which require calculations that eventually determine the RSF.
Level 3 assessment is frequently used for weld misalignments and shell distortions. Because of the type of geometry, these flaws often fail Level 1 and 2 assessments. Misalignment can be assessed using linear analysis and FEA. Shell distortions should be assessed using nonlinear analyses. In some cases, stability analyses may be required.
Assessment of Crack-Like Flaws Section 9 covers the assessment of crack-like flaws, which includes planar cracks, lack of fusion, lack of penetration, sharp groove-like corrosion, and environmental cracking. Other flaws that may be treated as crack are aligned porosity or inclusions and deep undercuts. Data requirements for assessment include flaw length and depth and numerous material properties listed in Appendix F.