Structural Integrity of Storage Tanks
As seen in the May 2011 edition of HydroCarbon Engineering
Click here to download a PDF version
Advanced integrity assessments based on inspection results can demonstrate that tanks and their components are fit for continued service, allowing tank operators to often postpone repairs until the next shutdown, eliminate the need for repairs or be exempt from hydrostatic testing.
Improvements in inspection and computing technologies, coupled with the emergence of fitness-for-service standards, have advanced tank integrity assessments. Advanced inspection technologies including ultrasonic testing (UT) and laser scanning provide detailed mappings of corrosion, settlement, bulging and denting of tank shells and floors. Quest Integrity Group assesses fitness-for-service and determines remaining useful life of tanks using finite element analysis (FEA) simulations incorporating the measured condition of the tank. Simulations include crack stability and propagation, shell buckling and floor settlement. This article presents three examples to demonstrate the benefits of advanced structural integrity assessments: hydrostatic test exemption, tank floor settlement and local thinning.
Hydrostatic test exemption
The purpose of a hydrostatic test exemption analysis is to establish that the repaired tank is fit for continued service, without the need for a hydrostatic test, by application of detailed stress analysis and fracture mechanics technology per API 653. Quest Integrity routinely performs hydrostatic test exemptions for floor replacement, shell repairs for removing nozzles and new nozzle installations.
A hydrostatic test exemption analysis has many benefits. It eliminates the cost and time of the hydrostatic test itself, as well as treatment and disposal costs of the water used for the test. The exemption is particularly well suited for tanks repaired during the winter, when cold temperatures make hydrostatic testing extremely difficult. Furthermore, eliminating a hydrostatic test allows an operator to return the tank to service earlier.
To determine if the repaired tank is fit for service without the need for a hydrostatic test, a fracture mechanics analysis is performed to establish the critical sizes of surface connected defects located in the repaired regions. The tank is considered safe if all flaws are reliably detected and eliminated before the flaws grow to critical size. Furthermore, inspection technologies are limited in that flaws below a certain size cannot be detected (the detectability limit). For conservatism, the repaired region is always assumed to contain flaws sized at the detectability limit and these flaws must be smaller than the critical size for the tank to be considered safe. Otherwise, flaws may grow to critical size and cause catastrophic failure prior to being detectable.
In order to perform a fracture mechanics analysis (FMA), the stresses imposed on the region of interest must be established. A simple finite element analysis model can be used to simulate the storage tank hydrostatic test when the tank is filled with water (which has a specific gravity that may exceed that of the contents during operation). An example of an FEA model is shown in Figure 1. Shell thicknesses obtained from inspection data can be incorporated in the finite element model.
For FEA incorporating shell elements, the resulting stress can be portrayed at the inside and outside sur-faces of the element. Figure 2 shows an example of hoop stress results. The locations of maximum hoop stress on both the outside and inside are easily identified (in the first course of the tank shell). At these locations, the corresponding stresses on the opposite side of the shell thickness were then identi-fied, thereby establishing a linear, through-thickness stress profile. Figure 3 illustrates another example, showing that the bending stress at the bottom fillet weld connecting the tank wall to the tank bottom is of concern. Figure 4 shows examples of highly stressed locations.
Figure 2- Hoop stress results
The material fracture toughness is required to perform an accurate fracture mechanics analysis. Often, no material samples are available for testing, and the material toughness is determined from lower bound estimates described in API 579/ASME FFS-1.
Once the applied stress and material toughness are established, critical defect sizes for regions of inter-est can be calculated using the Signal™ Fitness-for-Service software package which is a Windows®-based program that implements methodologies described in the British Standards BS-7910 as well as API 579/ASME FFS-1 Part 9 Level 3. Signal™ Fitness-for-Service is a widely-used, commercial software package, developed and sold by Quest Integrity. For this work, the flaw is assumed to be a semi-elliptical surface crack in a flat plate with a surface length, “2c” and a depth, “a”. For a semi-elliptical flaw, there is no unique critical flaw size (2c and a). Instead, combinations of 2c and a resulting in critical flaws are plotted as a line on a graph of surface crack length (2c) versus crack depth (a).
Figure 5 shows an example critical flaw size curve (shown in blue). In the modified tank, the new welds will be inspected visually and by magnetic particle testing. Assuming that ASME standard procedures and acceptance limits are used, linear indications are only considered relevant if the length is greater than three times the width. Relevant rounded indications greater than 3/16 inches or 0.1875 inches are cause for rejection. A rounded indication is one of circular or elliptical shape with a length equal to or less than three times the width. For this example, using these criteria for acceptance, it is observed that the critical defect size curves established by fracture mechanics analysis methods lie entirely above the relevant rounded indication size. The finite element analysis and fracture mechanics calculations have shown for this example that the repairs made to the tank do not require a hydrostatic test. The tank is considered fit for service so long as no defects of the critical size or larger exist in the region of the new repairs. This statement can be made since the critical defect sizes exceed the defect sizes that would be cause for rejection during inspection.
The weld residual stress is also incorporated into the analysis for a more accurate assessment. Weld residual stress profiles are calculated using the Signal™ Fitness-for-Service which implements profiles described in Annex E of API 579/ASME FFS-1.
A hydrostatic test exemption does not waive the inspection of welds. New vertical and horizontal shell butt-welds should have complete penetration and fusion and finished welds should be fully radiographed to ensure that this is the case.
Tank bottoms and shells can settle for various reasons. API 653-2009 suggests periodically measuring settlement. Figure 6 shows an example of elevation measurements for a tank bottom. For edge settle-ment exceeding permissible values, API 653-2009 stipulates “…all shell-to-bottom welds and bottom welds should be inspected visually and with magnetic particle examination or liquid penetrant examina-tion. All indications should be repaired or evaluated for risk of brittle fracture and/or fatigue failure prior to returning the tank to service.”
Often, assessments demonstrate that the repairs are unnecessary, thereby allowing the operator to return the tank to service sooner than would have been possible with repair. The fitness-for-service assess-ments for tank settlement follow procedures that are nearly identical to the procedures previously de-scribed for the hydrostatic test exemptions.
Stresses in the bottom and shell are computed using a finite element model (FEM) incorporating an ax-isymmetric profile of the tank, settled soil and ring wall (if present). Modeling the contact and separation between the tank bottom and soil as well as the yielding of steel is necessary for an accurate analysis. Several fill cycles are simulated to assess if the tank is susceptible to fatigue or plastic ratcheting. Through-thickness stress profiles for key locations are obtained from the FEM. This stress profile is used to establish critical defect sizes of varying dimensions. The data is plotted on a graph of surface crack length (2c) versus crack depth (a). The dimensions for measured flaws and the minimum detectable flaws are plotted on the same graph to determine if these flaws are stable. Appropriate factors of safety should be incorporated to ensure a conservative assessment.
Figure 7 shows radial stress for a tank with edge settlement. In such cases, the critical location is often the underside of the tank bottom opposite the chime weld. Magnetic particle and liquid penetrant testing cannot identify flaws at this location. Measuring flaws at this location requires inspection techniques such as phased array UT.
The objective of a thinning assessment is to determine if the corroded tank is fit for service per API 579/ASME FFS-1, Part 4 (Assessment of General Metal Loss) or Part 5 (Assessment of Local Metal Loss). A common, rudimentary approach is to compare the minimum measured thickness against a code-specified minimum required thickness (for example, in ASME Section VIII). This is a rapid but over-ly conservative assessment that leads to many unnecessary repairs because the tank can still safely con-tain fluids.
API 579/ASME FFS-1 describes three assessment methods (Levels 1 through 3) with increasing levels of analysis sophistication. These methods calculate the pressure carrying (or rupture) capacity of the mate-rial with localized thinned areas. Portions within the thinned areas may yield, which in turn, redistributes stresses to adjacent parts. The methods in API 579/ASME FFS-1 incorporate this phenomenon leading to higher allowable internal pressures compared to simple calculations based on minimum thickness. Both Level 1 and 2 assessment methods were developed based on large databases of experiments and FEA simulation results of thinned areas. The two methods differ in the idealizations of the thickness pro-files of thinned areas. The Level 1 assessment method assumes a rectangular area circumscribing the thinned area with a “bathtub” depth profile, while the Level 2 assessment method incorporates the profile measured on a grid resulting in a “river bottom” profile (see Figure 8). Levels 1 and 2 assessment meth-ods are implemented in the SignalTM Fitness-for-Service software package.
For Level 3 assessments, stress redistribution is simulated using finite element models. The model is constructed incorporating measured shell wall thicknesses in the corroded area. Figure 9 shows exam-ples of FEA models incorporating wall thinning. For this type of assessment, Quest Integrity generates the mesh using proprietary mesh generation software that incorporates nozzles, heads, repads and cor-roded wall thickness profiles. The simulated strains are compared to strain limits stipulated in API 579/ASME FFS-1 to assess if the tank wall has sufficient capacity to resist hydrostatic, gravity, thermal and other loads.
If the tank does not satisfy Level 3 fitness-for-service requirements, rerated (decreased) fill height re-quired to satisfy the acceptance criteria can be calculated. Alternatively, if the assessment was based on assumed material strengths, the tank may be reassessed using material strengths measured on the actual vessel material (using in-field positive material identification techniques or cut-out samples). This typically results in higher yield and ultimate strengths and reduces conservative assumptions inherent in ASTM-specified minimum property values.
Advanced integrity assessments based on inspection results can demonstrate that tanks and their com-ponents are fit for continued service, allowing tank operators to often postpone repairs until the next shutdown, eliminate the need for repairs or be exempt from hydrostatic testing. Three examples were presented in this article, demonstrating the benefits of advanced integrity assessments: hydrostatic test exemption, tank floor settlement, and local thinning. Brief overviews of the assessment methodologies were also described. Advanced integrity assessments described here are not limited only to tanks, and are often used to demonstrate fitness-for-service of other structures such as pressure vessels, piping, furnaces, heat exchangers, coke drums and rotating equipment. Ultimately, the use of more advanced integrity assessments allows operators to more efficiently and cost effectively allocate their integrity management budget, concentrating resources and budget dollars as truly needed across their asset portfolio.