Slip Blinds at Pressures Causing Permanent Deformation.pdf

Jaan Taagepera Chevron Energy Technology Co., 100 Chevron Way, Richmond CA 94802 e-mail: Trevor G. Seipp Fluor Canada Ltd., 55 Sunpark Plaza, SE, Calgary, AB, Canada T2Y 3X6 e-mail: Slip Blinds at Pressures Causing Permanent Deformation Slip blinds are frequently used for hydrotesting piping. In addition, when maintaining equipment or piping, the equipment or piping must be isolated to ensure a safe working environment. Separating fl ange pairs and inserting a blind fl ange against the process side prevents hazardous substances from entering the work area. Slip blinds are often used for this type of service. However, slip blinds are generally limited to low-pressure service since at excessive pressures the blind will become dished and may leak or become impossible to remove. For this paper, slip blinds of various sizes and thicknesses were hydrostatically tested to determine their deformation as a function of pressure. Nonlinear fi nite element analysis (FEA) was used to analytically determine the deformation of slip blinds. The goal of the testing and FEA was to determine allowable pressures that would limit permanent deformation of the blinds to specifi ed values. ?DOI: 10.1115/1.2716428? Introduction Piping spools are often isolated for hydrotesting in the fi eld by breaking a fl ange pair and inserting a blind to provide the isolation required for the hydrotest. Slip blindsround plates which are placed inside the bolt circle of standard piping fl angesare fre- quently preferred over blind fl anges for several reasons: Blind fl anges are more costly than slip blinds. Slip blinds are lighter and thus are easier and safer to handle, particularly in piping acces- sible only from scaffolding. Finally, slip blinds can fi t where fl ange pairs cannot be separated enough for standard blind fl anges to fi t. Vessels and heat exchangers are routinely taken out of service for inspection and maintenance. A safe working environment must be provided for personnel entering the equipment. One aspect of a safe working environment is breathable, nonhazardous air. Ensur- ing this environment requires a means of isolating the equipment from the piping normally attached to it. If a piping spool piece cannot be removed, other approaches for equipment isolation may have to be utilized. Block valves, when present, do not provide a 100% leak-free guarantee. Maintenance personnel working in a vessel isolated by only a single block valve is at risk of being exposed to potentially toxic chemicals. If the piping system is suffi ciently fl exible, a blind fl ange can be installed downstream of a single block valve to provide the required isolation. If a blind fl ange cannot be installed, a “double-block-and- bleed” setup may be used. A double-block-and-bleed setup con- sists of two block valves with a bleeder valve in between. The bleeder is left open to ensure that the downstream block valve sees no pressure. This provides reasonable assurance that no fl uid will enter the isolated equipment. In some cases, a pipe spool cannot be removed and the fl anges cannot be spread far enough apart to fi t a fully rated pressure blind fl ange into the gap. Only a single block valve is present or perhaps one of the two block valves required for a double block and bleed is inoperable. In these situations, a relatively thin slip blind may be used. Slip blinds, when properly installed, are recognized by the U.S. Occupational Safety and Health Administration ?1? as a lockout device for “Lockout/Tagout” programs. However, thin slip blinds are not usually considered to have adequate thickness to contain even moderate pressures. Slip blinds are used for both hydrotesting and equipment isola- tion. However, hydrotesting in the fi eld, as shown in Fig. 1, is probably the greatest cause of dished slip blinds. Although not likely to burst when exposed to moderate pressures, the slip blind may permanently deform by yielding at the inside diameter of the fl anges and become dished. A dished slip blind may be impossible to remove when the time comes to return the piping or equipment into service. When engineers are consulted for allowable pressures on slip blinds, limited resources are available to make this determination. The available resources include textbooks and piping codes that provide formulas to ensure that the blind does not yield. In an effort to determine allowable pressures higher than those allowed by linear calculations for 0.250 in., 0.375 in., and 0.500 in. thick slip blinds, fi ve sizes of slip blinds, 4 in. nominal pipe size, 8 in. NPS, 14 in. NPS, 20 in. NPS, and 24 in. NPS were hydrostati- cally tested. Nonlinear FEA was utilized to model the blinds that were hydrotested. Matching the modeling to the physical tests validated the FEA models. Allowable pressures for 0.250 in. thick blinds have been devel- oped that limit the deformation to 0.125 in. and 0.250 in. The target deformation of 0.125 in. was chosen since the deformed thickness of the blind would become 0.375 in. This is roughly the minimum gap between fl anges, which is reasonable in order to insert a gasket and slip blind. When the equipment is to be placed back into service, the gasket can be removed fi rst, thus making room to remove the now deformed blind. The 0.250 in. deforma- tion target was chosen simply to match the plate thickness. For the 0.375 in. and 0.500 in. thick blinds, target deformations were cho- sen as 0.125 in. ?based on gasket considerations?, half the plate thickness, and plate thickness. The slip blinds in this effort are assumed to be fabricated from commonly available carbon steels with a specifi ed minimum yield strength ?SMYS? of at least 35 ksi. For this effort, these steels are limited to SA36, SA515-65 and 70, and SA516-65 and 70 ?2?. Furthermore, the blinds are assumed to be used at ambient tem- peratures. This temperature assumption is clearly valid for hy- drotest situations. For equipment isolation uses, a slip-blind loca- tion would typically be downstream of a valve. Even if the valve were to leak and pressurize the slip blind, there would not be enough fl ow to heat up the blind signifi cantly above ambient tem- peratures. Linear Analysis Engineers asked to provide maximum allowable pressures for slip blinds would typically use a formula originally solved by Poisson and presented by Timoshenko ?3? for a circular plate with Contributed by the Pressure Vessel and Piping Division of ASME for publication in the JOURNAL OFPRESSUREVESSELTECHNOLOGY. Manuscript received January 23, 2006; fi nal manuscript received November 10, 2006. Review conducted by Dennis K. Williams. Paper presented at the 2005 ASME Pressure Vessels and Piping Con- ference ?PVP2005?, Denver, CO, July 1721, 2005. 248 / Vol. 129, MAY 2007Copyright 2007 by ASMETransactions of the ASME Downloaded From: / on 01/27/2018 Terms of Use: /about-asme/terms-of-use fl at edges. This formula is incorporated into the ASME B31.3 piping code ?4? and this same equation from B31.3 is used by B16.48 Steel Line Blanks ?5? in developing the thicknesses used for blinds in that standard. For a uniformly loaded circular plate with clamped edges, and assuming that the plate remains linear in geometry and materials, the maximum defl ection w occurs in the center of the plate with w = qa4 64D ?1? where q is pressure, a is radius, and D is the plate constant defi ned as D = Et3 12?1 v2? ?2? where E is Youngs modulus, t is the plate thickness, andvis Poissons ratio. The maximum bending stress occurs at the boundary of the plate: ?= 3 4 qa2 t2 ?3? Equation ?3? is based on the plate resisting the pressure prima- rily through bending. As the blind deforms, however, it begins to resist the pressure load through membrane stress in addition to the bending stress. Young ?6? recommends limiting Eq. ?3? for stress to situations where the defl ection does not exceed half of the plate thickness. Alternative, iterative formulas are provided for those situations where the defl ection exceeds half of the plate thickness. A comparison of results using Timoshenkos approach ?3?, Youngs ?Roarks? approach ?6?, and hydrotesting results was pre- sented in a paper by Taagepera and LaBounty ?7? in 2004. Hydrotesting Six test rigs were fabricated in order to run hydrostatic tests on various sizes of slip blinds. Figure 2 shows the test rigs used for the 24 in. and 14 in. tests. The rigs were fabricated as shown in Fig. 3 of carbon steel components. The slip blinds were cut from SA516-70 material with mill test reports ?MTRs?. Some of the material had yield stresses of 65 ksi, substantially higher than the specifi ed minimum yield strength of 38 ksi ?2? for that material. These high yield stresses were incorporated in validating the FEA models. Two raised face weld neck fl anges were used to sandwich a 1/16 in. thick corrugated metal graphite-coated ?CMGC? gasket and the slip blind. One fl ange had a pipe cap with pressure test connections welded to it. The optimal fl ange bolt torque to use as recommended by the gasket supplier ?8? results in a bolt stress of ?60 ksi. The hydrotest rigs were torqued to a bolt stress of ?45 ksi. This is a more realistic value for fl anges bolted up in the fi eld. Defl ection was measured directly with a dial indicator from the nonpressured side of the blind through the fl ange. Each blind was then pressurized, depressurized, and the re- sidual defl ection measured. The test pressure was incremented up ?increments varied for different sizes?, and the process repeated. Pressure was increased until a permanent deformation of between half the original plate thickness and 1 in. was achieved. In gen- eral, later testing was carried out to higher defl ections than earlier testing. Fig. 1Slip blind used for hydrotesting an exchanger Fig. 224 in. and 14 in. NPS slip-blind hydrotest rigs with Cl- 300 fl anges Fig. 3Typical slip-blind hydrotest rig Journal of Pressure Vessel TechnologyMAY 2007, Vol. 129 / 249 Downloaded From: / on 01/27/2018 Terms of Use: /about-asme/terms-of-use Nonlinear Analysis: FEA Validation A nonlinear fi nite element analysis was performed in order to provide better insight into the behavior of blinds in pipe sizes that were not hydrotested. The sizes that were tested were modeled fi rst in order to validate the FEA. Flange confi gurations are notoriously diffi cult to model accu- rately. Flange rotation, gasket material nonlinearities, bolt loading at assembly, the mechanical properties of the plate, and several other factors could all be accounted for in the model. In this case, however, some simplifi cations are possible. The blind is clamped between two fl anges that are substantially stiffer than the blind itself. Only one side of the blind has a gasket; the other side is in direct contact with the other fl ange. Thus, it is reasonable to model the blind as clamped on both edges without consideration for the fl anges and gasket. Furthermore, since the defl ection of the blind is toward the side without the gasket, it is unlikely that the gasket will contribute much to the overall stiffness of the blind. Therefore, the models consisted of only the blind. The blind was modeled as two-dimensional ?2D? axisymmetric using qua- dratic elements. The mesh for each model was chosen such that the results were independent of the mesh within 2%. On the top and bottom of the blind, vertical defl ection is restrained from the fl ange i.d. to the gasket o.d. Radial defl ection is not restrained. A typical model, showing typical mesh discretization and boundary conditions is shown in Fig. 4. The nonlinear material model that best fi t the hydrotest data was a linear elasticperfectly plastic ?EPP? material model, with the von Mises yield criteria used. The yield stress used was equal to the yield stress from the mill test report ?MTR? for each of the hydrotests, as shown in Table 1. Results of the FEA and the hydrotest for the 24 in. blind are shown in Fig. 5. The results for the 14 in. also closely tracked the data in a similar manner, as shown in Fig. 6, while the 4 in. results follow the general trend of the experimental data, as shown in Fig. 7. Although the 4 in. FEA results consistently underestimate the experimental results, at these pressures the defl ection as deter- mined both experimentally and through FEA is well over an order of magnitude below the defl ection limits for the recommended pressure listed in Fig. 8. As can be seen from the results shown in Figs. 68, the fi nite element model provides a reasonable approximation of the hy- drotest defl ection. Therefore, a high degree of confi dence can be placed in the FE model. Application of Validated FEA Model Having established the validity of the FE model using the MTR yield stress values, the yield stress used in the modeling was re- duced for the models that were run in order to establish maximum pressure recommendations. In order to capture a large portion of the available carbon steel plate material that might be used for slip blinds, a yield stress of 35 ksi was used for these models. This value would be appropriate for SA36 ?which has a SMYS of 36 ksi? and SA515-65 and SA516-65 ?which have a SMYS of 35 ksi?. It would also be applicable for SA515-70 and SA516-70 ?which have a SMYS of 38 ksi?. Results and Recommendation Figures 810 provide recommended allowable pressures for slip blinds fabricated from SA36, SA515 grades 65 or 70, or SA516 grades 65 or 70 plate installed between raised face fl anges. These recommended pressures were derived by reducing the FEA calculated pressure required to permanently deform a blind to the indicated amount by a factor of two-thirds. A design margin of 1.5 ?equivalent to reducing the pressure by 2/3? was chosen because it is the same design margin used against yield in the typical refi nery ASME standards ?B31.3, Section VIII, Division 1 and 2? ?4?. Exposing blinds to the pressures presented in Figs. 810 should result in residual defl ections, after depressuring, of less than the deformation shown. These pressures are compared to allowable pressures as calculated in accordance with Eq. ?15? of paragraph 304.5.3 of the B31.3 ?4? piping code. This calculation is based on SA516-70 or SA515-70 material and uses an allowable stress of 23.3 ksi and a blind diameter based on the inside diameter of the gasket. The allowable stress is determined as the lesser of 1/3 specifi ed minimum ultimate strength ?SMUS? or 2/3 SMYS. In this case, the values are 23.3 ksi and 25.3 ksi and the allowable stress is governed by the SMUS. The use of different design criteria leads to the signifi cant dif- ference in allowable pressures between this paper and B31.3 ?4?. Although the B31.3 criteria ?4? is to avoid yielding of the blind, the criteria used in this paper accepts yielding but limits perma- nent deformation. Accepting different design criteria would be expected to yield different results. It should be noted that the B31.3 formula ?3? compares well to the experimental results when using the yield stress of the material as the allowable stress and fi rst yield as a design criteria. Burst Testing After suffi cient pressure and defl ection data had been collected in the initial phase, attempts were made to determine the burst pressure of the 0.250 in. thick 14 in. and 24 in. slip blinds. These tests were conducted in a remote area with restricted access for safety. The dial indicator, which has a 1 in. range, was removed. Since the 24 in. rig had a convenient cross bar, which was used for both rigging and as a dial indicator mount, a simple guide was attached to the cross bar. This allowed a ruler to be placed on the blind in order to gather defl ection data while the blind was being pressurized. The fi rst attempt was limited not by the burst pressure of the blinds but by the available compressor that powered the hydrotest Fig. 4Typical FEA model Table 1Actual mill-test-report yield and ultimate stresses for slip-blind material 0.250 in., 4 in., some 14 in., and 24 in. NPS 0.250 in., 8 in., some 14 in., and 20 in. NPS0.375 in.0.500 in. Yield stress ?ksi?61625455 Ultimate stress ?ksi?76827475 250 / Vol. 129, MAY 2007Transactions of the ASME Downloaded From: / on 01/27/2018 Terms of Use: /about-asme/terms-of-use pump. This testing setup was able to generate between 1000 psi and 1100 psi pressure. Even with a blind thickness of only 0.250 in., however, this pressure was insuffi cient to burst the blinds. Residual defl ection was measured at ?2.7 in. for the 24 in. blind. Next, the 14 in. blind was tested to similar pressures with residual defl ection measured at ?0.8 in. A second burst test was then performed using a more powerful compressor. In this case, fl ange leakage limited the testing. The 0.250 in. thick 24 in. blind was pressurized to 1200 psi, which resulted in a residual defl ection of over 3 in. The 0.250 in. thick 14 in. blind withstood 2500 psi before the test rig experienced a fl ange leak. This resulted in a residual defl ection of roughly 1.7 in. FEA Burst Evaluation To further ensure the acceptability of the pressure recom- mended in Figs. 810, several additional FEA models were cre- ated. As with the models used to generate Figs. 810, the yield stress was set at 35 ksi. However, in the burst model FEAs, a nonzero tangent modulus was used and was set to 0.100 times the elastic modulus. The von Mises yield criteria and the isotropic hardening rule were also used, and geometric nonlinearities were also considered. Since all of the physical nonlinearities are in- cluded, upper-bound behavior would be expected, and these mod- els would be reasonable models against