Abstrat MECHANICAL INEGRIY ASSESSMEN OF A LARGE HORIZONAL NGL PRESSURE VESSEL: CASE SUDY A methodology for assessing the strutural integrity of a large horizontal NGL (Natural Gas Liquid) vessel has been developed. he general analysis proedure, stress analysis and remaining useful life evaluation are desribed. Reommendations for dealing with anomalies deteted during assessment are also presented. he methodology employed an be applied to other, similar, pressurised vessels in the oil and gas, hemial and petrohemial plants. FERNANDO VICENE Maintenane and Reliability Integrity Engineer ABB Servie, Argentina INRODUCION In the oil and gas industry, pressure vessel integrity is a major onern. After internal and external inspetions various anomalies or defets an be reported and repairs ould be required in order to restore a pressure vessel to its original ondition. he fi rst question for an engineer, operator or manager is: an we keep operating at this pressure level? Is it safe, or do I have to stop the proess to arry out repairs? Strutural integrity assessment an be a useful tool for determining the suitability of a vessel for servie, and good maintenane management an redue the inspetion ost and extend the equipment life within safety standards. Pressurised equipment, suh as a large horizontal vessel in a typial gas plant an experiene in-servie damage. he vessel ondition deteriorates due to various fators mehanial, proess-related and orrosion-indued. he integrity assessment methodology inludes analyses of fitnessfor-servie and of remaining useful life, based on non-destrutive examination results and operating onditions. his ase study desribes how strutural integrity assessment methodology has been developed for appliation to a large horizontal NGL vessel, and the analysis proedure, stress analysis and remaining useful life evaluation are disussed. Reommendations for dealing with anomalies deteted during assessment are also presented. MEHOD he methodology applied by ABB Servie to an NGL gas plant aimed to maximise the pressure vessel's reliability and availability. he proedure aimed at identifying its mehanial behaviour under different proess ondition, understanding the potential damage mehanisms and obtaining aurate results from nondestrutive inspetions. he methodology used in this analysis onsisted of five steps, viz. (A) Creating a qualitative risk matrix and seleting equipment that required a deeper analysis; Figure 1 he methodology steps 36 Sept/Ot 2010 ME maintenane & asset management vol 25 no 5
Mehanial Integrity Assessment (B) Carrying out the analysis of equipment (stress analysis, potential damage mehanism, failure modes, proess ondition and maintenane strategy); (C) Quantifiation of inspetion results; (D) Fitness-for-Servie analysis; (E) Failure analysis. Key fators were to omplete every step orretly and to respet the sequene A to E (see Figure 1). Probability 5 4 3 2 1 Step A: Qualitative Risk Ranking In this fi rst part a qualitative risk analysis of the pressure equipment needed to be performed. his would result in the deletion from the analysis of muh equipment due to the low risk presented and some equipment would be onsidered for other types of analysis. he rule of thumb was that 20% of the equipment would aount for 80% of the risk, so the idea was to fous on that vital 20%. In this partiular study the qualitative risk presented by equipment was alulated following the standard speifiation from API 580 and API 581 'Risk Based Inspetion' [1,2], where the risk is defi ned as the produt of likelihood and onsequene, e.g. Risk = Likelihood x Consequene For this analysis the large pressure vessel had a low hane of suffering a failure, but the onsequenes (fi re and explosion) were high, so the risk was medium. Figure 2 shows the qualitative risk of the equipment. NGL Vessel A B C D E Consequene Figure 2 Qualitative risk analysis of the large horizontal vessel During the qualitative assessment phase some tehnial and maintenane management aspets were reviewed. o assign the ategory into whih the probability of failure falls the damage mehanism, failure mode, proess ondition, type of inspetion and equipment design Figure 3 were assessed. o alulate the onsequene of a failure ategory basi safety aspets were reviewed, suh as the volume enlosed, the toxiity, risk of fi re and explosion. In the ase disussed here the large horizontal NGL pressure vessel required a deeper analysis. Step B: Assessment One the risk of equipment had been determined qualitatively a deeper analysis ould be required or not, depending on the risk level assessed. A detailed analysis was arried out for this partiular pressure vessel. In this part of the proedure three tehnial aspets were reviewed, i.e. Mehanial behaviour of the large horizontal pressure vessel v(a stress analysis) Potential damage mehanisms Maintenane strategy Mehanial behaviour analysis of the large horizontal pressure vessel RISK High Medium High Medium Low Medium Risk=Require analysis of equipment Stress diagram of a large horizontal pressure vessel he aim here was to identify all ritial setions of the equipment: where the maximum stress was loated; what types of stress ould be developed during normal operation. From the strutural point of view large horizontal pressure vessels (Length/Diameter > 3) are different from vertial vessels and require more attention. Zik [3] onsiders a large horizontal pressure vessel as a beam supported by two-saddle supports resisting the shell plus liquid weight (reating a longitudinal bending stress at mid span) and the internal pressure. here were shear and irumferential stress onentrations at the horn of the saddle (see Figure 3). o simulate the normal operational ondition of the large horizontal NGL pressure vessel a linear fi nite element analysis was performed. he normal operational pressure, operation temperature, liquid and shell weight were onsidered for the stress analysis (see Figures 4, 5 & 6). Pressure vessel data Material: A516 Gr 70 N hikness: 70 mm Insulated: Yes Length: 31.000 mm Diameter: 5.000 mm Figure 4 Figure 5 Finite element model of pressure vessel. [Operating pressure 2.3 Mpa (23 bar)] Maximum hoop stress 83mpa Figures 5 and 6 indiate that maximum stress (90 Mpa) was loated between supports. High stress onentration on the saddle support was found and the hoop stress ating on the shell was 80 Mpa. Based on these results areful attention would be foused on these ritial points during internal and external inspetion. vol 25 no 5 maintenane & asset management ME Sept/Ot 2010 37
able I: Internal inspetion ativities ype of damage Damage mehanism Behaviour Non destrutive tehnique Inspetion effetivity Loss hikness CUI Internal orrosion General, loalized, pitting Spot thikness measurement and U san B Full visual inspetion Highly effetive (90%) Figure 6 Von Mises Stress, 90 Mpa Surfae-breaking flaw Mehanial failure due to overload. Visible deformation on saddle support. Visible deformation Liquid penetrant applied on seam weld loated between saddle support Highly effetive (80%-100%) Potential damage mehanisms A key fi rst step in managing the safety and reliability of equipment is to identify and understand the relevant damage mehanisms. Corret identifiation is very important when applying the risk-based approah to the maintenane of proess equipment. he ND tehnique employed needs to be appropriate to the nature of the damage mehanism and its failure mode. Information on this may be found in API Reommended Pratie RP 571, whih overs situations enountered in the refi ning and petrohemial industry in pressure vessels, piping, and tanks, and whih ategorises the failure mehanisms as follows [4] Mehanial and metallurgial failure Uniform or loalised loss of thikness High temperature orrosion Environmental assisted raking In this part of the proedure material onstrution, type of proess fluid, design onstrution praties (welding proess, nondestrutive manufaturing report, odes) and operational ondition are analysed. able II: In-servie inspetion ativities ype of damage External Loss hikness Surfaebreaking flaw and deformation Damage mehanism CUI Internal orrosion Overload due to proess ondition quantifying eah potential damage mehanism (identified in the previous step) via non-destrutive testing. Auray of the results was a key fator, so qualified and trained personnel are required on site. Behaviour General, loalized, pitting Visible deformation Non destrutive tehnique U spot thiness measurements Insulation visual inspetion + thermography inspetion if water entrane is suspeted Visual inspetion and liquid penetrant on saddle support (high stress onentrations loation) Inspetion effetivity Fairly effetive (50%) Highly effetive (90%) Ultrasoni thikness measurements and U San B Internal ultrasoni thikness measurements and San B were arried out on the large horizontal pressure vessel. Ultrasoni For the analysis reported here, i.e. of a large horizontal vessel, the potential damage mehanisms identified were Corrosion under insulation (CUI) Mehanial deformation Loss of thikness due to internal orrosion Analysis of the proess ondition and the operational history indiated that both fatigue and abnormal loading ould be disounted as potential damage mehanisms. Maintenane strategy One potential damage mehanisms were identified a maintenane strategy based on in-servie and out-of-servie (internal) inspetion was proposed (see ables I and II). Step C: Quantifying the inspetion results he aim of this step was to determine the atual ondition of the equipment, Figure 7 U San B for pressure vessel. hikness measured 70.63 mm; there was no loss of material (blak side). 38 Sept/Ot 2010 ME maintenane & asset management vol 25 no 5
Mehanial Integrity Assessment Figure 8 Weld seam inspetion. No disontinuities were deteted he pitting size was determined (pit diameter 2 mm, pit depth 1 mm), and a remaining life and fitness-for-servie assessment were then required. Step D: Fitness-for-servie and remaining life assessment Fitness for servie assessment (FFSA) may be defi ned as the quantitative analysis of the adequay of a omponent to perform its funtion in the presene of a defet. FFSA must inlude an evaluation of the remaining San B is a tehnique in whih the results are presented on a sreen type B, in whih the thikness ross-setion an be visualised. Using this type of ultrasound tehnique, performed from inside the equipment, orrosion under insulation (CUI) ould be deteted without removal of insulation. A orrosion rate of 0.04 mm/year was determined from ultrasoni thikness spot measurements, whih also indiated that there was no orrosion under the insulation (see Figure 7 bottom right opposite). a) Pitting orrosion in the bottom b) Pitting orrosion Liquid penetrant inspetion Liquid penetrant testing is a non-destrutive method of revealing disontinuities that are open to the surfaes of solid and essentially non-porous materials. A wide spetrum of flaws is detetable regardless of the onfiguration of the workpiee and regardless of flaw orientations. For this partiular ase liquid penetrant inspetion was foused on the weld seam loated on the shell, between the saddle supports (see Figure 8) and looking for anomalies that an develop in-servie or during the ereting phase. Visual inspetion Visual inspetion is a non-destrutive testing tehnique that provides a means of ) Pitting orrosion in the joint Figure 9 deteting and examining a variety of surfae flaws, suh as orrosion, ontamination, in the surfae fi nish, and surfae disontinuities on joints. Visual inspetion is also the most widely used method for deteting and examining those surfae raks whih are partiularly important beause of their relationship to strutural failure mehanisms. In this ase an internal visual inspetion was arried out. During the internal inspetion pitting orrosion in the bottom of the pressure vessel was deteted (see Figures 9 (a), (b), () and (d) above). d) Pitting orrosion in the bottom Pitting orrosion in the bottom of the pressure vessel life of a omponent. A damaged omponent may be aeptable at the present time, but its remaining life must be established. his assessment is needed to establish inspetion intervals and a basis for reliability-based inspetion (RBI) and it will help to determine the risk priorities relative to other plant that needs to be opened during the next turnaround. For this partiular ase the remaining life was alulated as reommended in API 510 and pitting orrosion was evaluated as reommended in Chapter 6 of API 579 [5, 6]. Figure 10 shows the antiipated future thikness redution. vol 25 no 5 maintenane & asset management ME Sept/Ot 2010 39
CR RL long term Remaining Life alulation ime between atual CR Where min initial initial atual and 70.52 mm 67.2 mm 83 years 0.04 mm/ year atual 70.8 mm 70.52 mm 0.04mm/ year 7years ( years) CR = Corrosion Rate RL = Remaining Life initial = initial wall thikness (mm) (he asnew thikness at fi rst measurement) atual = thikness (mm) measured during most reent inspetion min = minimum thikness required by pressure or strutural load, omputed at the design stage Step 2: Determining the wall thikness used in the assessment using the equations t t trq FCA 70.42 1.27 69.15mm Figure 11 Pitting hart Grade 2. Atual damage state. 71 hikness redution vs time Step 7: Comparing the photograph of the pit damage area with standard pit harts. mm 70 69 68 67 66 65 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 Figure 10 Years Fitness for servie assessment Remaining life 83 years Wall thikness redution of large pressure vessel In this part of the proedure pitting orrosion damage was evaluated applying the Level 1 assessment proedures of Chapter 6 of the API 579 ode [6] whih an be utilised to evaluate metal loss from pitting orrosion. Pitting is defi ned as loalised metal loss, and an therefore be haraterised by a pit diameter and a pit depth. he Level 1 proedure is simplified in that it does not aount for the orientation of the pit-ouple with respet to the maximum stress diretion. Results are onservative and based on pitting harts. Step 1: Determining the following parameters D = inside diameter of equipment, 5000 mm Loss = thikness loss, 0.38 mm (70.8 70.42) FCA = future orrosion allowane, 1.27 mm RSF a = allowable (non-dimensional) Remaining Strength Fator, 0.9 rq = wall thikness (mm) measured at the time of assessment hikness Minimum Allowable hikness Step 3: Loating area on the omponent that has the highest density (number of pits) of pitting damage (using photographs inluding a referene sale). Step 4: Determining the maximum pit depth w máx 1 mm Step 5: Determining the ratio of remaining wall thikness R t FCA w máx wt t Step 6: Determining the Maximum Allowable Working Pressure for the omponent using the Step 2 thikness Step 8: Determining the RSF from the table aompanying the pit hart and from R wt RSF = 0.99 Level 1 assessment would be aepted only if 1. R wt > 0.2 rue (see Step 5) 2. RSF > RSF a rue (see Steps 8 and 1) For the ase onerned the pitting damage was aeptable for the atual operating ondition, i.e. the MAWP of 4.4 Mpa. Step E: Root Cause Analysis (RCA) he purpose of RCA is to identify and understand the basi root of problems that affet the equipment performane and its integrity. By understanding how anomalies an originate these failures, or re-ourrene of suh problems in similar plant, an be avoided in future. For this reason it is very important, during pressure vessel inspetion, to analyse every sign or evidene that an be tested in laboratory. For the analysis reported here a sample of orrosion produt was taken from the pressure vessel bottom and subjeted to X-Ray diffration analysis (whih explores the sample's rystal struture). Sample results are shown in able III and Figure 12. 69.15 1.27 1 1.0039 69.15 2 S t 2163 69.15 MAWP( Mpa) 4. 2 R t 2 (2500 1.27 0.38) 69.15 4 Mp 40 Sept/Ot 2010 ME maintenane & asset management vol 25 no 5
Mehanial Integrity Assessment Deteted element Atomi % O 74 Al 0.48 Si 0.37 S 1.54 Fe 24.04 able III Figure 12 he diffration analysis revealed the presene of orrosion produts suh as goethite [FeO(OH]) and magnetite (Fe 3 O 4 ) with sulphur ontent (S). Both of these an be reated by the CO 2 and/or H 2 S that are typially found in NGL. Corrosion by CO 2 was disounted however, beause FeCO 3 was not identified during laboratory analysis. he presene of sulphur in the orrosion deposit was a strong indiation of Sulphate Reduing Bateria (SRB) attak. he presene of sand deposits at the bottom of the pressure vessel had ontributed to reating an environment for Mirobially Influened Corrosion (MIC). Figure 13 Corrosion produt analysis X-Ray diffration analysis Pitting orrosion under tuberle Mirobially Influened Corrosion [7] MIC orrosion is not fundamentally different from other types of aqueous eletrohemial orrosion; the differene is that the aggressive environment is produed by miro-organisms as produts of their metabolism. he most important group of bateria assoiated with orrosion is that of the sulphate-reduing bateria (SRB). In pratie, the great majority of MIC failures are related to the ativities of SRB, anaerobi (oxygen-free) bateria that obtain their required arbon from organi nutrients and their energy from the redution of sulphate to sulphide. Pitting is reated under tuberle deposits (see Figure 13 below). Root Cause Analysis for the large horizontal pressure vessel Water was left in the vessel after hydro-test when the vessel was fi rst put into servie. It may be that there was a period between the hydro-test and startup when some pitting ould have started under deposits, or in the open as normal rusting ourred. he fat that all the damage was in the vessel base onfi rmed the presene, in the base, of water whih had drained and remained there. Sulphate in the water would provide the nutrient for the SRB. Water does not aumulate in the vessel during normal operation. When the vessel is put bak into servie there should be no water present and none should be able to enter the system from outside, but SRB bateria ould be generated if even very small vestiges of water were to remain after hydrotesting. Engineers require water speifiation for suh testing and good heating praties to ensure water removal. Reommendations Grinding out the pits to give a smooth surfae without going beyond the orrosion allowane was reommended. his would remove loal stresses and remove all ontamination and traes of the moisture whih ould allow orrosion if any pitting, or ontained deposits, remained. CONCLUSIONS A good mehanial integrity programme for pressure vessels is ruial for those plants that need to redue turnaround time and inspetion ost within safety standards. When surveying large horizontal pressure vessels speial are should be taken when internal and external inspetions are arried out on the shell between the saddle supports. Visual inspetion should be undertaken very arefully at the horn of the saddle where the effet of irumferential bending stress may be signifiant. In the partiular ase reported here the fitness for servie assessment permitted the large NGL pressure vessel to operate at its design performane, i.e. an MAWP of 4.4 Mpa and a Maximum Allowable Operating Pressure (MAOP) of 2.3 Mpa. Even though the degree of pitting orrosion was aeptable, it was suggested that the existing pitting should be removed by grinding. Water trapping in an NGL pressure vessel is unlikely during servie. In this ase, however, it was shown that bad water speifiation and inadequate pre-start-up heating had affeted the mehanial integrity of the vessel. An MIC orrosion mehanism had been generated due to the vestiges of water remaining after hydro-testing, reating an environment favouring the growth of sulphate reduing bateria and then the pitting orrosion indued by the metabolism of suh bateria. When the vessel is put bak into servie there should be no water present and none should be able to enter the system from outside. BIBLIOGRAPHY 1. API RP580, Risk Based Inspetion (1st Ed), Washington DC, May 2002 2. API 581, Risk Based Inspetion, Base Resoure Doument (1st Ed), Washington DC, May 2000 3. Zik L P, Stresses in large horizontal pressure vessels on two saddle supports, Welding Journal Researh Supplement, Sept 1951 4. API RP 571, Damage mehanisms affeting fi xed equipment in the refi ning industry, De 2003 5. API 510, Pressure vessel inspetion ode: in-servie inspetion, rating, repair, and alteration (9th Ed), Washington DC, June 2006 6. API RP 579, Fitness for servie (2nd Ed), Washington DC, Marh 2006 7. ASM, Metals Handbook Volume 13, Corrosion, ASM International 1992 vol 25 no 5 maintenane & asset management ME Sept/Ot 2010 41