TAYLOR FORGE ENGINEERED SYSTEMS, INC. 208 N. ron * Paola, Kansas 66071 Tel: 913-294-9331 + Fax: 913-294-5337 E-mail: engineered@ties.com February 23,2010 ‘Wayne Geyer STUSPFA 944 Donate Court Lake Zuri, TL 60047 Reference: Modem Flange Design Dear Wayne, Please find atached a copy of Modem Flange Design. This has been # kind of bile throughout the flange world for decades. Unfortunately, since Taylor Forge no longer manufactures flanges, this bulletin is no longer in print. ‘There ae earlier versions of essentially the same thing of which I've attached a cover of one from 1951. But the Modem Flange Design is the most complete and tll litle history in the front |Thope this helps. Let me know if you have any questions ‘Very tly yours, ‘TAYLOR FORGE ENGINEERED SYSTEMS, INC. Michael G. Kilkenny President ‘Traditionally DependableMODERN FLANGE DESIGN Bulletin 502Foreword: American Spicel Plpe Works was organized at the frst of the ‘century to produce spiral riveted pipe. This unique product was ‘widely used in water eupply systems Tor city equaducts, surtace mining, and penstocks in this country and overseas. A serious problem developed with this pipe when flanged joints were re- fuired. At that Ime the Nanges were made of cast iron ane were ‘ften broken during shipment and fabrication. J. Hall Tayler and his staff solved his problem by developing processes and equip- ment to manulacture a superior, tually unbreakable product— the forged flange. Then under a new name, Taylor Forge ané Pipe Works, the company pressed on to become a leader In Tange manufacturing ‘They did not stop with that, but topped their achievement by veloping analytic flange design formulas fr industry and Code tse, Professor £. 0. Watere of Yale University spearheaded tis work and joined. Taylor to.co-author the technical papers which Ted into the Tayler Forge publication Modern Flange Design. Industry Immedately accepted this convenient system of ange Gesign and used It worldwide This up-dated version of Modern Flange Design carries on the teadlion with a skilful Blend of data, graphs, and text. We sin- Cerely hope that this iaeue wil prove to be the mos! Useful of alTable of Contents Foreword, Introduction Side-step Designing Part A Designs. INTEGRAL * LOOSE * GPTIONAL WELDING NECK FLANGE DESIGN. SLIP-ON FLANGES RING FLANGES. BUND FLANGES. REVERSE FLANGES. ‘SLOTTED FLANGES UNUSUAL SHAPES EXAMPLE? WELDING NECK FLANGE DESIGN Full Face Design. EXAMPLE 2” SUP ON FLANGE DESIGN—FLAT FACED. Face and Gasket FACTCAS en y.5, Load and Moment. HARD AND SOFT GASKETS Bolts AREA SPACNG Stress FLANGES Baus Technical Data. ABLE OF CONTENTS. TABLES ANO GRAPHS BLANK FORMS. Symbols and Definitions Bibliography Appendx | — FV. f.FL, VL Program Listing Appendx lI — Bolt Torque Chart. Regional OfficesIntroduction Bulletin 502 was frst published in 1898 by Taylor Forge end Pipe Works es Modern Flange Design Ithas since Decome a standard celarence manual for the design of bates flanges—now called Part ‘A Flanges by the CodeO These are defined ae having gaskets wholly within the circle enclosed by the dot holes and no point of contact beyond this cite Part A rules limit the analysis to con- Bideration ofthe flange moment that reels from: + boltload + gasketioad + facepressuretoad + hysrostatic end load ‘The analytical aspect ofthis bulletin covering Part ‘A Flanges is based on paper number FSP-50-4 “Formulas for Stresses in Bolted Flanges Connec tions" This paper considered the effect of ta- ered tubs and built on an earler work by EO. Waters and J. Hall Tayior "The Strength of Pips Flanges? These two papers sill form the tech ‘ical base of the Code rules for flance design Later, the authors published the derivation of the flange formulas in a bookie titled, “Dovelopment ‘of General Formulae for Bolted Flanges." This saventh edition covers code revisions through the winter 7@ addenda, includes revised design lormulas for reverse ‘langes, hae a rofer- fence-orly paragraph with design shees for full face designe, adds a bibliography and acomputer program isting fo FV, 1, Fy, Vi, shape factors for hhubbed flanges. ‘The PVRC Subcommittes on Bolted Flarged Con- nections and the ASME Subgroup on Openings Cooperatvely developed design methods for flat face flanges In metal-to-metal contact. Rules for analyzing identical pars of such flanges were frst publshed in Mandatory Appendix Il of the 1971 Editon of the Code. These rules give a mathod Analysis that satisfies all conditions of qullbium, provides tor compaolly of rotation and transla. tion between elements. (hub and flange), and ‘accounts for the racal effect of pressure acting {on the pipe, hub and flange. The flanges designed by thego rules wore identified as "Parl B” te die- Lnguish tem from “Part A" ang ‘The theory used as the basis forthe rules inthe (Code for identical Part B flange palre wae e tended on a consistent basis to also cover non ideniical anges. Because such an analyse is laborious, a simpler method suitable for analrzing both identical and nonidentical pairs wae dovel- ‘ped and published es ASME Code Case "828: “m Simpliied Method for Analyzing Flat Face Flanges with Metal-To-Metal Gontact Outside the Bolt Crcia”" The Case was approved by Council ‘January 8, 1878, ater the rules for Part 8 flanges hea been transtorre trom Mandatory Appendix I {fo Nonmandatory Appendix ¥ in te Summer "77 Code Addenda ‘The simpler rules of Code Case 1828 resulted ‘rom Sesigning tor tangential contact between the flanges a their outside clameter, Algo many ol the assumptions and chars that apply equaly to Part ‘RorPart flanges wore used. It is beyond the scope of this Bulletin to cover the design of flat face flanges in meta-o-retal Contact, nowever, the work leading up tothe Case Is fuly documented in the open technical iter ture. For further information reter to bibliography items 6 through 12, PART A FLANGES: GASKET COMPLETELY INWARD FROM BOLTHOLES — NO OTHER CONTACT. reac ae aSidestep DesigningSidestep Designing ‘Any flange design work costs time and etfor, and leads one to ask, "Is there @ way I can sidestop esigning? ‘The answer for Part A Flanges is yos—two way. One way ie t0 consider that the Code recognizes the proper use of existing tlange dimensional standards. These are fly prepared and ready to be specified within the limits of their scope and rating. The Code incorporates ANSI B16.5 "Stes! Pipe Flanges . . "!" AP! 605 "Large Diameter Carbon Steel Flanges", ané ANSI B16.1 "Cast Tron Flanges... "Not included, but designed per Code rules ae industry Standards Classes 75, 178, and 950,* and MSS-SP4t "Pipe Line Flanges.” Note thatthe B16.5 ratings above ambient do not apply tothe large diameter NSS-SP44 flanges a6 thay are intended for use wih thin-wallhighyiels strength line pipe. These shuld be checked for Code applications; sizes over 36” In Classes 300, and greater were designed using higher than Code allowed stress levels. Flanges greater than 24” size that are required to match valves often must use ANSI BY6.1 cast iron dimensions. This may mean adjustments in mato- Fil, Boling, and facing. All such changes affect the rating, and require Code Inspector approval He will ften ask for supportig calculations. OPTIMUM INFORMATION INPUT ‘The other wey to. sidestep the task of designing is to call on TAYLOR FORGE for help. When shaped and Worbolt flanges departing not too greatly from the cieclar. For bath fypes, the bot spacing should be checked end an allowance made for malcistibution ‘of the momest, The spacing factor can bo less than is required for zicular Ranges since the metal avaliable inthe corners tends to spread the bolt load and even out Special care Is needed when the effective (highly svessea) par ofthe hub i interrupted. This may occur ‘with channel flanges on heat exchangers where large fozzle openings are loceted in a portion of the hub. Here's a case where the moment i introduced uniformly but it cannot be absorbed evenly. Inereased moment oF tocal reinforcement can help compensate for this con dition ‘The Code rules for flange design provides an alternate method of calculating hubs stresses for small diameter flanges. Whe the inside diameter Is less than 20 times the Rub thickness (20), 2, may be used for B in the formula for longitudinal hub stress, Sy. Calculate B, as follows: B, Bs 6, for loose hubbed flanges and integral flanges wnere the point corresponding to &./ae ‘andi falls below tho f= Lin. See the values ff chart, page 34. for Integra langes where f > 1 ‘This option ie not shown on sheets A & B— itis quoted here for reference, Soe the Cade paragraph UAT.1 paton conorions 2 ‘nse ae ese FE°10 X30 x he | 3418 OD ifie BE THK, ASB. COMP. = n= BOLE: nme 172 e277. evan rs 320077 a d3na hd 3 One = tb asfe = 20037/ HEE CAIGHATON- oping a win oR a= 73a PR = OOF =1Ae = 2.777 | S0a7 = ole = 47727 h 008 (2240 h=in = 0.192 = 3350), ee ayaa = 00.47 waa 390" ELBE Me HfFull Face Design je —a ef ca ee a Ry Cea on eee feat ee er ee ee eee aera erent ee er nes a wees Sen eae ae ed ee ee ner rs Peer ee eee en Meter ete eee eet area Snare ees Seer tie Panter Se ene eet er ey ere Seren ere eee ey retry Cee eee} Pree een eter ee ee ee eee er ed a eee eeiare Pereearin res CoUi oe ec cr arate ferred eeFick PAE QBEL 754 gwee Dibomerée. [sans a nor exoaTOR eerta = TS wewine 67 99¢ resets = Bs be 2079 Tae ZS S352 20-4 seanla -SW ae Ee CL CTFace and Gasket Figures to the right pitute @ variety of flange facing styles Used in industry, Spectic recommendations for facing/gasket ‘combinations are not within the scope ofthis manual, but pres ure, temperature, thermal shock, eyclle operation, and the fluid hancled should be considered. The various flange stand- ards give valuable suggestions In this area, along with dimen- sional data “Two general catagories are shown, The fist seals by bol force squeezing the gasket and this includes raised face, tongue and groove, ring type joint, mala and female, and lap jolat. The fongve face can be made with a small nubtin for extra gasket arp without bolt load penalty. The rubbin style also reduces Band is used with motel clad gaskes. Casket materials used inthis fist category include the various sheet stocks of rubber, cork, and asbestos compositions, Ietalic-lastomer mixes as the spiral wound, and solid metal Fings mage 98 flat washers or ring-yBe-Jint forms, ‘The other category is called sel-energizing or pressure actu- ated. This means that ital sealing is achieved by the gasket! facing geometty without significant bolt load, and the gasket eating force increases with pressure. O-rings made of elastomer materials are used most offen inthis style seal. Calculations for flanges vsing euch gaskets otten omit gasket loads and gasket Seating stresses, Was, Hy, and W, a8 being negligibly smal, Metallic sel-energizing ¢eals, not shown, are delta ring, double ‘cone, lens join, metalic O-ings, API AX. and BX ring joints.” ‘Allof these require very fine surface finishes, tight tolerances, and meticulous care in assembling. wd TONGUE 6 GROOVE a RING TYPE JOINT MALES FEMALE Ld LAP JOINT(FACTORS mb ‘The problem of speciying acceptable gasket factors has been a long-standing concern of the ASME Boller and Pressure Commities, ‘m= factor needed to held a geal under internal pressure, dimensiones. _y = nit seating load psi = etective gasket wich, In In 1941, Rossheim and Mark reviewed values then In use, made Changes, posentod m and y tablas for commony Used gasket mate- tials, and suggested rues for finding b. twas known that m increased {5 the inal ott load was given lower values —when the initial pressure was zero m wae vety high. Therefore, the authors decided that y cout be the m value required fora bolt load that would seal at Zero internal pressure, and proposed the empiical formula y= 180 (2it— 1)", Altrough Rossheim and Markl mage no claim other than they hoped to stimulate research, their m and y values were adopted by the Code. Gasket tests show leakage usually happens wall before the internal pressure has relleved the gasket reaction Suppose the ratio of the compressive stress of the gasket to the internal pressure ofthe fui trying to escape Is designated mat the instant wnen leakage start, then fora very poo" gasket ’ would be | alarge number. For good quality commercial gaskets it might range Between @ and 1.5, Fora theoretically perfect ease a leak would not start until the fluid pressure just exceeded the gasket pressure, with I=L, Assuming that reliable values of m’ are avaliable, the equllo~ fium conditions of Figure 4, can be expressed: bolt oad = (uid pressure) X (area subjected to Mulé pressure) +m’ (ld pressure) K (area, subjected to gasket pressure). In formula form this Is Wa, 9/4 G P+ 26Gmp where Wa isthe tot load (operating). ‘The fist term on the right sie is the Total hydrostatic pressure load acting on the effective gasket diameter G, ard the second is the {gasket reaction over an annular area 2) wide or the same G. Sealing theory requires that the load 2b-Gmp be resist by the gasket when ‘the internal preseure equals P. But the gasket also resists the total load, Wa Wien pressure Pie applied, load ~/« GFP Is removed and ZbeGmP remains as the theoretical load required to hold a light joint at . Fector mis derived trom m’ and has an added safety margin. Nether rm or y values have theoretical standing and those now in use are Based on practical experience and some fornel experimentation, ‘They have a direct effect on flange design and have been discussed {or years without reaching ix valves that could Be made mandatory. Many variables are involved, and much time Is required to make a single test¢ [At the present time, leakage criteria Is getting a “hard look” and esearch programs are under study by the Pressure Vessel Research Committee to determine if m and y values can be set up in relation to specified leak rates, For example, aint tat held for one minute without the escape of one drep of water was considered “tight” ‘This equals a leak rato of 10% co. per secord and would not be allowed In many industries. Possibly a leakage of 104 ce, per sec, Corresponding fo & drops per Nour, would be acceptable for liquids, While 10" would be necessary for heavy gas and 10" for lethal sub- ‘tances, This concept may lead to multiple istngs of gasket factors fora given material—withadiferent ovel of mand y values assigned to each leak cate an“The Code requires the analysis of two distinct load systems. The fist Is gasket seating — when there Is no internal pressure. To achieve 2 seal, all facing eurface iregularies must be filed with gasket material, This is done with direct force by bolting the flanges. The fequled seating toad Ie represented by: Was = bsGy andi Ik {edi Figure where the bolt ina (seating) is balancedby the gasket reaction ‘The other load system relates to Internal pressure and has four forces: Hy — the hydrostatic endtforce Hy — pressure orce onthe lange ace Hi, — the total gaaket load required to maintln geal W —the bolt load The firs three forces work to separate the flange pair. They are balanced by the fourth, W, which helds the assembly together. This |s iustated in Figure 6 ‘The hydrostatic end force Hy comes tothe lange from the closed fend of the pipe sysiem to which iis walded. Tha end force reaches the flange through the hub, and pulls onthe ring portion othe flange imid-hub ate large ond if ts a tapered hub. The fluid pressure force Hy ects direct on the face of the flange vihere ii exposed. For a gasket covering the entire raised face In Figure 6, Hy would equal 2er0, but as @ conservative allowance, laakage is assumed to be possible as far as G. The Hy force acts on a circle hall-way between B&G. The gesket load, H, = 2b-OmP whore gasket factor m relates the required gasket strets af design pressure P to that design pressure, For example, en m factor of 3 means thatthe residual gasket stress at P must be at least SP forthe Joint toe tight. In order to calculate the moment acting on a flange, the forces are multiplied by the appropriate lever arms which are measured trom the point of force application tothe bolt circle. The forees and the lover arms indicated in Figure 6 are for an Integral flange. ‘The gasket reaction load is goneraly assumed to decrease as internal pressure ls applied. The actual change is affected by flange rotation, bolt stratch, and the gaskets ablily to resist and recover from com= pression. Bolt loads frequently change too, for the same reason They must be retightened, especialy when gasket relaxation or creep occurs. All ofthese variables may be explained by the following ilutration, Leta par of flanges be represented by two steel bars placed aide by ‘sido at In Figure 7, page 20; they are separated by blocks, as shown, Which represent the gasket, and forces W at tho ends of the bars represent the bolt leads, which are balanced by opposite reactions atthe blacks. Under the action of the “bolt oad” the bars are pulled {ogether at thelr ends, the amount of deflection can be calculated by formulas in engineering handbooks. Consider now two cases: Load and MomentHARD GASKETS Hard Gaskets, Tho blocks are completely rigid, andthe bars deflect, Before intemal pressure is applied, the bolt load balances the gasket Feacton, and pulls the ends of the bars towards each other so that their separations reduced as the bolts are ghtened. For a cantilever, the ends of each bar would deflect a distance proportional to WL"/3, but due fo slight rotation as the bars pivot about the blocks, this I increased to (W2/3) (I-11. l/l). Therefore, Inthe inital boltup, the nuts are wrenched down unl the rod end aro brought closer together by tice this amourt, proportional to (Wh2/3) (2-+ h/t) Now, when pzessure is eppied to ihe bars between the blocks, equal inamount to Pl, and tending to spread the bars aper, the "bolt load W may be kepi constaat But the "gasket load" on the blocks Is Teduced; this may continue until in the ease of areal ilange jot, leakage ocours In any event, under the most favorable conditions P may be Increased until our ilustrative example, the pressure on the blocks ig completely ralieved, and Pl, = 2W. At this point, calcul tion of the detection ofthe ends of each bar indicates that iis more ‘than before, the total reduction of space betwaon the bars now being proportional to (WL/3) (2+ alll, + 05 (l/l) 2). In other words, In order to maintain ull bolt load while Internal pressure is applied, the effective bolt lengths must be shortened by wrenching down th rte til further. SOFT GASKETS Soft Gaskets. The bars are flexible, and romain straight and parallel ‘during all changes of loading. During bolv-up the blocks are com- pressed a small distance and the rod ends are pulled together the ‘Same amount; the "bot loa” = W. Then when pressure Plas applied between the block, the "gasket load” can any be relieved by lting the blocks return to thle erginal thickness; this means that the bars mist be separated by the same amour, from end to end, and the bolt lengine increased to avoid stressing the lt beyond the erginal value of W. ‘These illustrations represent two extremes, between which an end less variety of conditions can occurBolts AREA Area, The minimum tota! bolt area required (Aq) equals the larger (of Way/Ss oF Waa/S, where Sis the allowable bot stress t operating ronperature and 5, at ambisnt. Use A, to select bolt size and nur bet This is easlly done for most flange designs, but thin flanges at low pressure may require excess bolt area because’ 4) Bolt sizes <1/2" are easly overstressed, avoid using them when possible. 2) The number of bolts should be specified in multiples of four te guide ftup and alighment 9) Bolts must be close encugh to hold the seal SPACING Spacing. When a very few bolts wll satsty the Code but give bolt spacing 20 wide that leakage may occur, adjust the design by Increas- ing the flange thickness, Increasing the umber of bots, or both. ‘The will help the flange seal the gasket—even between bolts. The best Botting practice ie o combine smallest practical siz with min imum spacing to carry the tad. in 1950, Irving Roberts completed a detailed study of the interaction ‘of bolts, gasket and flanges Investigating the gasket factor mand the relation of Bolt spacing o leakage, he recognized that gasket ‘latcty dsibutes the load between bolts. He also derived a bolt spicing curve in terms of lango thickness, indicating a 5 percent loss of gasket pressure haf way belween bolts. Mr. Robert's work leads toa formula for “normal” bolt spacing when a workable gaskotto- fHange compression ratio Sis used, Normal bolt spacing = «SY, where 'S= compression of gasket/compression of tlange, or B (gasket thickness) (flange area)/, (lange thickness) (gesket area), E,E,= modulus of elasticiy of flange and gasket, respectively. Eis the uncertain factor in tie formula, but gasket manufacturers ‘of miscellaneous material tables might be data sources, ‘Ths formula for maximum bolt spacing, alge recognizes the impot~ tarce of gasket compression by using m. Maximum bolt spacing 20+ 6t/ (m 4 0.8) N aseumes proper load distriution when bolt spacing = 2x + t. Minimum bolt spacing is controlled by the require= ments of put ane wrench clearance—ee Table 1, page 28. If special tus and wrenches are Used to permit very close spacing, the flange Section may be seriously dillad away, and the weakening effect wil not be exposed by Part A analysis, discussed pp. 918.ee at oe le et) ABOVE CODE ALLOWANCE: EXPERIENCEPattFLANGES Example 1, Page 12 shows the stresses computed for Part A tlanges, Each stress must be equal of Tess than the Code allowable except forthe longl- tudinal hub stees which may reach 1.5 x the allon- ble. This is permitted because Sy is a bending Stress. The formula gives a maximum value which Only exists on the inside and outside suraces ofthe hub, and decreases 10 zer0 at a point half way Between, Ifa slight overstress in the hub causes Yielding, the load shits to the ring portion of the Tange, The ring is also subjected primarily to bend- Ing end thos able fo abeor> the additonal Inad 0 that a new equiiorivm within sate limits is estab- ished. When localized yielding extends fo the point whore the lange and hub cease to act elastically fan integral structure, the flange takes a perma ent set or "dishes and the joint usualy cannot Be uisecsombled and remade In esatistactory man- fer. This i ikly fo oecur when an integral flange hhas a thin straight hub Because bending stress and direct pressure stresses must be carried at the Same time, If there i eyelle loading, fatigue, or operation at high temperature in the creep range, withthe stress above ine yield point over a consic- trable portion ofthe hub, the use of Sy = 1.5 allow: fable should be reviewed. Itis resonable to expect that stress concentrations higher in magritude could be withstood betore the fienge ceases to act olastically but at this time ‘methods for seting thie crtcal level accurately are ‘ery laborious. These limitations apply to ductile materials having elongations of 15% or mare (in2") nd yield poinis of SO% to 70% of the ulimate Tensile strength, High strength materials may have Yield points very close to the ultimate tensile Strength or no abservebie yield point, and with such mmeterals the higher hub stress should ony be used ater earetul consideration “The existence of localized stresses, stress concen trations, and discontinuity strosses of 6 relatively high orderin all pressure equipment is well known. ‘Typical examples are strese concentrations around nozzle and branch connections, stiffener fiat head attachments. The Code Committe the Pressure Vessel Research Commit are cur- ‘ently studying such effects in an effort to arive at {better method of evaluating sate limits for design. This not likely at any simple answer will be found. “The character ofthe stress and the area over which it exists must be evaluates in adaltion tothe capac- liyof the adjacent structure to carry increased loa. ‘The Code accounts for localized stresses by using Compensating factors io the design formulas for In flange design, the most obvious example of 128 concentation oscurs with hubbed flanges at the comer where the outside eudace of the hub meets the back face of the flange. Integral flanges provide fr this with filet radius that ls required to be 0.28 gg minimum — But never less than 0.188BOLTS ‘The allowable bolt stesses listed in the Code may eam te be foo conservative, but they were very Ceretully selected by the Commitiees. Particular felerence wae made to stresses in flange Botting It was recognized that Dols are often stressed fabove the tabulated values, and that relaxation ‘curs in many service applications due to creep in the gasket, flange or bolt materi [Appendix S of the Code distinguishes between the design valve of Bolt stress and that which might Sclually be needed for both design and test con- Siions. The Inia tightening of the bolts prior to hhycrostatic test may be the most severe load that they wil receive. Ite sometimes thought that bolts stressed 1% times the allowable will pass the test, but thse not necessarily true. As internal pressure is applied, flange rotation and gasket properties ‘may actually lower elt load. To offst this the bolts must be tightened as internal pressure is added, or there wil be leakage. After the test, the flange will tend to relax to Its original flatness, the gasket will Fecover a large part of its original thickness, and the twice tightened bolts may be loaded beyond the Gesign valve, Bolt capacity fer higher intial loading Js desirable ana experience bears this out Code allowable bolt steses also account for se- Vere torsion and bending during service. The actual tension achieved In each bolt Ie never exactly un form whether by hand wrenching with plain or torque wrenches, power driven impact wrenches, ‘or hydraulio bolt tensioners. Tests show that man= {al tightening develope bolt stresses as high a6 145,000/ a pl. This incieates why a potential prob tem exists when bolts 34" and less are specified. ‘Allowable bolt stress ie related to bolt temperature and in uninaulated piping the bolts are cooler than the contained medium. it Is suggested that the higher stresses permited at lower temperature be voided unless fllabla test data ls available and Specifications permit their use, See ANSI 891.9 Paragraph 301 3.1Technical Data fora oN Me ee CLS FACTORS Foe e NS totes olen otes erterer RAN e aa rates) NAM aslo Weare aa ce WOVE tela wearer) NIV ==fa a Oeste) VALUES GF f [Ua S16|GRAPH Sera MY otal tata ete Sea None ees agetal ieee asics) Saget ae ee Sate aan aaNet Waa ie Saga to aot e ee) a ey Welesa et ea ee et) een Shae eae eneThreads | Root | foot Racial | Eos | alee Me | Sece | aver | acon | Acro | soir | outa | oitance| Russ | wreac ae | etn | oan | Fit | dar pia bet eo | Stoner a a6 20F * Te to EH 30, BL RL oe + ae | 159 = |e | Be 5 Pr en 1388 we | oR | Be > in | 33 20 Be | |e ois ian me Le om Pasig Tne 1% 4 5 | dais Ba we oh ay a | So iss me | he ae Bm | ow | ais is Sha | Bt ah mm | a) see wae] sx | 28 | am wh iB. bi K=AlB ‘The values of 7, Z, Y, and U in Table 2 have been Computed using Poisson's ratio= 03 and cover {00% H < 8.00. The incremanta of K are chosen {fo provide caleulation accuracy that is consistent with formulas ofthe Code. Vales of 7, 2, Y, and U fre listed with theee of four significant figures, and the diference between successive tabulation is not greater than #% i the range 1.100 = K 6.00. Linear interpolation is adequate for values of Z, ¥, and U when 1.000 < K's 1.020 where T is almost ‘constant When K ranges less than 1.020, the values of Z, Y, tnd U change rapialy and the Interpolation Formu: las should be used 2s follows: v=s0m[? + s90~4 “au 103) Tas below: where r= K=1 Wen K ls beyond the scope of Table 2, or @ mate- rial with a diferent Poisson's ratio is required, teleuiate 7, Z, ¥ and U according o the following formulas. These are equal to those in the Code Figure UA — 511 when Poisson's ratio U =03. as mo) y Tao F ee) BO Rt yey Ht yea-ou (1+ 46082 2£¥ tog,yK) —1 “To OI) KD AFD vPease is Penner| evans | BUSSE | 8b8Es | Conse] Z=nee | Goecs| sasua|sous pees| s82e3| 82e,Taste va49 GASKET MATERIALS AND CONTACT FACINGS! Gast Factors in) fr Operating Codtons ard Mii Deson Seating Stes (y) |= Sa Enennn Typ Sgt A Bagman re ° hot bie or high percentage of into, ois wey Sietoyand etal, ah — ~ ceo ar | re nan, rior Comapsed ese Fle ntl jacered ae “ ing, dsp ar nt. tradi, a niin, an ae Nones: oe - SVR ese atm conn ed ae aia nt a Sits geting wit rn let henge (2) Terie of wpaket hun «apo tbe amat e mab 238 SeBreer TABLE UA492 EFFECTIVE GASKET WIDTH (saw GZ NOTE hen be $ Kin Cia we NOTE: The gases Emily ithe theVALUES OF F.(UA-61.4)3 mone ukeseats"psi conorons Corse te Wry + Ws 3 [me Onertin [or at CALCULATION Operating nto aesSYMBOLS & DEFINITIONS ee Part Aisa type of flange having the gasket wholly delta — stress formula factor = #74 within the ciele of the bolt holes and no contact penitence B= raéial distance rom C10, in. ¢ = hub shape factor; for integral flanges «: PartBis atypect lange wth meta-to-metal contact = outside the bolt circle, identical pairs only LET OL F = hub shape factor for integral flanges nce Cae 1029 Farge Toe wih ntatemeta Con atta Wa cl hauing lente or, = hub shape ltr orlose nonin ga, & dean by Be silted |” iy syesscorecton tat for Iepal Manage wen the chat rena" oes (A = cuside diomaar of Hangs. For sated ten < une fanget the dot! of he cele tangent {Se nnr ape st noe = iamoter at location of gasket load reaction, In. Except as noted in sketch e of Figur t, actual total cross-sectional area of bolts at Gis detined 20 follows: root of tread of saction of least diameter Under stress, in? a When by < % In, = mean diameter of gasket contact face int When by > Win, = outside of gasket contact face less 22 iamoter ‘Aq = total required bolt cross-section ar ‘& = nominal bot diameter, in = alpha — stress formula factor, te-+1 41% chews of hb of salen in ee ee {t= thickness of hub at back of lange, in 1 EDO) 1 = gamma—siress formula factor=a/7, o! Te e a {or reverse flanges for reverse lange B= toll hydrostatic end force = +/4 GP, Ib. B= inside diameter of tenge, in. When 3 is Hy = hydrostatic end force on area inside of flange = «/4 BE, is asket load, operating =| Tess than 20g, tls optional forthe designer to substitute B, for B in the Code formula {or longitudinal hub tress Sy = Wa = Hy Ibs 1B, = B+ g, for loose hubbed flanges and also He = gasket load, seating = W, Ibs {or integral flanges when {<1 pinto Randee wien (= ol H, = total jint-contact-suriace compression B= ge lor integral anges when {21 ee Gane wane b= effective gasket or joint-contact-surtace seating wit, In Hy = difference between total hydrostatic end force and the hyérostatic end for an 2b = elective gasket or jolnt-contact-surtace rol ere Drossure width, in, CELE by basic gasket seating width, in eee | | hy = radial distance from Co the circle on which P= Bota — stress formula factor= 4/3 te +1 Botcicle diameter in. Hp act, in radial distance from @ to. © = (C—a)22,10 Ie ub shape factor, fr itr anges d= Whoa, for loose flanges. U of for Ure a= Uhh he he K = radial distance from Ca the circle on which Hy acts, in = A/8, A for reverse anges reverse flanges d= Uae 8 flanges a = Ua Bolt hole dla Xo(OY A= lambda — stress formula fastor—y +8 mg = M/B—unit lone, operating, Bfy/B' for e- verse flanges, Ibs. mp = Mo/B—unt load, seating, Me/B’ for e- verge flanges, Ibs ‘Mp = component of moment due to Hy Inch-pounds ‘Mo = total moment acting on the tlange, seating ‘Wh ineh-pounds Mz = component of moment due te Ho, operating Hos inch-pounds AM, = total moment acting upon the tange, for ‘operating conditions, inch pounds, Hoh, [My = component of moment du t> Hy = Hyhr m= gasket factor N= width, used to determine the basic gasket Seating wiath by, based upon the possible contact width ofthe gasket, n n= number of bolts + = Nu—Poisson’s ratio= 03 or steet P= cosign pressure, ps! R= ragal cietance from bolt circle to point of Intersection of hub and back of flange for Integral and hubbed flanges In. = radial stance from bolt circle to bore, ring flange, in. ‘Sa = allowable bolt stress at atmospheric tem- perature, psi allowable bolt stiess at design temp, pal Sw. = allowable design stress for material of flange, nozzle neck, vessel or pipe wall, at atmospheric temperature, pal Sq = calculated longitudinal stress in hub, pst Se = calculated radial str 8 in lange Sp = calculated tangential stress h flange, es} 1 = Kescior Ty = K-factor for reverse flanges f= flange thickness, in. = plpe wall thickness, in. U = Keactor Un = Ketactor for reverse flanges V__ = hub shape factor or integral flanges Vy, = hub shape factor for loose flanges w= width used to determine the basic gasket seating wicth by, Based upon the contact ‘width between the flange facing and the (asket (see Table UA-492), in W- = flange design bolt load for operating or seating condtions, 28 may apply, Ib. require bot load, operating conditions ibs. rin 914. bot load for gasket seating, Ibs Y= Keactor -R-tactor for revarae flanges y= gasket of joint-contactsurtace unt seating | toed, ps! 2 = KeactorWaters, €, 0, Wesstrom, D. B., Rossheim, D. 8., and Willams, "Formulas for Stresses in Bolted Flanged Connections," ‘Trane. ASME, Volume 59, 1937, Pages 161-168, Waters, . 0, and Taylor J H, “The Strength of Pipe Flanges, Mechanical Engineering, Volume 49, Mid:May 1927, Pages 53t- E. 0, Rossheim, D. B., Wesstrom, D. 8. and Willams, F.S Gy Development of General Formulas for Bolted Flanges, ‘Taylor Forge and Pipe Works, Chicago, line, 1948 5. Taylor Forge Catalog 871 13, 14 15. 16. ‘Schneider, A. W., Flat Face Flanges with Metal-to Metal Contact Beyond the Bolt Circle," Journal of Engineering for Power, Trans. ‘ASME, Serles A, Vol 80, No.1, Jan, 1868, pp. 82-88 ‘Waters. ©. and Schneider, R.W., "Axisymmetric, Non-identical, Flat Fece Flanges with Metal-to-Metal Contact Beyond the Bolt Circe,” Joural of Engineering for Industy, Trans. ASME, Series B, Vol $1, No.3, Aug. 1969, pp. 615-822. ‘Waters, E. 0, "Derivation of Code Formulas for Part 8 Flanges," [WRC Bullets No, 168, Oct. 1971, Schneider, R. W. and Waters, E.0., "Some Considerations Ro atrding the Analysis of Part 8 Code Flanges — 1974," ASME Paper No. 75-°VP-48, 1875. ‘Schneider, A.W. and Waters, E.0., "The Background of ASME Code Case 1828: A Simplitied Method of Analyzing Part B Flanges,” Tras, ASME, Journal of Pressure Vesse! Technology, Vol. 100, No. 2, May 1878, pp. 215-218. Schneider, R. W. and Waters, E. 0, "The Application of ASME Code Cate 1828," Trans. ASME, Journal of Pressure Vessel ‘Technology, Vo. 101, No. 1, February 1878, pp. 87-94 A Simpliied Method tor Analyzing Flat Face Flanges with Mota- fo-Metal Contact Outside the Bolt Circle, ASNE Code Case 1828, ‘Steo! Pipe Flenges and Flanged Fitings, ANSI Standard B1ES. Large Diamoter Carbon Steo! Flanges, API Standard 605 Pipe Lino Flarges, MSS-SP44, (Cast Iron Flaages and Flanged Fitings, ANSI Standard 16.1 Kent, G. A. “Gelecting Gaskets fer Flanged Joint,” Chemical Engineoring, May 27,1978, Raut, H.D., snd Leon, G. F. "Report of Gasket Factor Test,” WRC Bulletin 233, Docomber 1977 ‘Specification for WelIhasd Equipment 6A, Ametican Petroleum Inattute. Rosshelm, D8. and Markl, A. A.C. "Gaskot-Loading Constants,” Mechanical Engineering, September, 1943, pp. 647-648 Roberts, |. “Baskets and Bolted Joints.” Journal ot Applied Mochanies, Trans. ASME Vol. 17, June, 1950, pp. 169-170.APPENDIX Computor Program Listing That Calculates F, ¥, 1, Fs, Ve ‘This program was writen by E. C. Rodabaugh of Battelle Memoria! Insitute and his permission to reprint Is gratefully acknowledged ‘The program is based on formulas developed in relerence (4), Minor editorial changes have been made to publish th listing inthis manual. The program was adapted to fun on a 370/145 under OS. C CALCULATE FLANGE FACTORS, Fy vs SHALL-Fy E-su! DET COLL+012401340214022+523,031 ,0324033) = 1 O11s0229033 + 91280230031 + 01340324921 2=(O31*022013 + 03240234011 + 033*D12¢021) 50 READ (5,1,END=100) G1GO, HHO T FORMAT (210.5) C_G1GO= GL/60, HHO = H/soRTisecoD To CONTINUE A= 6160-1. B= 43. 608HnO9ea CAs ALPHA, B= KAPPA PRINT 2, C160, HHO FORMAT (9H ee LPLELOL5 8H H/HO=, . Az = ABA AS = Aves Dilei./3-* A/I2e cui=oi1 O12=5./42. + 1T.#A/336+ Ciz=0i2_ D13=1./210- + A/360~ Le AND v=sue—L Digs 11.7360. + 59-7040, + (1s43.¥AI 78 4 0151.90. + 5-BA/IO08. = (1s ¥ATR¥378 —| | £15=0i! Ol6=1./120. + 17-*8/5000- + 1.78 |—___thewpig | Datecie 022=215,/2772. * S1.#A/I2S2. + 160.7%. + 325-487 Toe Lr 1s.ehgit, + S20hd/2) 18 c2z=022 a + 128.#8/65045. + (64/2. + 154A/Ts + 12.082 _| Te setas/iles 7B 23-023 024=533.730040, + 653.807 13900, + 1e/es + SS0RATIO® TaneD34 + 025=29./3780, + 3.#A/706. ~ (1a/2. + 330#8/16, 6 O14882/204 Te tesassize) 7 8 2520)026314 /6048. + 1763.44/665280. + (Le/2e + 6488/7. Ls 15.42/28, + 5,8a3/4241 / 8 26-026 D3=cL ca1=031 D328ca: 32032 £33003: 033=1-/2925. + T1-*A/300300. + (8.735. + 18+¥A/35) Ls 15éc0n2 7 305. 4 6.843/55.1 7 8 116622006 + (14/3! T+ [1.8h2/10, + 3.883/10+1 7 8 6342034 - D35=19T, 415800, + 103.*A/3392040, = 1s/35. rb.#A7ISS Lt 11.82/70. ¢ 03/1001 7 8 35-035, 0362233, T+ A2/ia. + 2.43/105-) 7 8 c36=036 DETL = DET(DILs012, DI3+D21 v022y023;031 032,033) |___o1i=cre aary oai=cae DALAAO = DETIDII,O12,D13 sD 210224023 031,032,033) piiscis i ozi=ces, + ostasa5e A135. ALAAL = DETIDII,D12,D130 23,031,03z,033) 1 DLL=c1s. bat=cz6 DaL=c36 DALGEO = DETLOIL,D12+013 +0 21s 0224023 021403200339 pyrers ozt=C2 piz=ci4 D22=c24 032=C34 DAZAAO = DET(O) p13 0; 3210331, D1z=c15 bea=ca! Daz=c35 DAZAA] = DET (011,012,013, 021,022,023, 0310324033; Dis=Clé Dazeca 032=C36 A288 = DET 3 6: 9240331 Die=cl2 D3z=C32 bia=c1 b23=C24 033503 DASAAO = DET(OI1,D12,013y 021,022,023, 031,032,033) |____piaacypD23=c25, D33=c3: OASAAL O1scu 023=C26 ‘DET (011,012,013, 021,022,023, 03100329033) ‘DA3BB0 = DET(CILyD1z,013, 021,022, 023;031 032,033) AIAUSDALAAD/OETL ALALSDALAAL /OFTL 180=DALBBO/OETL ‘A2AU=DAZAAO/OETI AZA1SDAZAM1/0ET1_ ‘A280=D42080/0ET ‘ASAL=DA3AAL/OETL 4380-043980/DE11 ~ T ABOVE COMPLETES DETE®RINATION OF AlvAZy END A3 IN TERNS OF AAT, AAO,560 CULO =(8/6.189.25 5.712. = ALAO + AZAD — (8/4. 18*,25*RIAO Us7T2. = ATGO + AZSO = (074.149 2548150 D2l= =(87%- 188.5 cai=paL “(97a 18.75 31-031 O32= 3.8872. + (87as #8, TS*ATRO 032 O33=1./2+ + (874.088, 7S*ALEO 332033, ETL — = DETIDIT C1Z,013 071 ,D22+023 031,032,033) AL = 1/2. + ALA ="azal + (6/4.089125 4 ALAL Az = Oe 43 > = "16/e.)ee,75 * ALA Ol = AL pai = O34 = AS DCS = DET 011 401240134021 y022 023 »031+0321033) Olt = cit Del = cai O31 = C31 O12 = AL Dez = a2 032 = 83 ‘DAD = DET{OIN,D12,013 sD21yD22,023s031,032 0337 D2 = ch _ Dee = t2e bap = c32 DLs = AL 23 = a2 33 = 83 980 = DET( 01140124013 02149221023 403110324033),DES = BCS7DETT DAD_= DAO/DETL _ ‘Dau = DBU/DETL PRINT 5; 0C5, DA0, 080 S FORMAT {1PSE15.5/77 ALS ALAGeOAG + AIAL ¢ a1nos080 2 = SeKOSDAD + AZAT + AZSO*OBO A3 = S340eD40 + 4341 + A3¥O*DBO GRETA = SAL =3.¥A2/2, ~ 83/5. + OAO/4. © I.7h. 6 DBO IZ. BL = (1a/2eth/ee Al © (La /astl1e#A/84 21842 ¢ (12/70. +A7105.)983 TTUT./120-48/26.+3 + FA7B19OA0 ~ (1-/40+ # A/T2e) ~ (16/005 # 2 A/120.—# 1. /5)4080 Fos sPL/UB/2eT21¢8.25 * (leh e93 7 BY Vo THETAY Uize asaya. 25 $ Cieenyeesy FS = 0AO 7 TietAy C_LCOSE Hus FLANGE; AA DBO=0, DAg=DbO. ‘AL = ALAOSDAO © ATAT + AIEO*OGD AZ = AZAGe0AG + Azal + aeBO-000 AS -= NGAOWDAD + AGAL + ASBORDBO THETA = Al =.442/2. — 03/5. # OAC/4, + 1a/4, ¢ DBO/12e = 980 +9 PLE Tis 72st /ocT#AL + (Le/4eti1-A/84.082 + (10/70. #A/ 105.1903 10111200 44/36.43 A/D ISORT ~ Uhe/4O. 6 A/T2I = 16/600 + 2 R/ TaD + 1.818036 cL =Pis(ia/2013)¥e.25 + ULeetyes3 / pp Vis THETA / (2s 73/a)44.25" (1-40)893) PRINT 3, Fy Vs FS) Fly VL S FORMAT (5H FH, IPLELO.5)5H Vey TELO.5y VAIN SMai-Fo, 1610.55 /y11m F-SuB-L=y LeLo.5s Pia V-SUb=L=y 1E10+5+ 7/7) 60_t0 50 TOO CALL EXIT ENDAPPENDIX II BOLT TORQUE CHART ‘The torque values above are suggested only duet the many variables that affect actual results, Thess loads oigraphile lubricant end the thread forms shown, Efficiency of non: lubrieated bolts may drop 50%. More accurate bolt tension is cal culated a9 follows: Divide total elongation by the bal length, and ‘multiply the result by E, the modulus of elatity. Bolt length is taken ‘mig-out 10 mid-nut, or an equivalent distance, Careful micrometer rmessurements are required re based on the use of heavyGretel ot)Res MODERN FLANGE DESIGN Bulletin 502