Proceedings of PVP2002 2002 ASME Pressure Vessels and Piping Conference August 5-9, 2002, Vancouver, BC, Canada BOLT PRELOAD CONTROL FOR BOLTED FLANGE JOINT PVP2002-1094 Hirokazu TSUJI Department of Intelligent Mechanical Engineering, Tokyo Denki University Saitama, Japan Makoto NAKANO Department of Intelligent Mechanical Engineering, Tokyo Denki University ABSTRACT Tightening method for flange joints proposed by ASME PCC--1 specifies that bolts are tightened by the cross-pattern sequence and tightening torque is raised with several steps. On the other hand, new tightening method has been proposed by Japan BFC committee, in which bolts are tightened in the clockwise-pattern sequence and tightening torque is 100% of the target torque in all steps after an install step with some snug torque. Tightening tests using flanges with various nominal sizes, performed in this study, show that the new tightening method achieves comparable uniformity in bolt preloads and in flange alignment with ASME PCC-4 procedure. This new tightening method is able to reduce both the work volume in the tightening operation and the possibility of human errors like missing the tightening order in it. Stepqike increment of the bolt tension under the repeated tightening with small increment of the tightening torque is also discussed. Experimental results show that continuous control of the bolt tension by the repeated tightening is impossible. The step-like increment of the bolt tension is observed in the tightening process of the actual bolted flange joints so that the excessive iteration of the repeated tightening round hardly improves the uniformity of the bolt preloads. 1. INTRODUCTION In the assembly of a bolted flange joint with a gasket, uniform bolt preloads and precise flange alignment are important to achieve the uniform distributions of gasket stress and leak-free service of the joint as well. Tightening method for flange joints has been proposed as ASME PCCA, Guidelines for Pressure Boundary Bolted Flange Joint Assembly, which specifies that bolts are tightened in the cross-pattern tightening sequence and tightening torque is raised in several steps. This tightening procedure is complicated, so that the joint assemblers are forced to pay much attention and spend considerable amount of working hours. This leads to further problems such as the increase of the cost and human errors. Japan BFC committee has proposed a new tightening method aiming at the development of a simple and effective tightening procedure to meet the requirement of practical assembly fields. In the proposed method, bolts are tightened in the clockwise-pattern sequence and tightening torque is 100% of the target torque in all steps subsequent to an install step (CW method). This proposal is opposite to a common belief that the cross-pattern tightening sequence in conjunction with torque increment steps is needed in order to achieve the uniform bolt preloads and the flange alignment and to avoid the effects of elastic interaction caused by the tightening process. Firstly, advantages of the CW method with respect to the uniformity of the bolt preloads and the simplicity of the tightening procedure are demonstrated. From the tightening tests of the various types of flanges to compare two tightening methods, it is shown that the uniformity of the bolt preloads and the flange alignment are comparable, while the number of total steps and the time required can be reduced when using the CW method. Next, the behavior of the step-like increment in the bolt tension under the repeated tightening of the flange joint is discussed. The repeated tightening operation is generally employed to achieve uniform bolt tension, i.e., Round 4 and Round 5 of the PCC--1 procedure. Experimental results indicate that excessive iterations of the repeated tightening round hardly improve the uniformity of the bolt preloads. 2. TIGHTENING METHOD FOR BOLTED FLANGE JOINTS 2.1 ASME PCC-1, Guidelines for Pressure Boundary Bolted Flange Joint Assembly The flange joint assembly guideline, ASME PCC--1, describes the bolt tightening procedure to provide the flange joint with leak--free service. Tables I and 2 show the tightening procedure and the tightening sequence given by ASME PCC--1, respectively. Figure 1 shows the bolt numbering for flange joints tested using 4 bolts and 8 bolts. 1 Copyright 2002 by ASME
Table 1 Step Install Round i Round 2 Round 3 Round 4 Round 5 Torque increment rounds specified in ASME PCC-1. I Loading Hand tighten, then "snug up" to 15-30 Nm (not to exceed 20% of Target Torque). Tighten to 20-30% of Target Torque. Tig.~hten to 50-70% of Target Torque. Tighten to 100% of Target Torque. Continue tightening the bolts, but on a rotational clockwise pattern until no further nut rotation occurs at the Round 3 Target Torque value. Time permitting, wait a minimum of four hours and repeat Round 4; this will restore the short-term creep relaxation]embedment losses. Table 3 Clockwise tightening method (CW method). Step [ Loading Hand tighten, then "snug up" to 15-30Nm (not to exceed Install 20% of Target Torque). Round 1 Round 2 Round 3 Continue tightening the bolts, but on a rotational Ro~d4 clockwise pattern until no further nut rotation occurs at Target Torque. Time permitting, wait a minimum of four hours and Round 5 repeat Round 4; this will restore the short-term creep relaxatinn/embedment oisses. Table 2 Examples of cross-pattern tightening sequence. Number of bolts 8 24 [ Sequence 1-3 -5-7~ 2-6-4-8 1--11-749, 4--16--10-22, 2--14-8-20, 5-17--11-23, 3-15-9-21, 6-18-12-24 2.2 New Tightening Method, CW method Plans to solve the above problems in relation to tightening method ASME PCC-1 are considered as follows: one is to decrease the tightening steps to reduce the working hours and the other is to employ a simple and easy procedure to avoid the human errors. Japan BFC committee has proposed a new tightening method (Clockwise tightening method: CW method), with the aim of a simple and effective procedure that achieves uniform bolt preloads and precise flange alignment. Table 3 shows the tightening procedure of CW method. The bolts are tightened to 100% of the target torque in all steps to decrease the number of tightening steps, except for the install step, in which the tightening toque is 10% of the target torque to prevent the flange misalignment by snugging and tentative tightening. In all steps, the clockwise-pattern sequence tightening is employed with the aim of a simple procedure to avoid the human errors. Aging for a minimum of four hours is needed between Round 4 and Round 5 as same as ASME PCC--1 procedure, which compensates the decrease in bolt tension resulting from the stress-relaxation of the gasket. 8 bolts 24 bolts Fig.1 Numbering of bolts on flange. Bolts are tightened by the cross-pattern sequence and tightening torque is raised for several steps from Round 1 to Round 3. In Round 4 and Round 5, clockwise-pattern sequence tightening of bolts is applied and in these tightening rounds tightening operations are repeated until no further nut rotation occurs. Aging for a minimum of four hours is needed between Round 4 and Round 5, which compensates the decrease of bolt tension by stress-relaxation of the gasket. The above complicated procedure has been arranged in order to avoid the effects of elastic interaction occurring in the tightening process. In other words, the tightening procedure intends to prevent the flange misalignment and to achieve the uniform bolt preloads. However, the tightening method involves several problems such as the increase in tightening hours, cost, and some possibility of the human errors. 3. TIGHTENING TEST OF BOLTED FLANGE JOINT WITH GASKET 3.1 Tightening Test of 4 inch Flange Joint 3.1.1 Setup for Tightening Test and Test Joint Figure 2 shows a setup of tightening test for 4inch bolted flange joint with gasket. Bolt tension is controlled by the torque control method where the tightening torque is controlled by a manual torque wrench. Axial tension of each bolt is measured using the strain gages at the bolt shank and the flange displacements are measured using four clip-gage type displacement transducers placed at the outer perimeter of the flange. Measured data are recorded by a personal computer through a digital multi-meter. Test flange is the 4B class 1501b (material: SFVC2A), raised face slip-on welding type flange specified in JPI (Japan Petroleum Industry). Test bolt (stud) is nominal diameter of M16 with nuts (bolt material: SNB7, nut material: $45C). Two types of gaskets shown in Fig. 3 are used for tightening test. The gasket shown in Fig. 3 (a) is a spiral wound gasket (SWG) made of non-asbestos filler with an outer ring made of stainless steel SUS304 (No.591, Nippon Valqua Co.). The gasket shown in Fig. 3 (b) is a compressed asbestos fiber sheet gasket (No.1500, Nippon Valqua Co.). 2 Copyright 2002 by ASME
I ~ @ : Amplifier 40 l ASME' method' ' ' L Test Gasket : Nonasbestos SWG 30F e:no.10:no.5 ^ A:No.2 A:No.6 20 0 1 2 3 4 5 Round No. Fig.4 Tightening test result of bolt tension obtained by ASME method (4 inch flange with SWG). (~ Torque wrench ~) Strain gage (~) Displacement transducer (~) JPI 4inch flange (~) Clamp lock ~1) Contact peaces (~ Double nut bolt (~) Aligner ~) Digital multi-meter @ Hexagon nut ~) Bearing (~) Personal computer Fig.2 Setup for tightening test of bolted flange joint (4 inch flange). Filler(Nonasbestos) t.... ~...~Hoop(SUS304) l L'127 ~ 149.4 ~ 174.8 (a) Non--asbestos SWG with outer ring. tl Asbestos fibers ~ 132.~ ~b 173 (b) Compressed asbestos fibers joint sheet gasket. Fig.3 Dimensions of gaskets used in tightening test (4 inch flange). 3.1.2 Test Conditions Lubricant used in the tests is dry coating type spray of MoS2 (molybdenum disulfide) used in many plants Tightening test procedures follow ASME PCCq. method and CW method. Figure 2 shows the bolt numbering and the measuring points of the flange displacements. Target bolt preload is 35kN which is calculated based on the gasket stress which becomes 49 MPa for the SWG 3.1.3 Results and Discussion Bolt Tension Figures 4 and 5 show the relationships between the bolt tension and the tightening step using the non--asbestos SWG following ASME method and CW method, respectively. Round 0 means Install step. 30 20 11"11-11 ~ / &:No,2 A:No,6- ~1 II/// I: No.3 El: No.7 ~///jl / V: No 4 V: No.8 Test Gasket : Nonasbestos SWG 0 1 2 3 4 5 6 7 8 9 Number of cycles Fig.5 Tightening test result of bolt tension obtained by CW method (4 inch flange with SWG). In the results of the SWG by ASME method, 8 bolts form two groups in their tension through Round 1 to Round 3. This phenomenon is caused by the elastic interaction between adjacent bolts, generally observed under the cross-pattern sequence tightening [2]. In the CW method using the SWG, the effects of the elastic interaction are reduced by the clockwise-pattern sequence tightening so that the scatter range of the bolt tension is small; therefore, the bolt tensions become uniform in the small number of the tightening. Decrease in the bolt tension is observed in the number eight and nine of the tightening. Under aging of four hours between Round 4 and Round 5, corresponding to the number nine of tightening, bolt tension is decreased due to the stress-relaxation in the gasket. Results of the bolt tension using the compressed asbestos fiber joint sheet gasket are not shown in this paper; however characteristics of the bolt tension indicate almost the same tendencies, except that the effect of the elastic interaction caused by ASME method is fairly smaller than that in the case of the SWG. The dispersion of final bolt tension is 6% in ASME method and 10% in CW method. The number of tightening steps is 16 in ASME method and 10 in CW method. Thus CW method is able to achieve a comparable uniformity of the bolt preloads with the decreased steps of tightening, compared with ASME method. 3 Copyright 2002 by ASME
I I I I ASME method 6.5 ~--<~---: Amplifier 6 5.5 Test gasket : Nonasbestos SWG 50 ~ I, I ~ I, I, 1 2 3 4 Round No. % Fig.6 Tightening test result of flange displacement obtained by ASME method (4 inch flange with SWG). 8: 8 i, i, i, i i, I ' I ' ~ ' CW method 7.5 ~ Test gasket : Nonasbestos SWG ",k\\ *:No1 7 ~,\\\ A:No.2",x%N\\\~\N,\N\\\\~",,~ ~\\\\\\\\\\\\\\\\NN '~~ 6" 6.5 5.5 50'1'½ ;'4' ;';'7'8'9 Number of cycles Fig.7 Tightening test result of flange displacement obtained by CW method (4 inch flange with SWG). Flanoe Displacement Figures 6 and 7 show the relationships between the number of tightening steps and the flange displacements obtained by ASME method and CW method, respectively. Test gaskets are non-asbestos SWG for both tightening methods. Round 0 means Install step. In the both eases of ASME and CW methods, the values of the flange displacement are scattered in certain range. In the case of CW method, the flange displacement, that is, compression strain of the gasket becomes large around the starting point of the bolt tightening sequence. This phenomena reflects the flange misalignment (non-uniformity of flange gap) resulting from the partial tightening of the bolts; however, for both ASME method and CW method, the scatters of the flange displacement are comparable and satisfy the alignment tolerance specified in ASME PCC--1 as within lmm in 200mm. 3.2 Tightening Test for 20 inch Flange Joint 3.2,1 Setup for Tightening Test and Test Joint Figure 8 shows a setup of tightening test for 20 inch bolted flange joint with gasket. Tightening torque is controlled by the manual torque (~ Torque wrench (~) Double nut bolt ~) Contact peaces ~) Torque multiplier ~) Hexagon nut (~ Potentiometer (~) Reaction arm (~) Strain gage ~) Digital multi-meter (~) JP120inch flange ~) Displacement transducer (~ Personal computer Fig.8 Setup for tightening test of bolted flange joint with gasket (20 inch flange). Filler(Nonasbestos) I L, 5Z5.5, ~b510 ~ ~b 577.8 ~ 651 654 (a) Nonasbestos SWG with outer ring (b) Compressed asbestos fibers joint sheet gasket Fig.9 Dimensions of gaskets used in tightening test (20 inch flange). wrench with a torque multiplier attachment by the ratio of 10:1. Axial tension of bolts are measured using strain gages mounted on the bolt shank, and the flange displacements are measured using each four displacement transducers placed at the outer and inner perimeter of the flange, respectively. Flange displacement at the middle perimeter is measured by four potentiometers and is used to calculate the flange rotation. Measured data are recorded by the personal computer through the digital multi-meter. Test flange is the 20B class 300 lb, raised face slip-on welding 4 Copyright 2002 by ASME
type flanges specified in JPI (material: SFVC2A). Test bolts (studs) are 1¼ 8UNC with nuts (bolt material: SNB7, nut material: $45C). Test gaskets are shown in Fig. 9; (a) is a spiral wound gasket (SWG) made of non-asbestos filler with an outer ring made of the stainless steel SUS 304 (No.591, Nippon Valqua Co.); (b) is a compressed asbestos fiber sheet gasket (No.1500, Nippon Valqua Co.). 3.2.2 Test C o n d i t i o n s Lubricant used in the tests is dry coating type spray of MoS2. Tightening test method follows ASME method or CW method. Figure 8 shows the numbering of the bolts and the measuring point of the flange displacements and the flange rotation. Target bolt preload is determined to 220kN based on the gasket stress which becomes 124 MPa for the SWG. 3.2.3 Results and Discussion Bolt Ter)sion Figures 10 and 11 show the relationships between the bolt tension and the tightening step following ASME method and CW method, respectively. The gasket is the non--asbestos SWG. Round 0 means Install step. Test results for the compressed asbestos fiber joint sheet gasket obtained by ASME method and CW method are omitted 7~ 300~ - 250! o 20c ~ 15C 00! ~' 5ff ' i i r, ASME method Test gasket: Nonaabestos SWG I ~ 3 4 Round No. -~-: No.1 --W.- : No.13 -~)-: No.2 -V- : No.14 -O-: No3 --~t- : No.15 -O-: No.4 -~-: No.16 --A--: No.5 -'O-: No.17 ',-~-: No.6 -~)-:No.18-4k--: No.7.-~ : No.19 -~.-: No.8 "O-: No.20 "B": No.9 --~-: No.21 -R-: No.10 -O- : No.22 41-: No.ll "r~.: No.23 -I.: No.12 -~-';-: No.24 5 Fig.10 Tightening test result of bolt tension obtained by ASME method (20 inch flange with SWG). 300 _~ 250 O 200.~ 150 100 ' " ' ' ~ ' ' ' ' ' ' ' ' " "O":No.1 CW method -1 -O-: No.2 ~ "O-: No.3 -~lt-: -O" : No.4 : No.5 -O-: -~r: No.6 : No.7-41-:.d~-: No.8 "O': "-13-: No.9 -r:l-: No.10 -n-: No.ll 1 2 3 4 5 6 7 8.m.: No.12 Number of cycles Fig.ll Tightening test result of bolt tension obtained by CW method (20 inch flange with SWG). No.13 No.14 No.15 No.16 No.17 No.18 No.19 No.20 No.21 No.22 No.23 No.24 In ASME method, 24 bolts form three groups in their tension through Round 1 to Round 3. This phenomenon is caused by the elastic interaction between adjacent bolts due to the cross-pattern sequence. In CW method, since the effects of the elastic interaction on the bolt tension is small, the uniform bolt tension may be achieved in reduced tightening steps as well as the case of the 4 inch flange with 8 bolts described above. The same tendency is observed in the compressed asbestos fiber joint sheet gasket. The dispersion of final bolt tension is 20% in ASME method and 19% in CW method. The number of the tightening steps is 11 in ASME method and 9 in CW method. Thus CW method is able to achieve a comparable uniformity of the bolt preloads with the decreased the tightening steps comparing with ASME method. Flanqe Displacement Figures 12 and 13 show the relations between the number of tightening steps and the flange displacement obtained by ASME method and CW method, respectively. The test gaskets are non-asbestos SWG. Values of the flange displacement are measured at four points around the outer and inner perimeter of the flange. Round 0 means Install step. Test results using compressed asbestos fiber joint sheet gaskets are omitted. 8 i i i ' i ASME method Test gasket: Nonasbestos SWG - ~ O: No.1 O: No.2 k: No.3 /k: No.4 I: No.5 []: No.6 - Outer diameter side ~ V:No.7 V:No.8 Inner diameter side 3 20 ' I ~ I h I, I h 1 2 3 4 5 Round No. Fig. 12 Tightening test result of flange displacement obtained by ASME method (20 inch flange with SWG). ~3 8 20 i l i i i i i CW method T e s t gasket: Nonasbestos SWG - Inner diameter side O: No.1 O: No.2 &:No.3 A:No.4- I I I I I I I 1 2 3 4 5 6 7 8 Number of cycles Fig.13 Tightening test result of flange displacement obtained by CW method (20 inch flange with SWG). 5 Copyright 2002 by ASME
Torsional compliance of tool Friction at bearing surfaces / of nut: Bearing surface torque / Torsional compliance of bolt~ / r, tiona, tto o i \ of thread: Thread torque x,,~, ~ ~ Fe o o.o N Fig.14 "/~ ~ "7 r/- (Axial tension of bolt) Axial tension of bolt Model of friction for bolted joint. Friction condition at fitted portion of thread Static friction Kinetic friction /' // / / / Nut turn angle Fig.15 Relation between nut turn angle and axial tension of bolt under static/kinetic friction. In both ASME method and CW method, the flange alignment measured at the outer perimeter achieves within the alignment tolerance specified in the guideline ASME PCC--1, therefore, CW method achieves comparable uniformity of the flange alignment to ASME method. The compression of the gasket measured at the inner diameter of the flange tends to be decreased in both ASME method and CW method, as the number of the tightening step increases. The decrease of gasket compression results from the flange rotation followed by the increase of the bolt tension [3] 14]. The increase of bolt tension also causes the partial separation of the gasket contact around the inner diameter of flange where the joint is assembled with the joint sheet gasket following both ASME method and CW method. 4. BEHAVIOR OF BOLT TENSION UNDER REPEATED TIGHTENING 4.1 Effect of Static/Kinetic Friction Coefficient on Tightening of Flange Joint The torque control method has been commonly used for the bolt preload control at the flange joint assembly. The torque control // / (~) Torque wrench (~) Bolt specimen (~ Digital multi-meter (~) Torque multiplier ~) Hexagon nut ~ Personal computer (~) Reaction arm (~ Bolt-tension meter Extension bar Strain amplifier Fig.16 Setup for tightening test of bolt. method has some problems such that the bolt tension varies due to the fluctuation of the friction coefficient at the fitted portion of the thread and the bearing surfaces of the nut. In addition, in the case of the flange joint assembly, bolts are tightened repeatedly with small increment of the tightening torque; accordingly the variation in the bolt tension may be increased by the difference between the static and the kinetic friction coefficients. In this paper, the tightening tests using a set of bolt and nut were done to account for the variation of the bolt tension due to the transition of the friction coefficient between static and kinetic. 4.2 Relation between Bolt Tension and Static/Kinetic Friction Coefficient It is commonly known that the most important factor to control the bolt preload by the tightening torque is the friction coefficient between the threads and that between the bearing surfaces. Increment of the bolt tension induced by small increment of the tightening torque may be greatly affected by the difference between the static/kinetic frictional conditions at the contact surfaces of bolt and nut. Figure 14 shows a spring model for the bolted joint in order to explain the increment of the bolt tension during the tightening process under the static/kinetic frictional condition. Springs simulate the tensile compliance and torsion compliance of the bolt, and a rotational motion along the helix of the screw thread is represented by a linear motion on the slope. Slip motions at the fitted portion of the thread and at the bearing surfaces during the tightening process are considered separately according to each their frictional conditions. For the kinetic friction at the bearing surface but the static friction at the thread portion, no increment of the bolt tension will occur in contrast to the nut rotating. The increment of bolt tension will occur when the kinetic friction conditions are satisfied both at the thread portion and at the bearing surface. The tightening torque required for the kinetic friction condition is less than for the static friction condition, resulting from the change of the friction coefficient. The repetition of this mechanism causes the sporadic increment of the bolt tension like a staircase shown in Fig.15. 6 Copyright 2002 by ASME
Z 600 500 ~r 400.~ 300 200 100 ' I I ' I 0: Type 1 7 6 12 18 Test steps 24 30 200 150 100 < 50... ' ' I ' I I -- : Axial tension Type 1... : Tightening torque....'thread torque,, p; h ~" I j'l l I t +~ I i +) II I II ii II i I I I, I,+P i I.. +. U I +. i I ~? III I ii i I II I I I I I I t I ii i ~ ~ i i J! Ii ~ IIII 1111 11 ilfll I I I i I I I II It i I I I I ii I I I I I ii I i1+1 i% 14/I I l!jjl ii I / I ii I I II II I I I II I '1 I II II I ;li,,,,,,..., I II I I I1 t i I I +I II ii, I ' 11 II, I I I ' il t + i, I I I I I I I Ii IIII ii I I I t Ii II I I I I I I I I ' i I {j J ~ I ' ' l I, ';,J ', ',' ' ' ' J It i, +I/I,.. II jll'.dl. 50 100 150 2d8 Time sec 700 600 5oo ~, 400 ~ o +oo++; 300 "" +" 100 Fig.17 Two types of test procedures for repeated tightening. Fig.18 Result of repeated tightening test obtained by Type 1 procedure 4.3 Testing Method for Repeated Tightening Figure 16 shows the setup of tightening test for a set of bolt and nut. Tightening torque Tf was measured by the strain gages on the extension bar, and axial tension of bolt Fr and thread torque Ts are measured by the strain gage at the shank of the bolt Bearing surface torque Tw is calculated by the equation of T w = Tf - T~. The specimen is a stud with nuts of lr,,~ 8 UNC (bolt material: SNB7, nut material: $45C). Prior to the tightening test, the stud and the nuts were lubricated using MoS2 dry coating. Two types of repeated tightening procedures are tested, namely, Type 1 and Type 2, shown in Fig. 17. Type 1 procedure, in which the tightening torque is gradually increased with the repetition of tightening, simulates the tightening process appeared in the flange assembly. 4.4 Test Results and Considerations 4.4.1 Relation between Tightening Torque Tr and Bolt Tension Fr Figures 18 shows the test results of Type 1. It is well known that the torque control method is used assuming the linear relation between Tf and Ff. In the case of Type 1 procedure, the value of Ff increases sporadically like a staircase with increase of the value of Tf. This results from the transition of static]kinetic frictional condition at the fitted portion of the thread, that is, the values of Ff keep constant where Tf does not overcome the static frictional torque of the thread portion. There is a possibility that the bearing surface is also under the static condition. 4.4.2 Changes In Bolt Tension and Torque in Type 1 Procedure Figure 19 shows the details of the test results shown in Fig. 18. The value of Fr increases with an increase of the tightening torque Tf and the thread torque Ts decreases at 118.8 second. This behavior is a transition from the static frictional condition to the kinetic one at the engagement thread portion. 150 140 ~ 130 ~ 120 110 10~ 1'8,,,, 700 60O ~z a ~ /"'... "i, 500 {+,,, +oo + / AFI i _... 4-7...... 300 - - :Axial tension... :Tightening torque! 100...,Thread torque \, 119,, 120 I, 121 ], 122, ' -2 nlx Time sec Fig.19 Change of bolt tension and torques under repeated tightening of Type i procedure (Detail of Fig. 18). Mechanism of this behavior is considered as follows: (a) the thread portion and the bearing surface are both under the static friction for the low Tf; (b) the bearing surface becomes kinetic friction with an increase of Tf while the bearing surface remains under the static friction, and torsional strain increases in the bolt shank; (c) with further increase of Tf, the thread portion becomes kinetic friction as well as the bearing surface; (d) a slippage at the fitted portion of the thread, i.e., the substantial tightening occurs, consequently the bolt tension Ff increases, and the torsion of the bolt shank Ts is released due to the lower friction coefficient under the kinetic friction. Above sequence is repeated under the tightening, which results in the sporadic increase of Ft. The bolt tension increment is dominated by this mechanism, especially under the repeated tightening with small increment of the tightening torque It should be noted that the resolution of the bolt tension by the torque control method is limited by this mechanism in spite of the careful torque control 7 Copyright 2002 by ASME
~ 12 ' I I I -: Calculation.0: Tightenh~g test / A/~,=0.006 8 ~a 6 ~o 4 0.~,u,= 0.005~ A/z, = 0.003. A ~,= 0.002' 100 200 300 400 Thread torque Nm Fig.20 Effect of friction coefficient change at thread portion on increment of bolt tension. 4.4.3 Increment of Bolt Tension The relation between the bolt tension Fr and the thread torque Ts under tightening is expressed as, where d2 is a pitch diameter of the thread (;29.393 ram) and Pis a pitch of the thread. Upon knowing the change in the friction coefficient at the thread portion A/IB =/1 s s-/.is K, the increment of bolt tension AFt is estimated by the approximate equation: a~ (2) The value Aps = 0.03 is obtained experimentally by the change of thread torque A Ts before and after the transition of the static/kinetic friction (Fig. 20). Figure 21 shows the relation between the increment of the bolt tension and the change of friction coefficient. The solid lines are calculated results for various Aps, and the experimental result is also plotted. The experimental result is close to the calculated line for A/~, = 0.03. It is shown that the increment of the bolt tension is estimated by the above equation. 4.4.4 Increment of Bolt Tension in Flange Joint by Repeated Tightening The tightening steps of Round 4 and 5 according to ASME PCC--1 employ the repeated tightening of the flange bolts until no further nut rotation occurs. Although the repeated tightening aims to achieve uniform tensions of the bolts, the continuous control of bolt tension is impossible due to the sporadic increment of tension, as discussed in the previous section. The bolt tension dispersion of a flange joint resulting from the minimum bolt tension increment is estimated to be almost 5% of the target preload, except for the effects of the elastic interaction on the bolt tension It is rather reasonable to provide an upper limit for the number of tightening repetition, taking account the efforts and time required for assembling the joints. To determine the upper limit, further experimental and theoretical analyses are required to take the influences of the flange types and the gasket types into account. 5. EFFECTIVENESS AND FUTURE ABOUT CLOCKWISE TIGHTENING METHOD Test results for the flanges of various nominal sizes show that uniformity of bolt preloads and flange alignment is comparable to that of ASME procedure; moreover tightening steps and the human error may be reduced. CW method is more practical method with effectiveness and simplicity. The future works to be solved are as follows: - To set the upper limit of the number of tightening repetition in Round 4 and Round 5, since the excessive repetition does not result in the improvement of the uniformity of the bolt preloads. - To demonstrate the validity of the proposed tightening method including the leak test of the assembled joint. - To establish a new tightening method to control the gasket strain which dominates the sealing performance of the gasket [5]. 6. CONCLUSIONS The advantages of the CW method proposed by Japan BFC committee as a counterproposal to ASME PCC--1 are demonstrated. These tightening methods are compared by conducting the tightening tests with the various types of flanges and gaskets. The results obtained are summarized as follows: 1. Clockwise tightening method (CW method) achieves a comparable uniformity of the bolt preloads, while the total tightening steps/rounds required for joint assembling is decreased, compared with ASME method. 2. CW method reduces both the cost and the human errors to avoid the complicated specification on the torque increment rounds and the tightening sequence, so that the method is practical with effectiveness and simplicity. 3. The mechanism of sporadic increment in the bolt tension like a staircase under the repeated tightening with small increment of tightening torque is explained by the transition of static/kinetic frictional coefficient at the fitted portion of the thread and the bearing surfaces of the nut. 4. The continuous control of bolt tension under the repeated tightening of flange is impossible due to the sporadic increment of bolt tension. It is reasonable to provide an upper limit of the number of the tightening repetition that aims to the uniformity of the bolt preloads. REFERENCES [1] ASME, 2000, Guidelines for Pressure Boundary Bolted Flange joint Assembly, ASME PCC-1. [2] Bibel, G D. and Ezell, R. M., 1996, Bolted Flange Assembly: Preliminary Elastic Interaction Data and Improved Bolt-up Procedures, Welding Research Council Bulletin, 408, pp.1--27. [3] Bibel, A. and Derenne, M., 1997, Distribution of the Gasket Contact Stress in Bolted Flange Connections, ASME PVP--Vol.354, pp.185--193. [4] Bouzid, A. and Chaaban, A. and Bazergui, A., 1995, The Effect of Gasket Creep-Relaxation on the Leakage Tightness of Bolted Flanged Joints, Transaction of ASME, Journal of Pressure Vessel Technology, 117, pp.71-77. [5] Tsuji, H. and Fujihara, S., 2001, Sealing Performance Evaluation Using Gasket Strain under RoTr/HOTr, ASME PVP--Vol.416, pp.83-89. 8 Copyright 2002 by ASME