You are on page 1of 17

Baragetti pp414-430 22/4/03 12:31 pm Page 414

414 Int. J. of Materials & Product Technology, Vol. 19, No. 5, 2003

111
2 Friction and tightening force of conical threaded
3 connections: experiments in various conditions
4
5
6 S. Baragetti*, P. Clerici†, S. Matteazzi‡
7
8 *Dip.to Ingegneria, Università Degli Studi di Bergamo, Viale Marconi
9 5, 24044 Dalmine (BG), Italy
10 †
Dip.to Meccanica, Politecnico di Milano, P.zza Leonardo da Vinci 32,
1 20133 Milan, Italy
2 ‡
Dip.to Ingegneria Meccanica, Università Degli Studi di Brescia, Viale
3 Branze 38, 25123, Brescia, Italy
4 Corresponding author: sergio.baragetti@mecc.polimi.it
5
6 Abstract: In this study the friction behaviour of conical threaded connections
7 was investigated in various conditions by performing full scale make up tests. A
new procedure for friction condition evaluation in threaded joints is proposed.
8 Several make up/break out (M/B) tests were carried out and the behaviour of the
9 threaded connections at different make up speeds was investigated; the results
2011 were summarised and organised through the use of probability charts. All the
1 results of the tests provide the user with a procedure, based on a large data spec-
2 trum, that enables the evaluation of whether or not to reject a connection before
inserting it in the perforation drillstring.
3
4 Keywords: Friction behaviour, make up speed, rotary shouldered connections
5 (RSC).
6
7 Reference to this paper should be made as follows: Baragetti, S., Clerici, P. and
8 Matteazzi, S. (2003) ‘Friction and tightening force of conical threaded connec-
tions: experiments in various conditions’, Int. J. of Materials & Product
9 Technology, Vol. 19, No. 5, pp. 414–430.
30
1
2
3
Abbreviations
4
5
ID Inside diameter
6
7 M/B Make up/break out
8 OD Outside diameter
9 rpm Rotations per minute
40
RSC Rotary shouldered connection.
1
2
3
Nomenclature
4
5
Ab Cross-sectional area of box
6
711 Ap Cross-sectional area of pin

Copyright © 2003 Inderscience Enterprises Ltd.


Baragetti pp414-430 17/4/03 1:51 pm Page 415

Friction and tightening force of conical threaded connections 415

111 E Young’s modulus of the material


2 dm Mean value of the bolt diameter
3
Dm Mean nut-plate connection area diameter
4
5 f Friction coefficient
6 Kb Box stress reduction factor under external axial loading
7 Kp Pin stress factor under external axial loading
8
Ks Contribution to the make up torque due to the shoulder contact
9
10 Kt Contribution to the make up torque due to the threads
1 Kt1 Contribution to the make up torque due to the helical effect of threads
2 Kt2 Contribution to the make up torque due to the contact among the threads
3 Kup Friction parameter, determining frictional torque due to RSC make up
4
5 Kout Friction parameter, determining frictional torque due to RSC break out
6 M Per unit length moment load
7 M* Moment for which the torque-turns diagram presents an evident slope variation
8 Ms′ Torque friction contribution due to the nut-plate mating surfaces
9
2011 Ms″ Torque friction contribution due to the helical effect and friction among threads
1 Mout Maximum break out torque
2 Mup Maximum make up torque
3 Mout* Maximum break out torque obtained from extrapolation
4
5 Mup′ Make up torque in presence of an external applied load
6 Mout′ Break out torque in presence of an external applied load
7 N Per unit length axial load
8 p Thread pitch
9
Qb Amplitude of box cyclic load due to applied external tensile loads
30
1 Qp Amplitude of pin cyclic load due to applied external tensile loads.
2 Qex External applied axial load
3 Qs Shoulder load
4
Qt Thread load
5
6 Qup Axial shoulder load induced by the make up torque
7 Qs Portion of the external load acting on shoulder
8 Qt Portion of the external load acting on threads
9
Ym Yield strength of the material
40
1 Rs Shoulder mean radius
2 Rt Thread mean radius
3 t Box bore zone thickness
4
* Helical thread inclination angle
5
6 out Angle of inclination of the break out curve in the torque-turns diagram
711 up Angle of inclination of the make up curve in the torque-turns diagram
Baragetti pp414-430 17/4/03 1:51 pm Page 416

416 S. Baragetti, P. Clerici and S. Matteazzi

111  Friction angle of threads


2  Thread angle
3
bp Box–pin relative rotation
4
5 x Deformation in the pipe axial direction
6  Deformation in the pipe circumferential direction
7 z Deformation in the pipe 45° deviated direction, with respect to the pipe axis
8
N Axial load stress
9
10 M Bending stress
1 x Stress in the pipe axial direction
2  Stress in the pipe circumferential direction
3
 Poisson’s coefficient
4
5 x Shearing strain
6
7
8 1 Introduction
9
2011 Rotary shouldered connections, used in the oil and gas industry, are subjected both to the
1 make up torque and the service loads (axial, bending and torsion) and these service loads
2 can reach high values and promote either static or fatigue damage. The make up torque
3 induces a pre-load between the two elements of the connection, pin and box, that is an axial
4 compressive load in the shoulder and links the connections to each other in the drillstring.
5 This way the drillstring itself can be assimilated to a monolithic pipe having the length of
6 the whole drillstring. The compressive load induced by the applied torque, the shoulder
7 load, is an internal load for the RSC and is restricted to the portion of the connection that
8 goes from the first threads engaged to the shoulder. This load is linked to the make up
9 torque by means of a coefficient Kup that depends both on the geometrical characteristics of
30 the RSC and on the friction coefficients of the contacting surfaces: Mup=Kup
Qup. Figure 1
1 shows a sketch in which the forces acting on the shoulder and on the threads, after the make
2 up torque application, are shown. The load acting on the threads and on the shoulder is the
3 same only if no axial external compressive or tensile load is applied. Otherwise the threads,
4 due to an external applied load, should bear a higher or lower load than the shoulder.
5 API Standards [1, 2] provide the user with a simple method to quantify the axial make
6 up induced load by using a formula in which one has only to insert the friction coefficient
7 of the contacting surfaces (shoulder faces and thread flanks) and the connection and thread
8 dimensions. Downhole working conditions are taken into consideration too by API
9 Recommended Practice 7G [2] in which the working limits (shoulder separation and
40 yielding of the connection in the weakest section) of the connections are proposed both
1 in an analytical and in a graphical way. With this procedure one has only to verify that,
2 for the applied make up torque and axial tensile load, the working limits of the drillstring
3 are not exceeded in any section.
4 The influence of the compound used to grease the flanks of the threads is also con-
5 sidered by API in terms of friction coefficient [3, 6] and the API Standard 7A1 [7] (the
6 friction coefficient is probably the most important parameter to be controlled in order to
711 have a reliable equipment).
Baragetti pp414-430 17/4/03 1:51 pm Page 417

Friction and tightening force of conical threaded connections 417

111
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
2011
1
2 Figure 1 Sketch of the forces acting on the shoulder and on the threads after the make up
3 torque application.
4
5 Besides simple API equations, in a number of papers the behaviour of the threaded
6 connections by using different techniques was studied (finite element models [8–12],
7 experimental tests [8, 13], analytical models [14]) in order to calculate the distribution of
8 loads in the threads and to evaluate the connection resistance and friction capabilities.
9 Even though API Standards give a method for a nominal correct use of new threaded
30 connections, and even if it is possible to find good literature studies and experimental
1 approaches that can help in preventing failures [1–4], nevertheless it is not so simple to
2 quantify the friction capabilities of used connections and not so unusual to find rig site
3 examples of twist-off tool joints, damaged threads, washout of threads and broken pins [13].
4 Bearing these last considerations in mind and, regarding the industrial necessity to
5 reduce working costs, that is to say, stoppages due to the damaged connections or break-
6 age of the drillstring in one of its sections, the development of full scale tests carried out
7 on standard API connections could help in evaluating the influences induced by different
8 working conditions. Among the most important parameters that should be monitored
9 before and during actual drilling operations, make up speed should be carefully analysed
40 in order to quantify the friction behaviour differences in various kinds of make up condi-
1 tions.
2 The aim of this paper is to propose an operative procedure for friction capabilities
3 evaluation of RSCs by means of a simple procedure that only needs the knowledge of the
4 maximum values of the make up and break out torques (recorded in the torque/turns
5 diagram). Practical values, based on full scale tests with make up speeds in the range
6 1–10 rpm, of the parameters that characterise the friction behaviour of the API NC50
711 rotary shouldered steel connections are presented too. The work developed through
Baragetti pp414-430 17/4/03 1:51 pm Page 418

418 S. Baragetti, P. Clerici and S. Matteazzi

111 several make up/break out experimental tests carried out on API NC50 steel strain gauged
2 connections by using a typical rig site tong. The same compound was used in all the tests
3 and the make up torque values and speeds of make up of the components of the connec-
4 tion were set to three different levels (1 rpm, 5 rpm, 10 rpm). The statistical data analy-
5 sis and the definition of the procedure that permitted us to quantify Kup for different make
6 up speeds, provides the rig site user with a powerful and practical method to reject the
7 damaged NC50 connections. Kup evaluation by means of strain gauge measurements,
8 enabled us to validate the trustworthiness of the proposed procedure.
9
10
1 2 Theoretical analysis of RSCs friction resistance and background of API
2 standards
3
4 Machine design equations enable evaluation of the relationship between make up applied
5 torque and axial induced load, in the case of made-up bolts, by using the well known
6 formulation [15]:
7
8
9 Mup = Ms′ + Ms″ = Qup  f Dm dm
–––– + –––– tan (* + )
2 2  (1)
2011
1 where Ms′ and Ms″ are the two torque friction contributions, depending both on the geo-
2 metrical dimensions of the bolt and of its thread, on the friction coefficient of the con-
3 tacting threads and the nut-plate contacting surfaces.
4 The application of the make up torque to a tapered connection, can be schematised
5 with the following mathematical model [2]:
6
7 Mup = Ks
Qs + Kt
Qt (2)
8
9 Coefficients Ks and Kt, by taking into consideration that the helical thread inclination
30 angle is very small, reproduce the same terms of Eq. (1), where the contribution of the
1 nut-plate friction moment is substituted with the RSC shoulder contact friction moment:
2
3 Ks = Rs
f; Kt = Kt2 ± Kt1 (+ for make up and for break out) (3)
4
5 p Rt
f
6 Kt1 = ––– Kt2 = ––––– (4)
2 cos
7
8
Coefficient Ks summarises the shoulder contact friction effects, while Kt1 and Kt2
9
40 respectively take into consideration the contribution to the make up torque induced by the
1 helical thread angle and the friction conditions of the contacting surfaces.
2 By using Eqs (2)–(4) it is possible to obtain the API Kup value:
3
4 Kup=Ks+Kt2+Kt1 (5)
5
f p
6 Kup = f
Rs + –––
Rt + –––– (6)
711 cos 2
Baragetti pp414-430 17/4/03 1:51 pm Page 419

Friction and tightening force of conical threaded connections 419

111 This way the expression of Mup becomes:


2
3 Mup = Ks
Qs + Kt2 ± Kt1
Qt (7)
4
5 Analytical expression Eq. (7) establishes a link between make up torque and axial pre-
6 load passing through the shoulder; this relationship only depends on the geometric para-
7 meters characterising the connection and the friction coefficient f that influences the
8 behaviour in correspondence of the threads and of the shoulder. In the case of NC50 con-
9 nections the estimated API Kup value is equal to 12.49 mm (friction coefficient f = 0.08).
10 In the general case, one will have a situation in which:
1
2 Qt = Qex +Qs (8)
3
4 If there is no external load applied:
5
6 Qex = 0; Qt = Qs = Qup (9)
7
8 So, by means of a substitution into Eq. (9), respectively in the make up and break out
9 cases, it is possible to obtain the following expressions of Mup and Mout:
2011
1 Mup = Ks + Kt2 + Kt1
Qup (10)
2
3 Mout = Ks + Kt2 Kt1
Qup (11)
4
5 Or, in a different notation:
6
7 Mup = Kup
Qup (12)
8
9 Mout = Kup 2Kt1
Qup (13)
30
1 Kup can be also calculated by dividing Mup and Mout [4]:
2
3 p
––
4
Kup = ––––––– (14)
5 Mout
1 ––––
6 Mup
7
8 In the case in which Qex 0 (Qex is positive if it is a tensile load) one has to evaluate
9 terms Qt and Qs in this new situation.
40
1 Qs = Qup Qs = Qup Kb
Qex (15)
2
3 Qt = Qup + Qt = Qup + Kp
Qex (16)
4
5 where Qup is the axial shoulder load induced only by Mup, while the term Qs stands for
6 the portion of the external load acting on the shoulder and Qt is the portion of the exter-
711 nal load acting on thread. Coefficients Kp and Kb are theoretically defined as follows:
Baragetti pp414-430 17/4/03 1:51 pm Page 420

420 S. Baragetti, P. Clerici and S. Matteazzi

111 Ab
Kb = –––––– (17)
2 Ab + Ap
3
4 Ap
5 Kp = –––––– (18)
Ab + Ap
6
7 Kp + Kb = 1 (19)
8
9 This way the make up and break out applied torques, in presence of an external
10 applied load, become:
1
2 Mup′ = Kup
Qup + Kt
Kp Ks
Kb
Qex (20)
3
4 Mout′ = Kup 2Kt1
Qup + Kt
Kp Ks
Kb 2Kt1
Kp
Qex (21)
5
6 Theoretically, RSC static characteristics are linked to each other as shown in Figure 2.
7 The main points in Figure 2 are defined by the following equations:
8 Compression strength capacity of the RSC (box end) made up with make-up torque
9 equal or lower T4 [Eq. (27)]:
2011
1 Q1 = Ym
Ab (22)
2
3 Tension strength capacity of the RSC (pin end) made up with make up torque equal
4 or lower T4 [Eq. (27)]:
5
6 Q2 = Ym
Ap (23)
7
8 Torsional yield of the RSC (box end) under make up torque:
9
30 T2 = Q2
Kup (24)
1
2 Torsional yield of the RSC (pin end) under make-up torque:
3
4 T1 = Q1
Kup (25)
5
6 Over-torque if a connection was first made hand-tight, pulled to pin yield, then torque
7 applied:
8
9 T3 = Q2 Kt1 + Kt2  (26)
40
1 Make-up torque which, if applied followed by external tension being applied to fail-
2 ure, would cause a failure mode of simultaneous shoulder separation and pin tensile yield:
3
4 T4 = Q2KbKup = Q4Kup (27)
5
6 Additional axial loads on box shoulder face and on pin threads due to applied exter-
711 nal tensile loads:
Baragetti pp414-430 17/4/03 1:51 pm Page 421

Friction and tightening force of conical threaded connections 421

111
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
2011
1
2
3
4 Figure 2 Sketch of the torque/applied tension load diagram (working limits) for a New NC50
5 165  83 connection.
6
7 Qp = Kp
Qex (28)
8
9 Qb = Kb
Qex (29)
30
1 Figure 2 shows the stressed state (and also operating limits) of NC50 (165 mm
2 OD  83 mm ID) connection under combined make up torque and tension loads. The line
3 OP1 represents shoulder separation for low make up torque, the line P1Q1 represents pin
4 yield, Q1 represents pin yield due to make up, the horizontal line Q*1P1 represents max-
5 imum tension load on the pin, Q2 (on the Qup axis) represents box yield due to make up.
6 The line P2Qupr (through the point Qupr and parallel to the line P1Q1) and the horizontal
7 line through the point P2 defines box cyclic load Qb while an external tensile (static or
8 dynamic) load Qex is applied. The pin cyclic load Qp is defined by the line OP1 and the
9 horizontal line through P2. Analytically Qb and Qp values are defined by the equations
40 (28) and (29). Some design factor should be taken into account for this plot in order to
1 provide some safety margins. This safety margin depends on actual drilling conditions.
2 The behaviour of RSCs under external axial loads is defined by the pin and box stress
3 reduction factors (Kp and Kb) that can also be evaluated from Figure 2:
4
5 Q4
6 cot = ––– = Kb (30)
711 Q1
Baragetti pp414-430 17/4/03 1:51 pm Page 422

422 S. Baragetti, P. Clerici and S. Matteazzi

111 Q1 Q4 (31)
2 cor = ––––––– = Kp
Q1
3
4 Axial loads on the shoulder and along the threads define RSC sealing and tensile
5 capacity (Figure 2):
6
7 Qt = Qup + Qp < Q1 (32)
8
9 Qs = Qup Qb > 0 (33)
10
1 The values of the principal geometrical parameters characterising the NC50 connec-
2 tion torsion and tension capacity are reported in Table 1.
3 The friction coefficient of the thread flanks can be evaluated by carrying out appro-
4 priate experimental make up tests under tension or compression without shoulder contact.
5 If there is no shoulder contact (this could be done by using the load cycle load/make
6 up/break out/unload) one can write:
7
8 Mup = Qex Kt (34)
9
2011 where Kt only takes into consideration the friction effects at the mating surfaces of
1 threads:
2
3 p Rt f
4 Kt = –– + –––– (35)
2 cos
5
6
7
3 Experimental procedures
8
9
The tested connection is the API NC50 made of AISI 4145 H steel (Ultimate tensile stress
30
960 N/mm2, Yield stress 860 N/mm2, Young’s modulus E=206000 N/mm2, Poisson’s
1
coefficient =0.3). Threads were greased with a 40% zinc compound [16] with friction
2
coefficient equal to f=0.08. The compound was renewed every 5 M/Bs.
3
Four different sets of test were carried out both on strain gauged and non strain gauged
4
NC50 connections. The first set of tests, carried out on non strain gauged connections,
5
enabled the verification of the value of the friction coefficient among the mating surfaces
6
of the threads. The connections were put in tension or compression, without shoulder con-
7
tact, and then the make up torque was applied [Eqs (34) and (35) were used to calculate
8
the friction coefficient]. The second set of tests, carried out on strain gauged connections,
9
enabled the determination of the Kup value in order to have a reference experimental value
40
1
2 Table 1 Geometrical parameters characterising the NC50 connection torsion and
3 tension capacity.
4 OD ID Ap Ab Kb Kp Ks Kt2 Kt1 Kup Kout
5 mm mm (mm2) (mm2) (mm) (mm) (mm) (mm) (mm)
6
168.3 98.4 4462 8283 0.65 0.35 5.97 5.51 1.01 12.49 10.47
711
Baragetti pp414-430 17/4/03 1:51 pm Page 423

Friction and tightening force of conical threaded connections 423

111 for the other tests; in these tests the make up speed was set equal to 1 rpm. In the third
2 set of tests different make up speeds were tested in order to quantify the trustworthiness
3 of a new procedure for friction capability evaluation of RSCs. In the last set of tests NC50
4 connections were subjected to tensile or compressive loads after the application of pre-
5 scribed make up torques (the load cycle consists first in the application of the make up
6 torque and then of the axial tensile load): these tests enabled the calculation of the Kb
7 coefficient (more details about this last set of tests can be found in [17, 18]).
8
9
3.1 Testing apparatus
10
1 Equipment was designed and constructed in order to simulate the real working conditions of
2 the connections: applied make up torque, external applied tension or compression. Only the
3 device for the application of axial tensile or compressive loads was designed and construct-
4 ed while the device for the application of the make up torque was a typical rig site tong.
5 New and used NC50 steel connections were equipped with strain gauges and the Qup
6 values, the values of the shoulder axial pre-load induced by the make up torque, were
7 measured and Kup was calculated by knowing the make up torque values: Mup=KupQup.
8 These tests enabled the quantification of the effects induced by external applied loads
9 too. In particular, the application of tensile or compressive loads, without making up the
2011 connection, permitted us to verify the value of the friction coefficient between the thread
1 surfaces while the application of tensile loads gave us the value of Kb coefficient.
2 A sketch of the test equipment for make up is shown in Figure 3(b). Torque was
3 applied by using rig casing tongs able to tighten the NC50 connections with a maximum
4 torque equal to 24 kNm.
5 The drawing of the test equipment for tension during make up is shown in Figure 3(a).
6 This equipment enables application of an axial compressive or tensile load to NC50
7 full-scale connections up to 2
106 N by using a hydraulic piston that drives a shaft posi-
8 tioned inside the connection itself.
9 The torque-turns diagram was recorded by using a JAM® acquisition system while
30 strain gauges measurements were developed by using the UPM100 data logger connected
1 to and controlled by a personal computer.
2 Two different kinds of constraints were used to connect the top of the NC50 connections
3 to the shaft put inside them: in the case of tensile axial applied loads, a pin connection was
4 developed by putting a sphere between the two elements while, for compressive applied
5 loads, it was decided to use an axial bearing mounted at the top of the whole structure.
6
7
3.2 Strain gauges measurements
8
9 As shown in Figure 4 the box of the connections was machined inside and outside the
40 bore zone and strain gauges were applied inside and outside the box machined end. Only
1 by means of this modification it was possible to record the bending and axial actions
2 values. On the strain gauged equipment it was possible to calculate both Kup and the axial
3 shoulder load in various conditions and combinations of make up applied torque and axial
4 external applied load.
5 Axial, circumferential and positioned at 45°, with respect to the pipe axis, strain
6 gauges were put onto the outer bore zone surface while only axial strain gages in the inner
711 surface of the bore zone of the box element.
Baragetti pp414-430 17/4/03 1:51 pm Page 424

424 S. Baragetti, P. Clerici and S. Matteazzi

111
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
2011
1
2
3
4
5
6
7
8
9
30
1
2
3
4
5 Figure 3 (a) Sketch of the equipment for tension and compression while make up and
6 (b) sketch of the casing tongs.
7
8 The values of deformation given by the external strain gages are as follows:
9 x in the pipe axial direction;
40  in the pipe circumferential direction;
1 1
z = –– (X +  + x) in the 45° deviated, with respect to the pipe axis, direction.
2 2
3 By using these strain values it was possible to deduce the shearing strain
4 x = 2z x y.
5 In order to evaluate the shoulder axial load induced by the make up torque and by
6 axial tensile or compressive loads, a rearrangement of the strain measured values, accord-
711 ing to the tubular shell theory [19] is required; as a matter of fact the knowledge of only
Baragetti pp414-430 17/4/03 1:51 pm Page 425

Friction and tightening force of conical threaded connections 425

111
2
3
4
5
6
7
8
9
10
1
2
3
4 Figure 4 Disposition of the strain gages on the box of the inside and outside machined
5 NC50 connection.
6
7 the internal and external axial strains and of the external circumferential strains is not
8 enough to deduce the shoulder load.
9 Equations defining axial and circumferential (internal and external) deformations are
2011 linked to the axial and circumferential, internal and external, stresses by means of the fol-
1 lowing equations (the convention that the external quantities are called “e” while the inter-
2 nal ones “i” is assumed):
3
4 1
e = ––– (e xe) (36)
5 E
6
7 1
8 xe = ––– (xe e) (37)
E
9
30 1
1 i = ––– (i xi) (38)
E
2
3 1
4 xi = ––– (xi i) (39)
E
5
6 The bore zone of the box, that is a cylindrical shell of a given thickness t, is subjected
7 both to axial and bending per unit length actions N and M with the following values of
8 the internal and external axial and circumferential stresses:
9
40 xe = N + M (40)
1
2
xi = N M (41)
3
4
––
5 e =  + M (42)
6 –– 
711 i =   M
(43)
Baragetti pp414-430 17/4/03 1:51 pm Page 426

426 S. Baragetti, P. Clerici and S. Matteazzi

111 where –– is the circumferential stress acting at a distance equal to t/2 from the external
2 or internal surface of the box shoulder bore zone.
3 By using Eqs (36)–(43) and substituting the values of the stresses provided by
4 Eqs (40)–(43) into Eqs (36)–(39), it is possible to obtain the following linear system in
5 the three unknown quantities,  ––,  ,  .
 M N
6



xi  2 1 1 ––

7 
8 1
e = –– 1 0  M (44)
9 E
10 xe  1 2 1 N
1
2 The solution of the linear system (44) enables the evaluation of the axial shoulder
3 stress (N and then both the calculation of the axial shoulder load and of Kup).
4
5
6 4 Results and discussion
7
4.1 Make up tests
8
9 The first series of tests enabled, once experimentally recorded the torque-turns diagrams,
2011 to evaluate the friction coefficient of the mating surfaces of the flanks of the threads. The
1 load cycle consisted in putting pin and box under compression without the application of
2 any make up torque and without shoulder contact, applying the make up torque, breaking
3 out and unloading. A 40% zinc compound was used to grease the threads; the compound
4 was renewed every 5 M/Bs and the make up speed was set equal to 1 rpm.
5 Evaluation of the friction coefficient among the flanks of the threads was made by
6 using the torque value of the first part of the torque/turns diagram (make up phase) record-
7 ed in correspondence of the point for which the curve presents an evident slope variation.
8 In Figure 5, in which a typical torque-turns diagram is shown, this point is marked with M*
9 and is part of curve “a” (in case of make up torque). Without shoulder contact, under com-
30 pression, the relative rotation between pin and box becomes non null as soon as the make
1 up torque reaches the value for which the mating surfaces begin to have relative motion.
2 This point gives the make up torque for the calculation of the static friction coefficient
3 while the horizontal portion of curve “a” would enable to determine the dynamic friction
4 coefficient in presence of relative motion. From the experimental tests it is possible to
5 obtain Mup= M* and Qex and later to calculate Kt by means of Eqs (34) and (35).
6 The results showed that the mean value of Kt is equal to 0.8 mm with a mean square
7 deviation equal to 0.02 mm; so these tests permitted us to verify that the static friction
8 coefficient, obtainable by using a 40% zinc compound, has a mean value equal to f=0.08
9 (in accordance with the API standard [1]).
40 In the second test phase, preliminary strain gauges measurements enabled to quantify
1 Kup for make up speed 1 rpm. The results showed that with the application of the nomi-
2 nal make up torque, equal to 24 kNm for the NC50 steel connections, it is possible to cal-
3 culate a mean Kup value equal to 12.5 mm. The value is almost the same as the theoretical
4 API one (Kup = 12.49 mm).
5 In the third test phase make up/break out tests were carried out on NC50 connections
6 with different make up speeds. The first tests were carried out by using a 1 rpm make up
711 speed. Later, tests were carried out with make up speeds respectively equal to 5 rpm and
Baragetti pp414-430 17/4/03 1:51 pm Page 427

Friction and tightening force of conical threaded connections 427

111
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8 Figure 5 Torque-turns diagram.
9
2011
1 10 rpm in order to give evidence to the trustworthiness of Eq. (14), that enables to estab-
2 lish a relationship between Kup and Mup/Mout by using only the recorded torque-turns dia-
3 gram, even for higher make up speeds.
4 Probability charts were constructed in order to explain the results and propose opera-
5 tive values of the parameters that influence the make up procedure.
6 All the results of the tests, in terms of Kup calculated by using Eq (14) with maximum
7 torque values (Mup and Mout in Figure 5), were put together and normal distributions were
8 drawn. In the probability charts the mean Kup value is the intersection point of the distri-
9 bution with the abscissa (Kup values axis), while the distribution dispersion can be viewed
30 as the slope of the drawn curves.
1 In Table 2 the results of the tests, obtained at different make up speeds and on differ-
2 ent NC50 steel RSCs, are reported; for each make up speed 50 tests were carried out.
3 Some tests were carried out on the same connection with different make up speeds.
4 The results are shown in Figure 6.
5 These results, summarised in Table 2, put in evidence that the mean Kup values are
6 rather different both in comparison to the API one and to the one measured by means of
7 strain gauges at 1 rpm. Data dispersion increases as soon as the make up speed increases
8 (for 1 rpm and 5 rpm the differences in the mean square deviations are lower then in the
9 case of 10 rpm); the mean square deviations in the cases of make up speeds equal respec-
40 tively to 1 rpm and 5 rpm are 1.84 mm and 1.15 mm while in the case of 10 rpm the mean
1 square deviation assumes the value 5.07 mm.
2 It is worth emphasizing also the presence of a bias in the mean Kup values recorded at
3 the three make up speeds taken into consideration; so, in the cases of make up speeds of
4 1 rpm and 10 rpm the mean value is higher than the API one, while in the case of 5 rpm it
5 is lower.
6 The analysis of the results shows that the behaviour of the NC50 connection is dif-
711 ferent for the speeds of make up taken into consideration and Kup, if evaluated by using
Baragetti pp414-430 17/4/03 1:51 pm Page 428

428 S. Baragetti, P. Clerici and S. Matteazzi

111 Table 2 Kup mean values and mean square deviations for the tested make up speeds
2 [Kup calculated with Eq. (14)]
3 Kup (API) = 12.49 mm 1 rpm 5 rpm 10 rpm
4
Kup mean value (mm) 14.28 10.66 16.85
5 Mean square deviation (mm) 1.84 1.15 5.07
6
7
8
9
10
1
2
3
4
5
6
7
8
9
2011
1 Figure 6 Probability charts of Kup [calculated with Eq. (14) obtained by using make up speeds
2 respectively equal to 1 rpm, 5 rpm and 10 rpm (same connection)].
3
4
5 Eq. (14) with the maximum values of make up and break out torques, depends on the rota-
6 tion speed during the make up torque application.
7 In case of low make up speed (1 rpm) one can obtain, from experimental tests and
8 analytical equations, similar values of Kup, while for rotation speeds of 5 rpm and 10 rpm
9 the differences become greater and one can obtain respectively a lower and a higher Kup
30 mean value in comparison with the theoretical API one.
1 Furthermore the behaviour is no longer the same if one considers the results obtained
2 for one connection tested at different rotation speeds as confirmed by the results reported
3 in Figure 6.
4 This behaviour can be probably explained by taking into consideration the changes
5 in the friction characteristics of the contacting surfaces when compound is subjected to
6 temperature increases due to higher make up speed in the case of make up speed equal
7 to 10 rpm. Compound redistribution at different make up speeds could represent an
8 explanation for this behaviour too. The effect is a higher results dispersion, in the case of
9 different tested connections, as reported in Table 2: mean square deviation in case of make
40 up speed 10 rpm is 5.07 mm.
1
2
3
4.2 Make up tests with axial applied loads
4
5 The experimental tests with axial tensile applied loads enabled us to evaluate the Kb coef-
6 ficient value. Kb coefficient was calculated by using Eq. (29). More details about these
711 tests can be found in [17, 18].
Baragetti pp414-430 17/4/03 1:51 pm Page 429

Friction and tightening force of conical threaded connections 429

111 In these tests the box shoulder was modified in order to be able to put strain gauges
2 and measure the axial induced load after make up and further tensile applied load (the
3 load cycle consists first in the application of the make up torque and then of the axial ten-
4 sile load; the box shoulder section results to be equal to Ab′ = 6107 mm2 and the API Kb
5 coefficient results to be equal to 0.581 (without the modification it would be equal to
6 0.654). The mean experimental Kb value is equal to 0.618 with a standard deviation equal
7 to 0.012; the value is quite far from the API one with a difference more than 6% [17].
8
9
10 5 Conclusions
1
2 The friction behaviour under make of API NC50 rotary shouldered connection was inves-
3 tigated. The results of the investigation provide the rig site user with an operative proce-
4 dure for evaluating Kup coefficient and practical values, mean and mean square deviation,
5 of the parameters characterising the friction behaviour of this kind of connection.
6 The following conclusions can be drawn:
7
8 ● Four sets of tests were carried out with and without axial applied loads; a 40% zinc
9 compound was used to grease the connections and the compound was renewed every
2011 5 M/B in all the tests
1 ● A first set of experimental tests gave the friction coefficient of the contacting surfaces
2 and the Kup coefficient that links the make up torque and the axial shoulder induced
3 loads, by means of load cycles of the kind load/make up/break out/unload. The mean
4 value of friction coefficient results to be equal to f=0.08 with 40% zinc compound
5 ● Make up/break out tests were performed for make up speeds equal to 1 rpm, 5 rpm and
6 10 rpm. Kup values were calculated by using a new formulation in which only the peak
7 make up and break out torque values, recorded in the torque-turns diagram, are needed.
8 The results showed that only for make up speeds up to 5 rpm Kup is close to the API one
9 even if a bias in the Kup mean values is always present and equals respectively 1.79 mm
30 in the case of 1 rpm make up speed and 1.83 mm in the case of 5 rpm make up speed.
1 For make up speeds less than 5 rpm the mean square deviation never exceeds 2 mm.
2 Bearing in mind this last considerations a practice recommendation, when using the pro-
3 posed procedure for the evaluation of the friction capabilities of the RSC through the
4 experimental measurement of the Kup coefficient, can be deduced and the user should
5 not make up the NC50 rotary shouldered connections over a 5 rpm make up speed.
6
7
8 Acknowledgements
9
40 The authors wish to thank ENI-AGIP S.p.A. for permission to publish this paper.
1
2
3 References
4
5 1 API Standard (1994) ‘Specification for rotary drill stem elements-specification,’ 7, 38th edn.
6 2 API Recommended Practice (1995) ‘Recommended practice for drill stem design and operat-
711 ing limits’, RP 7G, 14th edn.
Baragetti pp414-430 17/4/03 1:51 pm Page 430

430 S. Baragetti, P. Clerici and S. Matteazzi

111 3 Baryshnikov, A., Ferrara, P., Schenato, A. and Curioni, E. (1995) ‘Makeup torque and rotary
2 shouldered connection reliability’, SPE/IADC 29352, IADC/SPE Drilling Conference,
pp. 197–204
3
4 4 Baryshnikov, A., Ferrara, P., Toffolo, G. and Donati, F. (1998) ‘Drillstring components ser-
viceability: Standards and Practices’, 2nd ARPO Convention Proceedings, ENI-Agip, Milan,
5 Italy.
6 5 Bailey, E. I. and Smith, J. E. (1992) ‘Testing thread compound for rotary-shouldered connec-
7 tions’, SPE/IADC 23844, IADC/SPE Drilling Conference, pp. 35–44.
8 6 Holcomb, R. T., Wilcox, V. T. and Oldiges, D. A. (1986) ‘The performance characteristics and
9 economics of rotary-shouldered connection thread compounds’, IADC/SPE 14794, IADC/SPE
10 Drilling Conference, pp. 595–602.
1 7 API Recommended Practice (1992) Recommended practice for testing of thread compound for
2 rotary shouldered connections. 7A1, 1st edn.
3 8 Baragetti, S., Baryshnikov, A., Guagliano, M., Terranova, A. and Vergani, L. (1999)
4 ‘Sollecitazioni in giunzioni coniche filettate: risultati numerici e sperimentali’, Convegno
5 Nazionale AIAS ’99, Vicenza, Italy.
6 9 Barry, W. R. (1975) ‘Finite element analysis of casing threads’, Proc. ASME Petroleum Mech.
Engng. Conf., Tulsa, Oklahoma.
7
8 10 O’Hara, P. (1974) ‘Finite element analysis of threaded connections’, Proc. Army Symp. Solid
Mech., 1974, Bass River, Mass.
9
11 Zhao, H. (1996) ‘A numerical method for load distribution in threaded connections’, Trans.
2011 ASME, 118, 274–279.
1
12 Zhao, H. (1998) ‘Stress concentration actors within bolt–nut connectors under elasto-plastic
2 deformation’, Int. J. Fatigue, 20(9), 651–659.
3 13 Farr, A. P. (1957) ‘Torque requirements for rotary shouldered connections’, The Oil and Gas
4 Journal, December, 109–114.
5 14 Sopwith, D. G. (1948) ‘The distribution of load in screw threads’, Inst. Mech. Engrs. Appl.
6 Mech. Proc., 159, 373–383.
7 15 Massa, E. (1990) Costruzione di Macchine, Milan: Masson.
8 16 Weatherford, Int’L. (1994) ‘Lube-guard high temperature-high pressure tool joint compound’,
9 Product Information, Oil Center Research Inc., USA.
30 17 Baragetti, S. and Baryshnikov, A. (2001) ‘Rotary shouldered thread connections: working limit
1 under combined static loading’, ASME J. Mech. Design, 123, 456–463.
2 18 Baragetti, S. (2002) ‘Effects of taper variation on conical threaded connections load distribu-
3 tion’, ASME J. Mech. Design, 124, 320–329.
4 19 Timoshenko, S. P. and Woinowsky-Krieger, S. (1959) Theory of Plates and Shells. New York:
5 McGraw-Hill.
6
7
8
9
40
1
2
3
4
5
6
711

You might also like