Research Article

Advances in Mechanical Engineering

2016, Vol. 8(7) 1–16
Ó The Author(s) 2016
Static metal sealing mechanism of a DOI: 10.1177/1687814016654821
aime.sagepub.com
subsea pipeline mechanical connector

Zongliang Wei1, Liquan Wang1, Yu Guan1, Shaoming Yao2 and Shaokai Li1

Abstract
The static metal sealing mechanism is essential for the structural design of subsea pipeline mechanical connector. By ana-
lyzing the sealing mechanism and contact pressure on sealing face, the critical condition of the seal is obtained. Based on
the superposition theorem of elasticity, the formula of critical mean contact pressure ensuring seal is deduced. The finite
element analysis result of critical mean contact pressure agrees with the analysis result very well and the critical mean
contact pressure decreases with the increase in pipeline wall thickness and sealing width; however, the radius of pipeline
has less influence on it. The critical mean contact pressure of 8-in subsea pipeline mechanical connector is 223.34 MPa
and subsea pipeline mechanical connector is designed with ANSYS. The internal pressure experiments indicate that the
sealing performance of final designed subsea pipeline mechanical connectors is higher than the design pressure of
4.5 MPa and all the initial designed ones cannot seal up to 4.5 MPa. By observing the sealing indentation on pipeline sur-
face, when the mean contact pressure is higher than the critical mean contact pressure and the contact pressure var-
iance is smaller, a full contact is seen to form a seal. Finally, the vibration experiments prove that the number of vibration
is more than 107 and final designed subsea pipeline mechanical connector meets the experiment specification.

Keywords
Mechanical connector, sealing mechanism, superposition principle, critical mean contact pressure, sealing performance

Date received: 28 January 2016; accepted: 23 May 2016

Academic Editor: Michal Kuciej

Introduction pipelines. The first type of the connectors includes com-

Because of the rapid growth of offshore oil and gas
exploitation, more and more pipelines are applied to
deep seabed to transport oil and gas,1 and the operation
will increase in pipeline laying, merger, and mainte-
nance.2 However, how to accomplish mechanical con-
nections between pipelines reliably as well as the related
equipment is the key problem to build a subsea oil and
gas production system successfully.3
For various operation depths and connection
requirements, corporations in the Unites States, such as
Oil state, Cameron, FMC technologies, Quality
Connector Systems, and Oceaneering, have developed
a series of mechanical connectors. Two typical connec-
tors are well known in the industry. One is used for the
flange-end pipelines and the other is for non-flange
mon flange connector, articulated clamp connector,4
and collet connector,5 which are mainly used for con-
nection in pipeline laying and merging. The installation
of the connectors is achieved by a special tool, and the
sealing principle of the connectors is to squeeze the gas-
ket between the flange ends by the axial clamping force.
1
College of Mechanical and Electrical Engineering, Harbin Engineering
University, Harbin, China
2
Advanced Manufacturing Research Centre with Boeing, The University
of Sheffield, Sheffield, UK
Corresponding author:
Zongliang Wei, College of Mechanical and Electrical Engineering, Harbin
Engineering University, 2062 Room, the 61st Building, 145 Nantong Road,
Harbin 150001, China.
Email: weizongliang1207@163.com

Creative Commons CC-BY: This article is distributed under the terms of the Creative Commons Attribution 3.0 License
(http://www.creativecommons.org/licenses/by/3.0/) which permits any use, reproduction and distribution of the work without
further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/
open-access-at-sage).
Downloaded from ade.sagepub.com by guest on August 17, 2016
2 Advances in Mechanical Engineering

The other type of connectors is applied in the connec- predicted that flange rotation had effects on sealing
tion of pipeline without flange. A flange can be fitted performance. In addition, finite element method was
at the end of the pipeline by mechanical connection or applied to research the distribution of the contact pres-
using double-grip mechanical connectors, such as sure on the sealing surface.18,19 Their work considered
hydraulic double-grip and seal connector,6 smart the influence of nonlinear characteristic of gasket and
flange,7 and pipeline connector for connecting a branch internal pressure; however, the analysis was based on
pipe to a carrier pipe,8 which are also used for des- the spiral wound gasket. Our previous article20 proved
tructed pipeline renewal. During the connecting opera- that the mean contact pressure and its distribution on
tion, destructed pipeline is cut off by the subsea cutting sealing surface had an important effect on static metal
machinery, replaced by an intact pipeline prefabricated seal. However, no relevant public research has been
on board, and connected together by the mechanical done on sealing mechanism of the mechanical connec-
connectors. The complicated design of the conventional tor without a gasket.
subsea mechanical connectors results in inefficient In order to obtain reliable sealing performance of
operation, and particularly most of the bolts have to be SPMC, we analyzed the sealing mechanism of static
preloaded. Then, a simple mechanical connector is metal sealing without a gasket and established the criti-
designed for the connection of water supply and drai- cal condition of the seal. Based on the superposition
nage pipelines under low or zero pressure on land.9 On theorem of elasticity, the formula of the critical mean
the basis of this connector design, Haelok Inc. in contact pressure is deduced, which can be used to calcu-
Switzerland supplies the connectors connecting pipeline late the minimum contact pressure of static metal seal.
less than 168.3 mm in diameter. To make this kind of Meanwhile, the effects of the key structural parameters
connector work well in complicated subsea environ- on the critical mean contact pressure are analyzed. The
ment, this article presents an improved design, the sub- 8-in SPMC is designed with ANSYS, and the internal
sea pipeline mechanical connector (SPMC), that can be pressure experiments, indentation experiments of the
used for a larger diameter and a higher internal sealing face, and vibration experiments are carried out.
pressure.
The seal of SPMC is a static metal seal without a
gasket, which is the key technique to realize static metal Structure and sealing mechanism of
sealing in the connector. The static metal seal is exten- SPMC
sively applied in subsea non-weld connections to pre-
The structure of SPMC is shown in Figure 1, which is
vent the leakage of the transportation medium, because
composed of a basic body and two press rings. The
of its advantages in condition with high pressure, corro-
external surface of the press ring is a straight cylinder,
sion, and complicated external forces.10 In the design of
and the internal surface is a profile that consists of
a mechanical connector with a gasket, the gasket factor
cylindrical and conical surfaces. On the external sur-
and minimum gasket seating stress must be determined
faces of the basic body, two convex rings are produced
by the design codes11,12 according to the internal pres-
on each side. The two ends of the basic body are grip-
sure and structural geometry, but these codes are based
per parts. On the internal surface of the basic body,
on a series of assumptions and hence may not calculate
two sealing rings are produced on each side.
the contact pressure and its distribution between flange
The installation of SPMC is shown in Figure 2. The
ends and metal gasket.13,14 ASME Boiler and Pressure
press ring is compressed by installation tool and moves
Vessel Code (BPVC) Special Working Group-Bolted
toward the middle of basic body along the axial direc-
Flanged Joints (SWG-BFJ)15 improved the design
tion. When the press ring moves toward the middle of
method of conventional bolted flange joints with gas-
SPMC, its internal conical surfaces will compress
ket, and new gasket constants were defined. However,
all conclusions were obtained by the assumption that
the contact pressure was uniform on the sealing surface.
Sawa et al.16 had made a research in which the distribu-
tion of contact stress which governs the sealing perfor-
mance is analyzed as a three-body contact problem,
using an axisymmetric three-dimensional (3D) theory
of elasticity. This work was further simplified by treat-
ing the pipe flange and the gasket as hollow cylinders.
The formula of gasket contact stress and its variation
through the gasket width was deduced on the basis of
nonlinear characteristic of the gasket materials,17 which
was supported by the experiment and finite element
analysis (FEA). Nonetheless, this method only Figure 1. Structure of SPMC.

Downloaded from ade.sagepub.com by guest on August 17, 2016

Wei et al.
Figure 2. Installation of SPMC.
sealing rings and gripper parts to the external surface of
the pipeline. The press ring is compressed further,
which will increase the contact pressure on the sealing
face. When the contact pressure exceeds the critical
contact pressure, enough deformation on sealing face
will happen, which forms a static metal sealing.
Meanwhile, gripper parts are embedded into the exter-
nal surface of the pipeline and connect the pipelines as
a whole piece, which resists the complicated external
force, protects sealing face, and improves reliability and
lifetime of SPMC in submarine environment. In the
installation of SPMC, Remote Operated Vehicle
(ROV) can operate other equipments, and operators
on board can observe work environment in underwater
by ROV.
Static metal sealing mechanism
The internal surfaces of the sealing rings are not per-
fect and its roughness is a little better than that of the
pipeline, as shown in Figure 3. If the deformation on
the sealing face is not enough, two sealing faces will
not fully contact each other, which tends to generate
the micro leak path,21 particularly when external
forces and internal pressure are applied, as shown in
Figure 4. In order to achieve a reliable static metal
sealing, the contact pressure on the sealing faces
should reach a critical value. The micro leak paths
will be blocked by reducing the height of micro-peaks
and the depth and diameter of micro-pits or increas-
ing the diameter of micro-peaks, which achieves a
reliable static metal sealing.22 The micro leak paths
are formed by linking micro chambers formed by the
overlay of micro-peaks and micro-pits on the sealing
faces. Because micro-peaks and micro-pits on the
sealing surface obey the normal distribution, the
increase in the sealing face width will effectively
reduce the probability of the micro leak paths genera-
tion. The contact pressure and sealing face width are
two important parameters in the design of static
Downloaded from ade.sagepub.com by guest on August 17, 2016
3
Figure 3. Micro surface.
Figure 4. Micro leak path.
metal seal.23–25 In order to generate reliable static
metal sealing, which will not be affected by the sur-
face roughness, the contact pressure on the sealing
face should be more than the Brinell hardness HB of
the softest material.26 Taking the design of Badische
Anilin-und-Soda-Fabrik (BASF) lenticular gasket of
Germany as a basis, the width of the sealing face is
not less than 1/16 in.27
Critical condition of seal
By considering subsea environment, the sealing rings
and pipeline should be in elastic deformation and have
enough elastic spring-back to resist the negative effects
of internal pressure, temperature, and external load for
seal.28 Assuming that the deformation of the sealing
part is always kept within the range of elastic deforma-
tion range in installation and operation condition. LAD
is the width of the sealing face. Because of the inter-
coordination of elastic deformation between the sealing
ring and the pipeline, the contact pressure on the
boundary of the sealing face is higher than that of the
central plane and the distribution is symmetric about
the central plane of the sealing face, as shown in Figure
5. On the basis of static metal sealing mechanism in sec-
tion ‘‘Static metal sealing mechanism,’’ the critical con-
dition to realize reliable static metal sealing is
4 Advances in Mechanical Engineering
Figure 5. Distribution of contact pressure.
8
< PB = HB
PC = HB
:
LAB + LCD = 1=16 in ’ 1:6 mm
where PB is the contact pressure at B, PC is the contact
pressure at C, LAB is the distance between A and B, and
LCD is the distance between C and D.
 is the
In order to realize the static metal sealing, P
mean contact pressure on the sealing face and should
be higher than the critical mean contact pressure
Lð
AD
P  = 1
P P(z)dz
LAD
0
where the contact pressure on the sealing face, P(z),
  is
meets the critical condition of equation (1) and P
the critical mean contact pressure.
Solution of critical mean contact pressure
The contact pressure on the sealing face includes two
parts, that is, the contact pressure from installation and
that from the internal pressure. Figure 6 shows the dis-
tribution of the internal pressure. The internal and
external surfaces of the pipeline are under the internal
pressure on the left of ‘‘W’’ line and there is no expan-
sion tendency for the pipeline. However, on the right of
W line, the internal pressure is applied on the internal
surface only and expansion of the pipeline tends to
happen, which increases the contact pressure on the
sealing face and improves the sealing performance of
SPMC. Based on the above analysis, the SPMC is a
self-tight seal. As long as the mean contact pressure
from installation is more than the critical mean contact
pressure, the SPMCs will maintain their high sealing
performance. Therefore, the mean contact pressure
from installation is analyzed to design SPMC.
Downloaded from ade.sagepub.com by guest on August 17, 2016
Figure 6. Distribution of internal pressure.
ð1Þ
Figure 7. Approximate model of contact pressure.
ð2Þ Equivalent model of contact pressure
The static metal seal of SPMC is realized by the squeeze
on the sealing surfaces. By the analysis of structural fea-
ture and force condition of the pipeline and SPMC, the
geometrical profile, constraints, and boundary condi-
tions of pipeline are symmetric about the central line,
and the mechanical model of the pipeline is a 3D axial
symmetric constitutive model.
The cylindrical coordinate system is established, as
shown in Figure 7. z-axis coincides with the central line
of the pipeline, r points along the radial direction, u is
along the circumferential direction, and r–u plane coin-
cides with the low end of band of the contact pressure.
In axisymmetric elastic problem, stress, strain, and dis-
placement of pipeline wall only relate to r and z.
Because the contact pressure is non-uniform on the
sealing face, we use a sectional approximation method
to solve the formula of the contact pressure on the seal-
ing face.
LAD is divided uniformly into n sections, and the
width of every section is li = LAD/n = l. If n is big
enough, we can assume that the contact pressure pi is
uniform with a width of li, and the equivalent mechani-
cal model is n bands with uniform contact pressure pi
with a width of li applied on the external surface of the
pipeline, where i = 1, 2,., n.
Wei et al. 5

Four stress components (sor 1 , sou1 , soz 1 , torz1 ) can be

expressed as
8  2

>
> sor 1 = ∂z∂ mr2 C  ∂∂rC2
>
>  
>
< sou1 = ∂ mr2 C  1 ∂C
∂z h r ∂r i
2 ð7Þ
>
> soz 1 = ∂z∂ (2  m)r2 C  ∂∂zC2
>
> h i
>
: to1 = to1 = ∂ (1  m)r2 C  ∂2 C
rz zr ∂r ∂z2

where m is the Poisson’s ratio, r2 is the Laplace opera-

tor, and
∂2 1∂ ∂2
Figure 8. Superposition of pressure. r2 = + + ð8Þ
∂r2 r ∂r ∂z2
If r2C = 0, the stress function C will meet biharmo-
Under band of uniform contact pressure pi with a nic function (6). Substitute C into equation (8), and the
width of li, the stress in the pipeline can be solved by result is
superposition method in Figure 8. The stress of the
∂2 C 1 ∂C ∂2 C
outer radius at z = 0 and z = l1 is pi/2 or pi, which will 2
+ + 2 =0 ð9Þ
not affect the distribution of stress in the pipeline, ∂r r ∂r ∂z
because the area is 0. Suppose C = T(r)cos(kz), the equation (9) can be
rewritten as

Solution of pressure band with width li ∂2 T (r) 1 ∂T (r)

2
+  k 2 T (r) = 0 ð10Þ
Superposition method. On z positive axis, the stress state ∂r r ∂r
in the pipeline under band of pi with a width of li is Equation (10) is correctional zero-order Bessel func-
equivalent to that under band of pi with a width of l1 tion. Because the pipeline is hollow, the general solution
which moves (i 2 1)l along the z-axis direction. In of equation (10) is
Figure 8, according to superposition principle of elasti-
city, the stress state in orz is equal to the sum of the T (r)= a0 I0 (kr)+a1 krI1 (kr)+b0 K0 (kr) + b1 krK1 (kr) ð11Þ
stress state in o1r1z1 and that in o2r2z2, whose equation
is Substitute equation (11) into C = T(r)cos(kz), the
stress function, C, can be expressed as
s1m (r, z) = som1 (r, z) + som2 (r, z) ð3Þ
C = cos (kz)½a0 I0 (kr) + a1 krI1 (kr)
The relational expression is obtained by theory of + b0 K0 (kr) + b1 krK1 (kr) ð12Þ
elasticity
where four coefficients, a0, a1, b0, and b1, are solved by
som2 (r, z) =  som1 (r, z  l) ð4Þ the boundary conditions of the pipeline, I0(kr) is the
first kind correctional zero-order Bessel function, I1(kr)
where s1m (r, z) is the stress in orz, som1 (r, z) is the stress in is the first kind correctional one-order Bessel function,
o1r1z1, and som2 (r, z) is the stress in o2r2z2; m = r, u, z. K0(kr) is the second kind correctional zero-order Bessel
function, and K1(kr) is the second kind correctional
one-order Bessel function; the expressions are
Solution of stress in o1r1z1. When the influence of gravity 8
is negligible, the equilibrium differential equation of > P
þ‘ krv + 2n
< Iv (kr) = 1
n!G(v + n + 1) 2
stress in o1r1z1 is n=0 n o ð13Þ
>
: Kv (kr) = lim p
( o o ½Im (kr)  Im (kr)
o
∂sr 1
o
∂t zr1 sr 1 su1 m!v 2 sin (mp)
+ + =0
∂ro
∂sz 1
∂zo
∂t zr1
o
t zr1
r ð5Þ
+ + =0 where v is equal to 0 or 1.
∂z ∂r r
Assume a0 = e1b1, a1 = e2b1, b0 = e3b1, and equation
To meet equation (5), the stress function C should (12) can be rewritten as
be the biharmonic function
C = b1 cos (kz)½e1 I0 (kr) + e2 krI1 (kr)
r4 C = 0 ð6Þ + e3 K0 (kr) + krK1 (kr) ð14Þ

Downloaded from ade.sagepub.com by guest on August 17, 2016

6 Advances in Mechanical Engineering

8
The stress function C meets equation (6) regardless ð
þ‘
< pi =2 z.0
 pi sin (kz)
of the value of k. Substitute b1 = b(k)dk the equation s o1 
r = dk = 0 z=0 ð19Þ
r = rout p k :
(14) can be rewritten as 0 pi =2 z\0

dC = cos (kz)½e1 I0 (kr) + e2 krI1 (kr) where rout is the outer radius of the pipeline and rin is
+ e3 K0 (kr) + krK1 (kr)b(k)dk ð15Þ the inner radius.
By solving equations (7), (13), and (16) with bound-
Equation (15) meets equation (6). Integrate equation ary conditions (17) and (19), the expressions (e1, e2, e3,
(15) within (0, +N), and the general solution of the b(k)) are
stress function C is 2 3 2 31 2 3
e1 G1 (krout ) G2 (krout ) G3 (krout ) Q(krout )
ð
þ‘ ð
þ‘ 4 e2 5 = 4 G1 (krin ) G2 (krin ) G3 (krin ) 5 4 Q(krin ) 5
C= dC = b(k) cos (kz)½e1 I0 (kr) + e2 krI1 (kr) e3 H1 (krin ) H2 (krin ) H3 (krin ) P(krin )
0 0 ð20Þ
+ e3 K0 (kr) + krK1 (kr)dk ð16Þ where
In o1r1z1, the force boundary conditions of the pipe- 8
> G1 (kr) = I1 (kr)
line are >
>
>
> G2 (kr) = krI0 (kr) + 2(1  m)I1 (kr)
 >
>
>
>
sor 1 r = rin = 0 >
>
> G3 (kr) =  K1 (kr)
 >
>
torz1 r = rout , r = 0 < Q(kr) = krK0 (kr)  2(1  m)K1 (kr)
in
8 ð21Þ
> ð17Þ >
> H1 (kr) = I0 (kr)  I1 (kr)=(kr)
 < pi =2 z.0 >
>
>
o1 
sr r = rout = 0 z=0 >
> H2 (kr) = (1  2m)I0 (kr) + krI1 (kr)
> >
>
: >
>
pi =2 z\0 > H3 (kr) = K0 (kr) + K1 (kr)=(kr)
>
>
:
P(kr) = (1  2m)K0 (kr)  krK1 (kr)
Because
8 b(k) =
ð
þ‘
< p=2 z.0 pi
sin (kz)
dk = 0 z=0 ð18Þ pk 4 ½ H1 (krout ) H2 (krout ) H3 (krout ) ½ e1 e2 e3 T  P(krout )
k :
p=2 z\0
0 ð22Þ
So Substitute equation (22) into equation (16), and the
stress function, C, can be rewritten as

C = pi
n o
ð
þ‘
½ I0 (kr) krI1 (kr) K0 (kr) ½ e1 e2 e3 T + krK1 (kr)
cos (kz)
n o dk ð23Þ
pk 4 ½ H1 (krout ) H2 (krout ) H3 (krout ) ½ e1 e2 e3 T  P(krout )
o

By solving equations (3), (4), (7), (8), and (23), s1r (r, z), s1u (r, z), and s1z (r, z) can be expressed as
8  

> Ð
þ‘ H1 (kr) H2 (kr) H3 (kr) ½ e1 e2 e3  P(kr)
1 sin (kz)sin½k(zl) 
T

>
> s1r (r, z) = pi 
dk
>
> p k
H1 (krout ) H2 krout H3 krout ½ e1 e2 e3  Pkrout
T
>
> 0  
>
> Ð
þ‘ T
>
> ½k(zl) I1 (kr)=(kr) (1  2m)I0 (kr) K1 (kr)=(kr) ½ e1 e2 e3 
>
> s1u (r, z) = pi p1 sin (kz)sin k
  T

dk
>
> H1 (krout ) H2 (krout ) H3 (krout ) ½ e1 e2 e3  P(krout )
>
<
0
Ð
þ‘
½k(zl)
pi p1 sin (kz)sin k
 (12m)K0 (kr) 
T

dk ð24Þ
>
> H1 (krout ) H 2 (kr out ) H3 (krout ) ½ e1 e2 e3  P(krout )
>
>
0  
>
> Ð 1 sin (kz)sin½k(zl) I0 (kr) 2(2  m)I0 (kr)  krI1 (kr) K0 (kr) ½ e1 e2 e3 T
þ‘
>
> s 1
(r, z) = p  
dk
>
> z i p k T
H1 (krout ) H2 (krout ) H3 (krout ) ½ e1 e2 e3  P(krout )
>
> 0
>
> Ð
þ‘
>
> ½k(zl)
: + pi p1 sin (kz)sin k
 2(2m)K0 (kr)krK1 (kr)
 T

dk
0 H1 (krout ) H2 (krout ) H3 (krout ) ½ e1 e2 e3  P(krout )

Downloaded from ade.sagepub.com by guest on August 17, 2016

Wei et al. 7

Based on elasticity method, physical equation ð

þ‘
(1 + m)r sinfk ½z  (i  1)lg  sin½k(z  il)
and geometric equation of 3D axial symmetric model uir (r, z) = pi
Ep k
are 0
½ I1 (kr)=(kr) I0 (kr) K1 (kr)=(kr) ½ e1 e2 e3 T
n o dk  pi
1 1  ½ H1 (krout ) H2 krout H3 krout ½ e1 e2 e3 T  Pkrout
e1u = su  m(s1z + s1r ) ð25Þ
E ð
þ‘
(1 + m)r sinfk ½z  (i  1)lg  sin½k(z  il)
u1r Ep k
e1u = ð26Þ 0
r K0 (kr)
n o dk
Using equations (24)–(26), the radial displacement of ½ H1 (krout ) H2 (krout ) H3 (krout ) ½ e1 e2 e3 T  P(krout )
the pipeline can be obtained as
ð28Þ
ð
þ‘
(1 + m)r sin (kz)  sin½k(z  l) Contact pressure on sealing surface
u1r (r, z) = pi
Ep k
0 According to the analysis of section ‘‘Equivalent model
½ I1 (kr)=(kr) I0 (kr) K1 (kr)=(kr) ½ e1 e2 e3 T of contact pressure,’’ the radial displacement in a pipe-
n o dk
½ H1 (krout ) H2 (krout ) H3 (krout ) ½ e1 e2 e3 T  P(krout ) line under n bands of contact pressure meets the super-
ð
þ‘ position principle, and based on equation (28), the
(1 + m)r sin (kz)  sin½k(z  l) radial displacement is
pi
Ep k
0
K0 (kr) X
n
n o dk ur (r, z) = uir (r, z) ð29Þ
T
½ H1 (krout ) H2 (krout ) H3 (krout ) ½ e1 e2 e3   P(krout ) i=1

ð27Þ If we choose n points at different locations of the

pipeline, the radial displacement ur at point j is
Verification of stress and displacement. In order to verify X
n
equations (24) and (27), stress components (s1r , s1z , s1u ) ur (rj , zj ) = uir (rj , zj ) ð30Þ
and radial displacement (u1r ) are compared with the i=1
results of FEA. Define l = (r 2 rin)/(rout 2 rin), when
Solve equation (30) with n points radial displace-
rout = is 109.5 mm, rin = 100 mm, l =10mm and pi
ments as the input, and contact pressure pi can be
=–1 MPa, the distribution of stress (s1r , s1z , s1u )
expressed as
and displacement (u1r ) along the z-axis are shown in
Figure 9. 2 3 2 31 2 3
Figure 9(a) shows the distribution of stress s1r along p1 a11 a12    a1n ur (r1 , z1 )
6 p2 7 6 a21    a2n 7 6 7
z-axis, Figure 9(b) shows the distribution of stress s1z , 6 7 6 a22 7 6 ur (r2 , z2 ) 7
6 .. 7 = 6 .. .. .. 7 6 .. 7 ð31Þ
Figure 9(c) shows the distribution of stress s1u , and 4 . 5 4 . . . 5 4 . 5
Figure 9(d) shows the distribution of displacement u1r . pn an1 an2    ann ur (rn , zn )
The stress components (s1r , s1z , s1u ) and displacement
component (u1r ) are symmetrical about the central plane where
and the absolute values decrease progressively when l  
 
ð
þ‘
decreases. When l is 0.4 or 1, the theoretical results (1 + m)rj sin k zj  (i  1)l  sin k(zj  il)
aji =
(s1r , s1z , s1u , u1r ) agree with FEA results very well. The Ep k
0
minor error is from the assumption of uniform pressure ½ I1 (krj )=(krj ) I0 (krj ) K1 (krj )=(krj ) ½ e1 e2 e 3 T
in a band as well as computation error of finite element n o dk
½ H1 ðkrout Þ H2 ðkrout Þ H3 ðkrout Þ ½ e1 e2 e3 T  Pðkrout Þ
method. Figure 9 also indicates that the expressions of
ð
þ‘  
 
stress and displacement are deduced correctly by (1 + m)rj sin k zj  (i  1)l  sin k(zj  il)

elasticity. Ep k
0
K0 (krj )
n o dk
Displacement. The radial displacement is equal to uir ½ H1 (krout ) H2 (krout ) H3 (krout ) ½ e1 e2 e3 T  P(krout )
the radial displacement u1r in the pipeline under band ð32Þ
of contact pressure pi (width is l1) moved (i 2 1)l along
z-axis. Solving equation (27), the radial displacement, When rj equals rout, equation (31) shows the relation
uir , is between radial displacement on pipeline surface and
contact pressure on sealing surface

Downloaded from ade.sagepub.com by guest on August 17, 2016

8 Advances in Mechanical Engineering

Figure 9. Stress and displacement: (a) distribution of s1r along z-axis, (b) distribution of s1z along z-axis, (c) distribution of s1u along
z-axis, and (d) distribution of u1r along z-axis.

2 3 2  3 2 3
p1 a11 a12    a1n 1 ur (rout , z1 )
6 p2 7 6 a21 a22    a2n 7 6  7
6 7 6 7 6 ur (rout , z2 ) 7
6 .. 7 = 6 .. .. .. 7 6 .. 7 ð33Þ
4 . 5 4 . . . 5 4 . 5
pn an1 an2 
   ann 
ur (rout , zn )

When P = [p1, p2,., pn] meets the critical condition

of equation (1), the critical mean contact pressure on
the sealing face is

Lð
AD

 1 1X n
P = P (z)dz ’ pi ð34Þ
LAD n i=1
0

Analysis of critical mean contact pressure

  , is the determi-
As the critical mean contact pressure, P
nant of SPMC’s sealing performance, the influence of
the structure parameters of the pipeline and sealing ring
(pipeline wall thickness T and sealing face width LAD)
on the critical mean contact pressure can be analyzed.
Figure 10. Relation of T and P .
When LAD is equal to 3 mm and rout is equal to 109.5
or 84.15 mm, the relation of the critical mean contact pres-
sure and pipeline wall thickness is shown in Figure 10.

Downloaded from ade.sagepub.com by guest on August 17, 2016

Wei et al.
The result indicates that the FEA result agrees with that
of analysis and both of them decrease nonlinearly with
the increase in pipeline wall thickness T. When rout =
109.5 mm, the critical mean contact pressure P   is
252.27 MPa at T = 5.5 mm and is 216.39 MPa at
T = 14.5 mm. When rout = 84.15 mm, the critical mean
contact pressure P  is 245.90 MPa at T = 5.5 mm and is
215.33 MPa at T = 14.5 mm. The radius of the pipeline is
increased by 30.1%, but the critical mean contact pressure
is increased by 2.6%, which shows the radius of the pipe-
line influences on the critical mean contact pressure on a
low level. Therefore, for different pipeline radius, the pipe-
line wall thickness T is an important parameter for the
critical mean contact pressure.
When T is equal to 9.5 mm and rout is equal to 109.5
or 84.15 mm, the relation of the critical mean contact
pressure and the width of sealing face is shown in
Figure 11. The result indicates that the FEA result
agrees with that of analysis very well and both of them
decrease almost linearly with the increase in the sealing
face width LAD. Under the same width of the sealing
face, the critical mean contact pressures of different
pipeline radial are nearly the same. When rout = 109.5
mm, T = 9.5 mm (rin = 100 mm) and LAD = 10 mm,
the critical mean contact pressure P   is 117.82 MPa.
When rout = 109.5 mm, the critical mean contact pres-
sure P   with LAD = 2 mm is 48.5% higher than that
with LAD = 6.5 mm. When rout = 84.15 mm, the critical
mean contact pressure P   with LAD = 2 mm is 48.3%
higher than that with LAD = 6.5 mm. Therefore, for
different pipeline radius, the width of sealing face LAD
is an important parameter for the critical mean contact
pressure.
Design of 8-in SPMC
Critical mean contact pressure
We take 8-in pipelines with API 5L standards (specifi-
cation for line pipe) as an example to demonstrate the
design of SPMC. The outer radius of the pipeline, rout,
is 109.5 mm, the inner radius, rin, is 100 mm, and the
yield strength, ss, is 235 MPa. According to the design
rules for static metal seal, the yield strength of sealing
ring should be higher than that of the pipeline, the
material of press rings and the basic body is Q345, and
its yield strength is 345 MPa. The Brinell hardness of
pipeline, HB, is HB158. With regard to comprehensive
factors such as sealing mechanism, lifetime, and load-
ing force in installation, the width of sealing ring LAD
is determined as 3 mm, so the value of z at B point is
0.8 mm and the value of z at C point is 2.2 mm. The
LAD is divided into 20 sections with equal spaces and n
is 20. Taking the above parameters into equation (34),
the critical mean contact pressure for 8-in SPMC, P   , is
223.34 MPa.
Downloaded from ade.sagepub.com by guest on August 17, 2016
9
Figure 11. Relation of LAD and P .
Finite element design
Structural parameters analysis. The contact pressure and
contact uniformity on sealing face affect sealing perfor-
mance of SPMC directly. In the design of SPMC, the
contact pressure is replaced by the mean contact pres-
sure and contact uniformity is expressed by contact
pressure variance. If the mean contact pressure is higher
  = 223.34 MPa, the smaller the contact pressure
than P
variance, the better the contact uniformity and the
higher the sealing performance of SPMC. By analyzing
SPMC’s structure, the interference D1 and offset D2
affect the mean contact pressure and contact pressure
variance on major sealing face, respectively. The inter-
ference D3 and offset D4 affect the mean contact pres-
sure and contact pressure variance on minor sealing
face, respectively, as shown in Figure 12. The variables
(D1, D2, D3, D4) are designed with ANSYS.
Finite element model. The structure of SPMC is axisym-
metric and two-dimensional (2D) finite element model
is established to save computing time. Because the
structure of SPMC is complicated, in order to obtain-
ing high-quality elements, the basic body and press ring
are divided into some quadrilateral planes which can
be meshed by a mapping method. The elements on the
sealing faces are refined to improve the accuracy. The
basic element is PLANE182, the contact element is
CONTA172, and the target element is TARGE169.
The finite element model is shown in Figure 13.
Results of FEA. Figure 14(a) shows the distribution of
the contact pressure on the major sealing face in the ini-
tial design. As the contact pressure variance is
40,316 MPa2, the contact of the sealing face is non-uni-
form, which explains the sealing face is line contact and
10
Figure 12. Key variables of SPMC.
Figure 13. Finite element model.
Figure 14. Contact pressure on major sealing face: (a) initial design and (b) final design.
tends to generate sealing failure. The mean contact
pressure is 171.9 MPa and is less than the critical mean
contact pressure of 223.34 MPa, which indicates the
deformation of the sealing face is not enough and tends
to form micro leak path. The distribution of the con-
tact pressure on the major sealing face in the final
design is shown in Figure 14(b). The contact pressure
variance is 17795 MPa2 and decreases by 55.86%. The
mean contact pressure is 245.2 MPa and is higher than
the critical mean contact pressure of 223.34 MPa.
Compared with initial design, the contact pressure
increases and contact uniformity is improved in the
final design, which can form a reliable seal.
Figure 15(a) and (b) shows the contact pressure of
the initial design and that of the final design on the
minor sealing face, respectively. The mean contact pres-
sure of initial design is 148.8 MPa and the contact pres-
sure variance is 20029 MPa2. The mean contact
pressure of the final design is 262.6 MPa and higher
Downloaded from ade.sagepub.com by guest on August 17, 2016
Advances in Mechanical Engineering
than the critical mean contact pressure of 223.34 MPa.
The contact pressure variance is 16912 MPa2 and
decreases by 15.56%. The final design improves the
sealing performance of the minor sealing face.
From Figure 16(a) and (b), the Mises stress of the
pipeline approximately equals 197.95 MPa, which only
makes elastic deformation on the pipeline. When
SPMC is under complicated external load, there is
enough elastic spring-back to keep the sealing. The
Mises stress of the sealing rings and press ring is less
than the yield strength, which will keep SPMC safe in
operation. After the design of SPMC, the values of the
variables are listed in Table 1.
Experiment
The design pressure of SPMC is 4.5 MPa. In order to
validate that the critical mean contact pressure can
form reliable sealing face for SPMC, the internal
Wei et al. 11

Figure 15. Contact pressure on minor sealing face: (a) initial design and (b) final design.

Figure 16. Distribution of Mises stress of final designed SPMC: (a) major sealing ring and (b) minor sealing ring.

pressure experiments, indentation experiments of seal- Table 1. Value of variables.

ing faces, and the vibration experiments are carried
out. These experiments have followed the experiment
specification of the mechanical connector.29 The experi-
mental devices include two initial designed SPMCs,
four final designed SPMCs, pressure gauge, shutoff
valve, water hydraulic pump, pipelines, Instron8801
(vibration source), and so on. The medium is water and
the temperature is 20°C.
The measurement range of the pressure gauge used
in the experiments is 0–16 MPa and the minimum scale
value is 0.5 MPa. The accuracy grade of the pressure
gauge is 1.6 and full-scale error is 0.256 MPa.
Internal pressure experiments
The schematic diagram of the internal pressure experi-
ments is shown in Figure 17. The internal pressure of
2 MPa is increased every 5 min by the water hydraulic
pump. When the internal pressure reaches the given
value, the shutoff valve is turned off to keep the pres-
sure. When the internal pressure is increased up to a
D1 (mm) D2 (mm) D3 (mm) D4 (mm)
2.1 2.2 2.65 6
certain value, the value of pressure gauge starts to drop
or leakage occurs, and this critical value is the sealing
performance of SPMC. Figure 18 shows experiment
facility of the internal pressure experiments.
Table 2 is the results of the internal pressure experi-
ments of the initial designed SPMCs and final designed
ones. The results indicate that the sealing performance
of the initial designed SPMCs is less than the design
pressure of 4.5 MPa and that of the final designed ones
is more than the design pressure. A pressure drop of
2.0 MPa can be seen in the experiment of the first one
of the initial designed SPMCs and the reason is that
there are some microdefects on pipeline surface and the
micro leakage paths are formed under a low mean con-
tact pressure. A pressure drop of 4.0 MPa is witnessed

Downloaded from ade.sagepub.com by guest on August 17, 2016

12 Advances in Mechanical Engineering

Figure 17. Schematic diagram of internal pressure experiment.

Figure 19. Cutting part of SPMC.

static sealing. The contact pressure variance is very

small and the sealing faces fully contact. Larger mean
contact pressure and smaller contact pressure variance
will block leakage paths entirely. In the beginning of
section ‘‘Solution of critical mean contact pressure,’’
when the internal pressure is higher, the sealing perfor-
mance of SPMC is higher. Therefore, the SPMC is a
self-tight seal and can resist a high internal pressure.
Meanwhile, the result shows that the design method
based on the critical mean contact pressure is valid
for SPMC.
Figure 18. Facility of internal pressure experiment.

Indentation experiments of sealing faces

in the experiment of the second one of the initial
designed SPMCs and the reason is that the sealing faces
cannot contact fully and is linear contact with a big
contact pressure variance, which tends to leak under
internal pressure. The final designed SPMCs are
designed with finite element method and its mean con-
tact pressure is higher than the critical mean contact
pressure of 223.34 MPa, which is enough to form a
To analyze the reasons to generate leakage for the ini-
tial designed SPMCs, the initial and final designed
SPMCs are divided by wire-electrode cutting method
and the indentation on pipeline surface is observed.
One half of SPMC is shown in Figure 19.
Figure 20 shows the indentation on pipeline surface
for the initial designed SPMC. The indentation locating
in ‘‘A’’ point of the major sealing face is not continuous

Table 2. Result for internal pressure experiments.

Class Group Keeping time (min) Pressure (MPa) Pressure gauge

Initial design First 5 2.0 Drop

Second 5 4.0 Drop
Final design First 5 14.0 No change
Second 5 10.0 No change

Downloaded from ade.sagepub.com by guest on August 17, 2016

Wei et al.
Figure 20. Indentation for initial designed SPMC.
Figure 21. Indentation for final designed SPMC.
Figure 22. Schematic diagram of vibration.
and the leakage occurs. When the internal pressure is
increased to 2.0 MPa, a pressure drop occurs. That is
because the mean contact pressure is less than the criti-
cal mean contact pressure, which cannot make micro-
structure on the sealing face generate enough
deformation and micro leakage paths be blocked
entirely. Although the indentation at ‘‘B’’ location is
continuous, it is not uniform. The indentation depth
locating at the bottom of the sealing face is more than
Downloaded from ade.sagepub.com by guest on August 17, 2016
13
that locating on the top of the sealing face and the real
contact width is less than the critical width of 1.6 in.
When the SPMC is under internal pressure, the seal
fails easily. The mean contact pressure and contact
pressure variance ensure the sealing performance
of SPMC.
Figure 21 shows the indentation on pipeline surface
for the final designed SPMC. The mean contact pres-
sure is higher than the critical mean contact pressure
14 Advances in Mechanical Engineering

Table 3. Result for vibration experiments.

Class Group Keeping time (h) Number of vibration Pressure gauge

Final design First 139.3 10,029,670 No change

Second 139.0 10,004,652 No change

performance but also it has high ability in resisting

vibration.

Conclusion
The article investigated sealing mechanism of SPMC
and established the critical condition of seal. On the
basis of the superposition principle of elasticity, the for-
mula of the critical mean contact pressure was deduced.
The SPMC was designed with ANSYS and the sealing
performance was verified by the internal pressure
experiments, indentation experiments, and the vibra-
tion experiments.
According to the analytical analysis, finite element
Figure 23. Experimental facility of vibration.
simulation, internal pressure experiments, indentation
experiments, and vibration experiments, the following
conclusions can be drawn:
and the indentation on pipeline surface is continuous.
On the partial enlarged observation, the sealing face 1. Based on static metal sealing mechanism, the crit-
has a similar indentation depth, the contact face is uni- ical condition realizing reliable static metal seal
form, and the width is 3 mm. The sealing performance for SPMC was established and the formula of the
of SPMC is improved observably. critical mean contact pressure was deduced.
2. By comparison of the analysis and FEA result
of the critical mean contact pressure, the critical
Vibration experiments mean contact pressure decreases nonlinearly
with the increase in pipeline wall thickness and
Because the initial designed SPMCs show a poor seal- decreases almost linearly with the increase in
ing ability in the internal pressure experiments, the sealing face width. The radius of the pipeline
vibration experiments are carried out only for the final has less influence on the critical mean contact
designed SPMCs. The schematic diagram of vibration pressure. The FEA result agrees with that of
is shown in Figure 22. Based on experiment specifica- analysis very well.
tion of the mechanical connector,29 the vibration source 3. The critical mean contact pressure of 8-in
(Instron 8801) provides the sine displacement wave, SPMC is 223.34 MPa which is used to design
S = A sin(2pft), where the amplitude A is 0.097 mm SPMC. The internal pressure experiments were
and the frequency f is 20 Hz. When the cycle number of carried out for both the initial and final designed
vibration is on the level of 107, leakage does not hap- SPMCs. The results indicate that the sealing
pen, which indicates the final designed SPMC meets the performance of the final designed SPMC is
experiment specification of the mechanical connector. higher than design pressure of 4.5 MPa and all
Before the experiments, the design pressure of 4.5 MPa the initial designed SPMCs failed in the experi-
is applied and the shutoff valve is turned off. Figure 23 ment. The sealing performance of the final
shows experiment facility of the vibration experiments. designed SPMCs is better than that of the initial
Table 3 is the results of the vibration experiments of designed ones.
the final designed SPMCs. The cycle numbers of vibra- 4. By the observation on the indentation on the
tion are 10,029,670 and 10,004,652, respectively, which sealing face, if the mean contact pressure is less
indicates the two final designed SPMCs meet the experi- than the critical mean contact pressure, it pro-
ment specification. Therefore, not only does the SPMC- duces discontinuous indentation and generates
based critical mean contact pressure have a high sealing micro leakage paths. Meanwhile, the large

Downloaded from ade.sagepub.com by guest on August 17, 2016

Wei et al.
contact pressure variance makes contact of seal-
ing face non-uniform and leakage will occur
under high internal pressure. The mean contact
pressure and variance of final designed SPMCs
meet design requirements, which forms uniform
and continuous sealing face. The SPMCs
designed by the critical mean contact pressure
have good sealing performance.
5. The vibration experiments show that the cycle
number of vibration of the final designed
SPMCs is more than 107. The final designed
SPMCs have good ability to resist both internal
pressure and vibration.
In the next work in the soon future, we will develop
auxiliary devices and further measure the contact pres-
sure on the sealing surface by ultrasonic technology. By
comparison of analysis, simulation, measurement, and
experiment, the analysis and design method can be
improved further. Once the SPMC is used to connect
the pipelines in seabed successfully, it will strongly sup-
port offshore oil and gas exploitation.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with
respect to the research, authorship, and/or publication of this
article.
Funding
The author(s) disclosed receipt of the following financial sup-
port for the research, authorship, and/or publication of this
article: This work was supported by the National Natural
Science Foundation of China (grant numbers 51279042,
51105088) and Royal Society International Exchange
(grant ref. IE141319).
References
1. Wang LQ, Wang CD and Liu J. Error analysis and pro-
totype testing of deepwater pipe flange connection tool.
Proc IMechE, Part C: J Mechanical Engineering Science
2014; 228: 1–16.
2. Wang WM, Wang LQ and Wang CD. Deep-sea struc-
ture pose measurement’s error analysis and experimental
study. Ocean Eng 2013; 70: 141–148.
3. Zhou CF, Jiao XD, Chen JQ, et al. New underwater
joining technologies applied in offshore engineering. J
Beijing Inst Petrochem Technol 2006; 14: 20–25.
4. Östergaard I. Articulated clamp connector. Patent
6499773 B1, USA, 2002.
5. Biester K. Connection device. Patent 7469931 B2, USA,
2008.
6. Reaux JR. Pipeline connection apparatus. Patent 5456501,
USA, 1995.
7. Moghaddam AS and Mohammadnia S. Three dimen-
sional finite element analysis of 4 inch smart flange on
offshore pipeline. Ocean Syst Eng 2014; 4: 279–291.
Downloaded from ade.sagepub.com by guest on August 17, 2016
15
8. Morris BE and Avery FL. Pipeline connector for connect-
ing a branch pipe to a carrier pipe. Patent 5590913, USA,
1997.
9. Gottfried H. Coupling device for establishing a permanent
non-removable pipe connection. Patent WO 2005/068888
A1, Switzerland, 2005.
10. Feng X, Gu BQ, Sun JJ, et al. Research on sealing
mechanism of metallic gasket. Lubr Eng 2007; 32: 97–99.
11. ASME. ASME boiler and pressure vessel code, section
VIII division I, rules for construction of pressure vessels,
2007, http://www.snct.org/export/sites/default/fr/_data/
media/documents/publications/asme_asme_bilingue/
ASME_2007_BILINGUE_SECTION_VIII_DIV_1_-
_SOMMAIRE.pdf
12. DIN 2696:1999. Lenticular ring joint gaskets for flanged
joints.
13. Murali Krishna M, Shunmugam MS and Siva Prasad N.
A study on the sealing performance of bolted flange joints
with gaskets using finite element analysis. Int J Pres Ves
Pip 2007; 84: 349–357.
14. Abid M and Nash DH. A parametric study of metal-to-
metal contact flanges with optimised geometry for safe
stress and no-leak conditions. Int J Pres Ves Pip 2004; 81:
67–74.
15. ASME. ASME boiler and pressure vessel code, subcommit-
tee on design, SWG-BFJ, non-mandatory appendix BFJ,
alternative rules for bolted flange joints with ring type gas-
kets draft report. New York: ASME, 2000.
16. Sawa T, Higurashi N and Akagawa H. A stress analysis
of pipe flange connections. J Press Vess: T ASME 1991;
113: 497–503.
17. Bouzid AH and Derenne M. Analytical modeling of the
contact stress with nonlinear gaskets. J Press Vess: T
ASME 2002; 124: 47–53.
18. Sawa T and Ogata N. Stress analysis and the sealing per-
formance evaluation of pipe flange connection with spiral
wound gaskets under internal pressure. In: Proceedings of
the ASME pressure vessels and piping conference, Vancou-
ver, BC, Canada, 5–9 August 2002, paper no. PVP2002-
1089, pp.115–127. New York: ASME.
19. Takaki T and Fukuoka T. Three-dimensional finite ele-
ment analysis of pipe flange connections: the case of using
compressed asbestos sheet gasket. In: Proceedings of the
ASME pressure vessels and piping conference, Vancouver,
BC, Canada, 5–9 August 2002, paper no. PVP2002-1095,
pp.171–177. New York: ASME.
20. Bouzid A, Chaaban A and Bazergui A. The influence of
the flange rotation on the leakage performance of bolted
flanged joints. In: Proceedings of the CSME forum, Mon-
treal, QC, Canada, 27–29 June 1994, vol. 1, pp.184–194.
Ottawa, ON, Canada: CSME.
21. Haruyama S, Nurhadiyanto D, Choiron MA, et al. Influ-
ence of surface roughness on leakage of new metal gasket.
Int J Pres Ves Pip 2013; 111: 146–154.
22. Polycarpou AA and Etsion I. Static sealing performance
of gas mechanical seals including surface roughness and
rarefaction effects. Tribol T 1998; 41: 531–536.
23. Murtagian GR, Fanelli V, Villasante JA, et al. Sealability
of stationary metal-to-metal seals. J Tribol: T ASME
2004; 126: 591–596.
16
24. Bouzid AH, Diany M and Derenne M. Determination
of gasket effective width based on leakage. In: Pro-
ceedings of the ASME/JSME 2004 pressure vessels and
piping conference, San Diego, CA, 25–29 July 2004,
paper no. PVP2004-2628, pp.105–111. New York:
ASME.
25. Choiron MA, Kurata Y, Haruyama S, et al. Simulation
and experimentation on the contact width of new metal
gasket for asbestos substitution. Proc World Acad Sci
Eng Tech 2011; 75: 283–287.
Downloaded from ade.sagepub.com by guest on August 17, 2016
Advances in Mechanical Engineering
26. Robbe-Valloire F and Prat M. A model for face-turned
surface microgeometry: application to the analysis of
metallic static seals. Wear 2008; 264: 980–989.
27. Hugo Bucher H. Industrial sealing technology. 1st ed. Beij-
ing, China: Chemical Industry Press, 1988.
28. Zhao HL, Chen R, Luo XL, et al. Metal sealing perfor-
mance of subsea X-tree wellhead connector sealer. Chin J
Mech Eng 2015; 28: 649–656.
29. ISO 21329:2004. Petroleum and natural gas industries—
pipeline transportation systems—test procedures for
mechanical connectors.