Shear Pin Vibration

Journal of Sound and Vibration ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Contents lists available at ScienceDirect
Journal of Sound and Vibration

journal homepage: www.elsevier.com/locate/jsvi
Transient response characteristics of a bolted flange

connection structure with shear pin/cone
Xu Lu, Yan Zeng, Yan Chen, Xuefeng Xie, Zhenqun Guan n
State Key Laboratory for Structural Analysis of Industrial Equipment, Department of Engineering Mechanics, Dalian University of
Technology, Dalian 116023, China
a r t i c l e i n f o abstract
Article history: Transient response characteristics of a bolted flange connection structure with shear pin/
Received 13 June 2016 cone under different types of load are studied by a simplified dynamical model with
Received in revised form material nonlinearity. It is found that different ratios of the longitudinal vibration fre-
27 January 2017
quency to the bending vibration frequency lead to superharmonic response or sub-har-
Accepted 9 February 2017
monic response. The computational results indicate that the classical linear model is in-
Handling Editor: Prof. L. G. Tham
adequate for studying the transient response of the bolted flange connection structure
when the ratio of the longitudinal vibration frequency to the bending vibration frequency
Keywords: is around 2:1. Moreover, to reveal the effect of inclination angle of shear pin, a contour
Bolted flange connection
map of the transient response of the connection structure is illustrated over an angle
Shear pin/cone
range from 0° to 40°. It is found that the inclined shear pin redirects a part of transversal
Nonlinearity
Shock vibration loads to longitudinal force.
Transient response & 2017 Elsevier Ltd All rights reserved.
Coupling vibration
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2. Non-linear dynamic models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1. Simplified spring models of substructures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2. Mass-spring model for bolted flange connection structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2.1. 2DOF Mass–spring model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2.2. 3DOF mass–spring model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3. Transient response of 2DOF mass-spring system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.1. Transient response of 2DOF system under harmonic excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2. Transient response of 2DOF system under impact excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.2.1. Impact excitation of triangular waves with same peak value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.2.2. Impact excitation of triangular waves with same equivalent initial angular velocity . . . . . . . . . . . . . . . . . . . . . . 9
4. Transient response of 3DOF mass-spring system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.1. Transient response of 3DOF system under harmonic excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.2. Transient analysis of 3DOF mass-spring system under impact excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.2.1. Impact excitation of triangular waves with a same peak value. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.2.2. Impact excitation of triangular waves with same equivalent initial angular velocity . . . . . . . . . . . . . . . . . . . . . 13
5. Effects of inclination angle of shear pin on the maximum tension force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
n
Corresponding author.
http://dx.doi.org/10.1016/j.jsv.2017.02.029
0022-460X/& 2017 Elsevier Ltd All rights reserved.
Please cite this article as: X. Lu, et al., Transient response characteristics of a bolted flange connection structure with
shear pin/cone, Journal of Sound and Vibration (2017), http://dx.doi.org/10.1016/j.jsv.2017.02.029i
2 X. Lu et al. / Journal of Sound and Vibration ∎ (∎∎∎∎) ∎∎∎–∎∎∎
5.1. Maximum tension force of 3DOF system under harmonic excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

5.2. Maximum tension force of 3DOF system under triangular wave excitation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
6. Conclusion . . . . . . . . . . . . .................. ............................................................... 17
Acknowledgement. . . . . . . . . . .................. ............................................................... 17
References . . . . . . . . . . . . . . . . .................. ............................................................... 17
1. Introduction
Bolted flange connection structures are widely used in many industrial fields, such as in civil engineering, mechanical
engineering, aeronautics and astronautics engineering, nuclear industry and national defense equipments. Many of these
structures carry dynamical loads or impact loads in complex working environments. Therefore, it is very important to study
the transient response of these structures under dynamic loading in structural design [1–4].
Many analytical works have been done for establishing dynamic models of the bolted flange connection structures, and
can be roughly categorized as two different ways [5–9]. The first way is using the classical models such as Iwan or Dahl
model [10–13]. Mignolet et al. [14] developed a stochastic Iwan-type model for the characterization of a particular class of
bolted joints, and verified this model by available experimental data with broad scatter on the stiffness and dissipation
behavior. Ahmadian et al. [15] developed a reliable procedure for identifying the parameters in the Iwan model by relating
the density function to the joint interface dissipated energy. Li et al. [16,17] developed a six-parameter Iwan model to
simulate joint experiments with different preload techniques. They further used this model to describe the phenomena of
residual stiffness and power-law relationship. Piatkowski [18] presented a method for determining the parameters used in
LuGre and Dahl dynamic friction models. Wojtyra [19] found a high parametric sensitivity of Dahl model when a rigid body
approach was used to analyse the static friction in closed loop mechanisms.
Another way is developing new models for bolted flange connection structures. EI-Zahry [20] derived a realistic linear
mathematical model from simple spring-mass system analysis. Using this model, he studied the response of a bolted joint
subjected to a harmonic excitation. Duffey [21] developed a simple spring-mass model for closure bolting systems by ig-
noring viscous damping of the bolt in response analysis of single-degree-of-freedom (SDOF) mass-spring system under
impact load with exponential attenuation. Based on this model, he proposed the corresponding relationship between bolt
prestress and stress peak value under special conditions. Swansona et al. [22] established a three-dimensional solid model
of bolt connection with contact and non-linear material characteristics, and studied the influence of bolt position and flange
thickness on the stress distribution of flange. Wang et al. [23] derived a practical design model for prediction of yield and
bending of the joints, to study the bending characteristics of connections of wing panel under bending moment. Scarselli
et al. [24] proposed a simplified physical model for analyzing the hysteretic response of bolt connection under cyclic loading.
Luan et al. [25] established a flange bolt connection model with different stiffness values for tension and compression by
simplifying the whole structure as a two-degree-of-freedom (2DOF) spring-mass system.
On the other hand, the dynamic behaviors of bolt flange connection structures have been studied by numerical methods
and experiments [26,27]. Liu et al. [28] studied the dynamic behavior of bolted joints with axial excitation through ex-
periments and three-dimensional finite element simulations. Yang et al. [29] investigated the nonlinear plastic deformation
behavior of bolted joints under cyclic loading through analytical method, finite element method and experiments. Soma-
sundaram et al. [30] proposed a finite element model based on Lagrangian smooth particle hydrodynamics method to study
the physical destruction and shock wave propagation in the bolted joints. Fu et al. [31] studied the dynamical response of
connection of bolted angle steel under sudden gravity loads based on a verification model. Shibue et al. [32] studied the
relaxation behavior of bolted flange connection under impact load and analysed the effect of the friction coefficient and
initial torque. Xiao et al. [33] studied response of disk spring bolt connection structure excited by step impact loads through
theoretical calculation and finite element simulation. He found a variety of vibration state of the nonlinear system when the
width variation of input load pulse is not consistent. Shaw et al. [34] investigated the harmonic and sub-harmonic vibrations
of a SDOF oscillator. Yang [35] found that the occurrence of sub-harmonic resonance and superharmonic vibration depends
on the system parameters and high frequency disturbances.
In this paper, we will establish a three-degree-of-freedom (3DOF) mass-spring model for bolted flange connection
structure with shear pin/cone, which can transform the structural nonlinearity problem into the material nonlinear pro-
blem. Using this model, we will study the coupling characteristics of response of 3DOF simplified dynamical model. The
inheritance of dynamical characteristics between 3DOF model and 2DOF model will be confirmed by studying vibration
responses of these systems under impact or circular loads. Additionally, the contour maps of maximum tension response
with respect to the inclination angle of shear pin/cone will be illustrated by studying the redirection effects of external loads.
X. Lu et al. / Journal of Sound and Vibration ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3
Fig. 1. Bolted flange connection structure with shear pin/cone: (a) assembly diagram; (b) sectional side views of shear pin/cone [36]; (c) sectional side
views of bolted flange connection [36].
2. Non-linear dynamic models
2.1. Simplified spring models of substructures
The bolted-flange connection structures with shear pins can be divided into two types of substructures: shear pin/cone
and bolted flange (see Fig. 1).
To validate the nonlinear character of bolted flange connection, a static experiment is carried out. The geometry and
dimensions of the specimen are shown in Fig. 2(a) and (b). The cylindrical shell is built with Q235 structural steel (E ¼
00 GPa, ν¼0.28, ρ¼7800 kg m 3), and the flange are fixed by eight M10 bolts (diameter: 10mm; S45 alloy steel: E¼210 GPa,
ν ¼0.27, ρ¼7850 kg m 3). The experiment is carried out by CSS-2210 electronic universal testing machine (See Fig. 2(c)) and
the axial displacement of connection structure is measured by VIC-3D. The result of axial loading-deflection from the ex-
periment and finite element analysis are shown in Fig. 2(d). As shown in the figure, there is a significant inflection point
when the loading state changes from tension to compression. Therefore, a uni-axial bilinear spring model with different
stiffness in tension and compression can be used to simulate the bolted flange connection structure.
In addition, when the bolted flange connection structures work under high-frequency loading and high humidity en-
vironment, the coefficient of friction on the contact surfaces can be assumed to be zero [37,38]. In this case, the shear pin/
cone can be also modeled as a uni-axial bilinear spring with assumptions of linear elasticity and frictionless contact [39]. The
bilinear force-displacement relationship may be affected lightly in a dynamic context because of the existence of hysteresis
and bolt loosening which may leads force-displacement relationship to be time-varying. To focus on the dynamic char-
acteristics of bilinear force-displacement relationship system, this time-varying effect is ignored here.
As a result, two assumptions of the simplified model are given as follows:
(1) The components of connection structures are in linear elastic state and their deformations are small;
(2) Under dynamical loads, the friction in the contact interfaces between flange pairs and shear pin pairs can be neglected.
Based on these assumptions, the following approximations are obtained:
(1) The longitudinal force-deformation curve of bolted flange substructure is bilinear;

(2) Force-deformation curve of shear pin is linear when the load on the contact interface is a compressive force;
(3) Bolted flange substructures under dynamical load only resist against the longitudinal force and the transverse force is
resisted by shear pin substructures.
Obviously, when shear pin substructures contact with each other on one side, they will be separated automatically on the
other side. In the separated side of the shear pin pairs, the contact force becomes zero. In the two dimensional case shown in
Fig. 2. The experiment for verifying the assumptions of theoretical model: (a) sectional view with dimensions; (b) vertical view with dimensions;(c) static
compression-tensile experiment set-up; (d) axial loading-deflection curve of joint.
Fig. 3. The schematic of shear pin.
Fig. 4. Simplified spring system of shear pin.
Figs. 3 and 4, the stiffness of the simplified spring can be calculated from the quasi-static mechanical analysis.
The stiffness k p* of the simplified spring for the shear pin can be modeled with different compression stiffness k p − and
tension stiffness k 0 , i.e.,
⎧ k0 = 0 , δ ≥ 0
k *p = ⎨
⎩ kp − ,δ<0
⎪
(1)
where δ represents status of the spring; δ < 0 if the spring is in compression; δ > 0 if the spring is in tensile; δ = 0 if the
spring is idle.
Similarly, one can also express the stiffness kb* of the simplified spring model for bolted flange substructure by different
compression stiffness kb − and tension stiffness kb + :
⎧ kb + , δ ≥ 0
kb* = ⎨
⎩ kb − , δ < 0 (2)
The corresponding schematics of Eqs. (1) and (2) are shown in Fig. 5.
In addition, this modeling method can be expanded into the case of large displacement and/or large strain by changing
force-deformation curves from bilinear into multi-linear.
2.2. Mass-spring model for bolted flange connection structure
2.2.1. 2DOF Mass–spring model

A schematic model is composed of a rigid mass block and two same bilinear springs as shown in Fig. 6. This model has
been initially constructed in Ref. [25] to investigate impact response of the joints. The properties of the joints are char-
acterized by bilinear springs with tension stiffness kb + and compression stiffness kb − (See Fig. 5(b) and Eq. (2)). The springs
are constrained at the bottom and connected to the block at the top.
The equation of motion of 2DOF undamped mass-spring model without loads can be expressed as:
Fig. 5. Bilinear force-deformation curve of spring models: (a) shear pin for transverse loading; (b) bolted flange structure for vertical loading [36].
Fig. 6. 2DOF Mass-spring model.
⎡ b⎛ * ⎞⎤
⎢ kb*1 + kb*2 ⎜ kb2 − kb*1⎟ ⎥
⎡ M ⎤ ⎛ ν¨ ⎞ ⎢ 2 ⎝ ⎠⎥ ν
⎢
⎣
⎥ ⎜ ¨⎟ + ⎢
J ⎦⎝ θ ⎠ ⎛ ⎞
⎢ b ⎜ kb*2 − kb*1⎟ b ⎛ *
2 ()
⎞⎥ θ
=0
⎜ kb1 + kb*2 ⎟⎥
⎣⎢ 2 ⎝ ⎠ 4⎝ ⎠⎥⎦ (3)
in which, M is the mass; J is the moment of inertia; v is longitudinal displacement of point O and θ is the rotation angle; b is
* and kb2
the width of the block; kb1 * are the stiffnesses of bilinear springs. Obviously longitudinal displacement of point O and
the rotation angle are coupled in Eq. (3).
To study the difference between nonlinear spring model and linear spring model of joints, it is assumed that connection
structures with nonlinear springs and equivalent linear springs have the same base frequency of rotation. So the linear
bending stiffness of the connection structure with nonlinear springs can be derived based on the same base frequency of
rotation as follows,
kM = ωθ2 J (4)
where ωθ = 2πfθ and fθ is base frequency of rotation. And then equivalent linear stiffness of springs can be derived from the
same bending stiffness of connection structure with linear springs,
2k M
k=
b2 (5)
When fθ = 10 Hz, b ¼ 2 m, kb þ ¼1 106 N m 1, kb- ¼2 107 N m 1, the equivalent linear stiffness of spring can be
obtained for different frequency ratio as shown in Table 1.
2.2.2. 3DOF mass–spring model

To comprehensively analyze the effects of bolted flange and shear pin on the structural dynamic response, the bolted
flange connection structure is simplified to a 3DOF mass-spring system with two kinds of nonlinear springs. As two parts of
the structure connected by bolted flange coupling system are symmetric, only the upper part is selected for the dynamic
response analysis as shown in Fig. 7. The mass m and rotational inertia J relative to the centroid O1 are supported by the
bolted-flange connections modeled as two bilinear springs (kb1 * ) on two sides. Shear pin is modeled as two different
* ,kb2
* ,k p2
bilinear springs (k p1 * ) connecting to center point O of contacting surface separately for its different moving direction. All
Table 1
Equivalent linear stiffness of spring and parameter of the system for different ωv: ωθ.
ωv: ωθ m (kg) J (kg⋅m2) k̄ (N m 1)
0.5:1 5500 1300 2563496

1:1 1350 1360 2681811
2:1 340 745 1469080
3:1 150 1070 2109954
Fig. 7. 3DOF mass–spring model: (a) without loads; (b) with loads of lateral impact [36].
these springs have different stiffness for tension and compression. The lateral displacement u of point O, the longitudinal
displacement v of point O and the rotation angle θ relative to point O are also shown in Fig. 7. Obviously this model can
simulate not only transversal and longitudinal vibration but also bending vibration under loads of lateral impact.
Consider the motion equation of three degree of freedom mass-spring system derived through coordinate transformation,
⎡ M − Mr ⎤ ⎧ u¨ ⎫ ⎡ C1 ⎤ ⎧ u̇ ⎫ ⎡ k11 k12 k13 ⎤ ⎧ ⎫
⎢ ⎥⎪ ⎪ ⎢ ⎥⎪ ⎪ ⎢ ⎥ ⎪ u⎪
⎢ M ⎥⎨ v¨ ⎬ + ⎢
⎪ ⎪
C2 ⎥ ⎨ v ̇⎬ + ⎢ k21 k22 k23⎥ ⎪
⎪ ⎪
⎨ v⎬ = 0
⎪
⎢⎣ − Mr Mr 2 + J ⎥⎦ ⎩ θ¨ ⎭ ⎢⎣ C3⎥⎦ ⎩ θ ⎭̇ ⎢⎣ k k k ⎥⎦ ⎩ θ ⎭
31 32 33 (6)
where r¼a/2, elements of damping matrix and stiffness matrix are shown as follows,
⎧
⎪ Cn = 2M⋅ω n⋅ζ n = 1, 2
⎨
⎪
⎩ C3 = 2 ( Mr )
2 + J ⋅ω ⋅ζ
3 (7)
⎧ k = k * + k * cos2 α
⎪ 11 ( p1 )
p2
⎪
k
⎪ 12 = k 21 (
= k * *)
p1 − k p2 cos α sin α
⎪k = k = 0
⎪
⎪
13 31
⎨ k22 = k *p1 + k *p2 sin2 α + kb*1 + kb*2

( )
⎪
⎪ b *
⎪ k23 = k32 = 2 ( kb2 − kb1)
*
⎪
⎪ k = b2 k * + k *
⎪
⎩ 33 4
( b1 b2 ) (8)
It can be concluded from Eq. (6) that there is a coupling term of lateral and bending directions in the given symmetric
mass matrix and elements of stiffness matrix are piecewise linear functions which depend on the tension/compression state
of bolted flange or shear pin. When r ≠ 0, the system is in coupling status. If r = 0 and α = 0∘ or 90∘ , the response in the
transversal direction decouples from that in the other direction. Only the longitudinal vibration and bending vibration are
coupling as mass matrix degrades to diagonal matrix and stiffness matrix degrades to symmetric matrix.
If all the springs are linear and have the same stiffness when r ≠ 0 and α ≠ 90∘ , the stiffness matrix becomes a diagonal
matrix and mass matrix becomes a symmetric matrix. In such case, the longitudinal direction decouples from the other
directions, and only the transversal and bending directions are coupling.
3. Transient response of 2DOF mass-spring system
3.1. Transient response of 2DOF system under harmonic excitation
Considering a 2DOF mass-spring system with a linear damping ratio of ζ ¼0.001 and an amplitude of A¼3500, this
Fig. 8. Maximum tension force of bolt spring of 2DOF system under harmonic excitation.
system is loaded by a harmonic excitation of bending moment M (N. m), where

M = A cos (2πtf − π ) (9)
The numerical models are solved by the ODE solver of Matlab and the total analysis time is 1 s. The maximum force of
each calculation step is extracted for analysis.
The maximum tension force of bolt spring in the 2DOF system is calculated for different excitation frequencies f, as
illustrated in Fig. 8. In which, 2DOF-lin-Lon is used for representing longitudinal excited response of 2DOF system with
equivalent linear bolt springs, 2DOF-lin-Ben is used for representing bending excited response of 2DOF system with
equivalent linear bolt springs, and 2DOF-non-Ben is used for representing bending excited response of 2DOF system with
nonlinear bolt springs. It is found that all the peak values of the maximum tension force are 50000 N approximately for
different ratios of the longitudinal vibration frequency to the bending vibration frequency (see dash line for 2DOF-lin-Ben).
When the bolt spring is nonlinear, however, the results are quite different. As shown in Fig. 8, all the peak values of
maximum tension force with nonlinear bolt spring are much smaller than that with linear bolt spring. The nonlinearity
leads to superharmonic response as shown in Figs. 8(a) and 8(b). Moreover, it also leads to double peak response as shown
in Fig. 8(c). The latter peak value at frequency 14.3 Hz is 17.11% greater than that of the first peak at frequency 9.5 Hz.
Therefore, the design based on simplified linear model of bolt spring would be failure to predict the excitation with a
frequency of second peak and thus leads to the risk of structure.
If one changes the action direction of the harmonic excitation from bending to longitudinal. The frequencies of peak
values of longitudinal excited response and bending excited response of equivalent linear system are consistent with given
ratio of frequencies as shown in Figs. 8(a) and 8(b). As the ratios of frequencies are reciprocals with each other, order in
which peak values of longitudinal and bending excited response of equivalent linear system appears is interchanged in
Figs. 8(a) and 8(c). Considering that the equivalent linear stiffness given in Eq. (4) and Eq. (5) are derived approximately
based on the bending vibration of nonlinear system, the frequencies of the longitudinal excited response are not exactly
Fig. 9. Triangular impact with same peak value.
equal to 2 times or 3 times of which of the bending excited response, as shown in Figs. 8(c) and 8(d). It is shown that
frequencies of peak values of 2DOF-lin-Lon and 2DOF-lin-Ben are included in Fig. 8(c) but not in Fig. 8(d). It is because the
other peaks with a higher frequence cannot be predicted by the linear model. It is also found in Fig. 8(d) that the maximum
tension force of equivalent linear system under longitudinal excitation is much larger than that of a nonlinear system when
mass value of system is small. Therefore, existence of error cannot be ignored in the response prediction of structure.
3.2. Transient response of 2DOF system under impact excitation
3.2.1. Impact excitation of triangular waves with same peak value

Consider the 2DOF mass-spring system which is loaded by an impact excitation of bending moment expressed as
⎧ 1
⎪ 7000⋅t⋅ft t≥0&t≤ T
⎪ 2
M=⎨ 1
⎪ 7000 − 7000⋅t⋅ft t > T & t ≤ T
⎪ 2
⎩0 t>T (10)
where ft ¼1/T, 0.0169 s oT r1 s (see Fig. 9). This problem is analyzed using the same method shown in Section 3.1.
The maximum tension force of bolt spring in the 2DOF system under the impact excitation of triangular wave is illu-
strated in Fig. 10. Similar to Fig. 8, the peak value of the maximum tension force predicted by the linear mode doesn’t change
with the ratio of the longitudinal vibration frequency to the bending vibration frequency.
As the excitation changes from harmonic to impact, difference results are found in the response of nonlinear system. It
can be found that the response of nonlinear system is very sensitive to the initial conditions and type of excitation (see
Table 2). Particularly, the responses of nonlinear bolt spring at ωv: ωθ ¼ 0.5:1 and ωv: ωθ ¼1:1 are significantly different from
that of linear bolt spring. It is shown in Fig. 10 that there are sub-harmonic vibrations in the low frequency region for all
these cases. It is found that the peak values of maximum tension force of nonlinear bolt spring at ωv: ωθ ¼0.5:1 or ωv: ωθ
¼1:1 drift to the lower frequency region.
If the action direction of excitation changes from bending to longitudinal, the frequencies of peak values of longitudinal
excited response and bending excited response of equivalent linear system are consistent with given ratio of frequencies as
shown in Figs. 10(a) and 10(b). Considering that the ratios of frequencies are reciprocals with each other, the order in the
peak values of 2DOF-lin-Lon and 2DOF-lin-Ben appears is interchanged in Figs. 10(a) and 10(c). It is also shown in Figs. 10(b),
10(c) and 10(d) that some frequencies of the peak value of response of equivalent linear system are consistent with that of
nonlinear system only in the low-frequency region. When the excitation frequency increases, the frequencies of peak value
of response of nonlinear system cannot be predicted by the superposition of response of the linear system. It is also con-
cluded that the error cannot be ignored in the response prediction of real structure under impact excitation.
3.2.2. Impact excitation of triangular waves with same equivalent initial angular velocity
We also study the response of the 2DOF mass-spring system under another impact excitation of bending moment, as
follows
Fig. 10. Maximum tension force of bolt spring of 2DOF system under impact excitation of triangular wave.
Table 2
The difference between linear and nonlinear of 2DOF system under impact excitation of triangular wave.
ωv: ωθ FL FN FN/FL
Peak values of max- Peak values of maximum tension
imum tension force force of nonlinear system (N)
of linear system (N)
0.5:1 2664.34 2582.33 96.92%

1:1 2664.39 2657.42 99.74%
2:1 2664.37 2677.84 100.51%
3:1 2664.38 2670.51 100.23%
⎧ 2 1
⎪ 4⋅I⋅t⋅f t t≥0&t≤ T
⎪ 2
M=⎨ 2 1
⎪ 4⋅I⋅ft − 4⋅I⋅t⋅f t t > T & t ≤ T
⎪ 2
⎩0 t>T (11)
where T ¼1/ft, I is the angular impulse which is simulated by giving an initial angular velocity θ ̇ as follows,
I= ∫t M d t = Jθ ̇
(12)
1 5 8
When θ ̇ ¼0.5 rad s , ft ¼1 10 Hz and time step is 1 10 s, the the maximum tension force of bolt spring in the
2DOF system under the impact excitation of triangular wave (Fig. 11) is calculated and illustrated in Table 3 for different
ratios of frequency of longitudinal displacement of point O and frequency of rotation angle. One can find that the maximum
Table 3
Maximum tension force of bolt spring of 2DOF system under impact excitation of triangular wave with ft ¼ 1 105 Hz.
ωv: ωθ FL FN FN/FL
maximum tension maximum tension
force of linear bolt force of nonlinear bolt
spring (N) spring (N)
0.5:1 20201.93 15296.23 75.72%

1:1 21134.33 13465.99 63.72%
2:1 11577.26 12840.56 110.91%
3:1 16627.74 15892.14 95.58%
Fig. 11. Triangular impact with same equivalent initial angular velocity θ ̇ ¼0.5 rad s 1.
tension force of nonlinear bolt spring is greater than that of linear bolt spring except in the case of ωv: ωθ ¼2:1.
4. Transient response of 3DOF mass-spring system
4.1. Transient response of 3DOF system under harmonic excitation
The parameters of 3DOF System with nonlinear bolt spring or equivalent linear bolt spring are given in Table 4. If α = 0∘ and
the stiffness parameter of the shear pin is kp- ¼1 1011 N m 1 which is significantly greater than stiffness of bolt spring, the
maximum tension force of nonlinear bolt spring of 3DOF system under same harmonic excitation as shown in Section 3.1 is
calculated and illustrated in Fig. 12. It is also compared with the results of equivalent linear bolt spring.
It is shown in Fig. 12 that the maximum tension force predicted by the bolt spring of 3DOF system is similar with that of
2DOF system. As shown in Table 5, the difference between results of 2DOF and 3DOF system at α = 0∘ is very small and the
effect of shear pin on the axial response can be ignored. Therefore, when stiffness of shear pin is much greater than that of
the bolt spring, the dynamical characteristics are consistent by releasing lateral freedom of system and adding shear pin in.
Note that stiffness of shear pin is always bilinear no matter whether stiffness of the bolt spring is linear or not in all these
cases.
When α = 0∘ and the stiffness parameter of the shear pin is kp- ¼1 106 N m 1 which is similar with that of bolt springs,
the maximum tension force of the bolt spring of 3DOF system under the same harmonic excitation as shown in Section 3.1 is
calculated and illustrated in Fig. 13. It is also compared with that of equivalent linear bolt spring of 3DOF system and that of
nonlinear bolt spring of 2DOF system. As shown in Fig. 13, the frequencies of the largest peak value of the maximum tension
force of the nonlinear bolt spring and linear bolt spring are very close, except for the case of ωv:ωθ ¼ 2:1. Compared with
Table 4
Equivalent linear stiffness of spring and parameter of the system for different ωv: ωθ.
ωv: ωθ r (m) m (kg) J (kg⋅m2) k̄ (N m 1)
0.5:1 0.421 5500 324.94 2563496

1:1 0.711 1350 677.48 2681811
2:1 1.18 340 271 1469080
3:1 2.258 150 305 2109954
Fig. 12. Maximum tension force of bolt spring of 3DOF system under harmonic excitation for stiffness parameter of shear pin kp- ¼ 1 1011 N m 1.
Table 5
Peak values of maximum tension force of 2DOF and 3DOF system.
ωv: ωθ System with linear bolt springs System with nonlinear bolt springs
F2L F3L F2L/F3L F2N F3N F2N/F3N

Peak values of maximum Peak values of maximum Peak values of maximum Peak values of maximum
tension force of 2DOF sys- tension force of 3DOF sys- tension force of 2DOF sys- tension force of 3DOF sys-
tem(N) tem(N) tem(N) tem(N)
0.5:1 51945.10 50828.67 102.19% 34345.80 33667.61 102.01%

1:1 51949.48 50833.28 102.20% 29517.77 28792.37 102.52%
2:1 51490.31 50845.50 101.27% 25965.87 27999.15 92.74%
3:1 51948.47 50829.09 102.20% 39675.56 38715.13 102.48%
that of 2DOF system, all the frequencies of peak value of maximum tension force of nonlinear bolt spring of 3DOF system
drift to higher frequency region. One can also finds that the superharmonic vibration is more obvious in the case of ωv:ωθ ¼
0.5:1. As shown in Fig. 13, frequencies of peak value of 3DOF-non-Ben drift to high-frequency region by comparing with
which of 2DOF nonlinear system. However, as shown in Figs. 13(c) and 13(d), the frequencies of peak value of 3DOF-lin-Ben
are only consistent with that of 3DOF nonlinear system in the low-frequency region. If one changes the action direction of
excitation from bending to longitudinal, the frequencies of peak values of 3DOF-lin-Lon and 3DOF-lin-Ben are not consistent
with that of 3DOF nonlinear system. It is due to the effects of bilinear stiffness of shear pin which still exist in the 3DOF
system with equivalent linear bolt springs.
Fig. 13. Maximum tension force of bolt spring of 3DOF system under harmonic excitation for stiffness parameter of shear pin kp- ¼ 1 106 N m 1.
4.2. Transient analysis of 3DOF mass-spring system under impact excitation
4.2.1. Impact excitation of triangular waves with a same peak value

For a 3DOF mass-spring system with α = 0∘ and kp- ¼1 106 N m 1, the maximum tension force of the nonlinear bolt
spring and the equivalent linear bolt spring under impact excitation of triangular wave (as shown in Eq. (10)) is calculated
and illustrated in Fig. 14. The results are also compared with that of the nonlinear bolt spring of 2DOF system under the
same excitation by using time history response analysis.
As shown in Fig. 14, all the peak values of the maximum tension force of the nonlinear bolt spring of 3DOF system is
smaller than that of 2DOF system. It is due to the existence of shear pin. However, the curve shapes of them are similar with
each other in Figs. 14(a) and 14(b). It is also found that the frequencies of the peak values of the maximum tension force of
linear bolt spring of 3DOF system drift to a higher frequency region, corresponding to that of nonlinear bolt spring of 3DOF
system when ωv: ωθ ¼ 0.5:1 and ωv: ωθ ¼1:1.
If one changes the action direction of excitation from bending to longitudinal, frequencies of the peak value of 2DOF-lin-
Ben and 3DOF-lin-Lon drift into high-frequency region as shown in Figs. 14(a) and 14(b). And the frequencies of peak value
of bending excited response are similar for 2DOF and 3DOF nonlinear systems. It is also shown in Figs. 14(c) and 14(d) that
the frequencies of the peak value of the bending excited response of the 3DOF system with equivalent linear bolt springs are
consistent with that of the 3DOF nonlinear system except for middle-frequency region from 10 Hz to 20 Hz.
4.2.2. Impact excitation of triangular waves with same equivalent initial angular velocity
For a 3DOF mass-spring system with θ ̇ ¼ 0.5 rad s 1 and ft ¼ 1 105 Hz, the maximum tension force of bolt spring in the
3DOF system under the impact excitation of triangular wave of Eq. (11) is calculated and illustrated in Table 6. As shown in
Table 6, the maximum tension forces of bolt spring of 3DOF system with shear pin increase significantly in the cases of ωv:
ωθ ¼ 0.5:1 and ωv: ωθ ¼ 1:1. One can also find the maximum tension force of nonlinear bolt spring is greater than that of
linear bolt spring in the 3DOF system with shear pin with exception for the case of ωv: ωθ ¼2:1.
Fig. 14. Maximum tension response of the system under constant peak value of triangle wave excitation.
Table 6
Maximum tension force of bolt spring of 3DOF system under impact excitation of triangular wave with ft ¼ 1 105 Hz and θ ̇ ¼ 0.5 rad/s.
ωv: ωθ FL FN FN/ FL
maximum tension maximum tension
force of linear bolt force of nonlinear bolt
spring (N) spring (N)
0.5:1 9946.86 8361.78 84.06%

1:1 14511.97 12241.19 84.35%
2:1 5883.47 6872.40 116.81%
3:1 7433.77 7228.67 97.24%
5. Effects of inclination angle of shear pin on the maximum tension force
The shear pin in the bolted flange connection structures leads to shearing resistance and causes a force redirection under
the traverse load. The effect of the inclination angle of shear pin was shown in Ref. [36], and the changes of the maximum
tension force of the bolt with the inclination angle of shear pin has been elaborated when the same transient impact loads
are applied to three different points O, O1 and O2. The inclination angle is the key physical parameter which has a significant
effect on the maximum tension force of bolt spring. In this part, the influences of the inclination angle of shear pin on the
dynamic characteristics of the system are studied for various frequencies of the excitations.
5.1. Maximum tension force of 3DOF system under harmonic excitation
Here we study the response of 3DOF nonlinear system with shear pin under harmonic excitation given in Eq. (9). As
shown in Figs. 15(a) and 15(b), inclination angle of the shear pin has little influence on the maximum tension force as the
Fig. 15. Maximum tension force of 3DOF system under harmonic excitation with respect to frequency of excitation and inclination angle of shear pin.
contour line of maximum tension force tends to be parallel to the horizontal coordinate with increase of the inclination
angle. However, the peak values of maximum tension force are affected by the inclination angle of shear pin for the case of
ωv: ωθ ¼ 2:1 and ωv: ωθ ¼ 3:1. When ωv: ωθ ¼ 2:1, the maximum of peak values of maximum tension force occurs at
inclination angle near 15° as shown in Fig. 15(c). But when ωv: ωθ ¼ 3:1, there are three peaks of maximum tension force
Fig. 16. Frequency response of 3DOF system under harmonic excitation for different inclination angle of shear pin.
Fig. 17. Maximum tension force of 3DOF system under triangular wave excitation with respect to frequency of excitation and inclination angle of the shear
pin.
with respect to different values of inclination angle as shown in Fig. 15(d).

The frequency responses at the peak point 1 (11°, 29.6 Hz) and peak point 2 (31°, 25.8 Hz) are further analyzed by fast
Fourier transformation. As shown in Fig. 16, the peak values and frequencies of response are all affected by inclination angle
of shear pin especially when ωv: ωθ ¼ 2:1 or ωv: ωθ ¼ 3:1.
5.2. Maximum tension force of 3DOF system under triangular wave excitation
For a 3DOF nonlinear system with shear pin under triangular wave excitation given in Eq. (10), inclination angle of the
shear pin has little influence on the maximum tension force when ωv: ωθ ¼ 0.5:1 as shown in Fig. 17(a). The peak values of
maximum tension force occur in low-frequency stage and increases when the inclination angle of shear pin increases to 14°
(see the red region in Fig. 17(b)). With the increase of inclination angle of shear pin, zone of the maximum of peak values of
maximum tension force gradually grows to cover a wider frequency range when ωv: ωθ ¼ 2:1 as shown in Fig. 17(c). When
ωv: ωθ ¼3:1, the frequency of peak values of maximum tension force is not affected significantly by the inclination angle of
shear pin if αo22°, but it increases as inclination angle of shear pin increases further if α 4 22° as shown in Fig. 17(d).
Two points 3 (25°, 9 Hz) and 4 (25°, 24.9 Hz) signed in Fig. 17(c) are studied through fast Fourier transformation to reveal
the zone growth of maximum of peak values of maximum tension force. As shown in Fig. 18, frequencies of peak values of
response are approximately the same although triangular wave excitations are with different frequencies.
Fig. 18. Frequency response of 3DOF system under triangular wave excitation with different frequency for same inclination angle of shear pin.
6. Conclusion
In this paper, theoretical formulas of the bolted flange connection structures with shear pin are introduced. The system is
simplified as a mechanical nonlinear block-spring model. Based on this simplified model, the nonlinearity originally from
structure is replaced by the nonlinearity from material. Then the coupling feature of system is demonstrated through the
derivations and conditions for decoupling equations of motion are introduced. Dynamical behaviours of 2DOF mass-spring
model under harmonic excitation or triangular wave excitation are studied. After dynamical characteristics of 3DOF mass-
spring model are validated for stiffness of shear pin far larger than that of bolt, maximum tension force of the 3DOF
nonlinear system under three different types of excitation are illustrated for four frequency ratio of bending and long-
itudinal frequencies. Results show that superharmonic, sub-harmonic components and the vibration with two dominant
peaks appear. It is also found that the existence of shear pin makes the 3DOF systems different with 2DOF ones. The effects
of inclination angle of shear pin on maximum tension force are further studied. The influences of inclination angle of shear
pin are concluded to reveal some interesting results.
Acknowledgement
The paper supported by National Science and Technology Major Project of the Ministry of Science and Technology of
China (No. 2011ZX02403), National Natural Science Foundation of China (No.11302035 and No.11272074) and the Funda-
mental Research Funds for the Central Universities.
References
[1] P.S. Symonds, J.F. McNamara, F. Genna, Vibrations and permanent displacements of a pin-ended beam deformed plastically by short pulse excitation,
Int. J. Impact Eng. 4 (1986) 73–82.
[2] N. Jones, Structural impact for marine structures, Advances in Marine Structures-II, Elsevier Applied Science publishers, London, 1991, 251–275.
[3] G. Muscolino, R. Santoro, A. Sofi, Reliability analysis of structures with interval uncertainties under stationary stochastic excitations, Comput. Methods
Appl. Mech. Eng. 300 (2016) 47–69.
[4] Y. An, B.F. Spencer Jr., J. Ou, Real-time fast damage detection of shear structures with random base excitation, Measurement 74 (2015) 92–102.
[5] B. Blachowski, W. Gutkowski, Effect of damaged circular flange-bolted connections on behaviour of tall towers, modelled by multilevel substructuring,
Eng. Struct. 111 (2016) 93–103.
[6] A. Yapici, G. Saracoglu, Fatigue analysis of bolted flange joints of a rotary dryer, Eng. Fail. Anal. 63 (2016) 182–190.
[7] L. Zhang, Y. Liu, J. Sun, K. Ma, R. Cai, K. Guan, Research on the Assembly Pattern of MMC Bolted Flange Joint, Procedia Eng. 130 (2015) 193–203.
[8] Z. Wang, Z. Tao, D. Li, L. Han, Cyclic behaviour of novel blind bolted joints with different stiffening elements, Thin-Walled Struct. 130 (2016) 157–168.
[9] A. Nechache, A.H. Bouzid, Creep analysis of bolted flange joints, Int. J. Press. Vessel. Pip. 84 (2007) 185–194.
[10] E. Piollet, D. Poquillon, G. Michon, Dynamic hysteresis modeling of entangled cross-linked fibres in shear, J. Sound Vib. 383 (2016) 248–264.
[11] S. Sitzmann, K. Willner, B.I. Wohlmuth, A dual Lagrange method for contact problems with regularized frictional contact conditions: modelling micro
slip, Comput. Methods Appl. Mech. Eng. 285 (2015) 468–487.
[12] J.L. Dion, G. Chevallier, O. Penas, F. Renaud, A new multicontact tribometer for deterministic dynamic friction identification, Wear 300 (2013) 126–135.
[13] M. Rajaei, H. Ahmadian, Development of generalized Iwan model to simulate frictional contacts with variable normal loads, Appl. Math. Model. 38
(2014) 4006–4018.
[14] M.P. Mignolet, P. Song, X. Wang, A stochastic Iwan-type model for joint behavior variability modeling, J. Sound Vib. 349 (2015) 289–298.
[15] H. Ahmadian, M. Rajaei, Identification of Iwan distribution density function in frictional contacts, J. Sound Vib. 333 (2014) 3382–3393.
[16] Y. Li, Z. Hao, A six-parameter Iwan model and its application, Mech. Syst. Signal Process. 68-69 (2016) 354–365.
[17] Y. Li, Z. Hao, J. Feng, D. Zhang, Investigation into discretization methods of the six-parameter Iwan model, Mech. Syst. Signal Process. 85 (2017) 98–110.
[18] T. Piatkowski, Dahl and LuGre dynamic friction models — The analysis of selected properties, Mech. Mach. Theory 73 (2014) 91–100.
[19] M. Wojtyra, Properties of the dahl model applied to modelling of static friction in closed-loop kinematic chains, ROMANSY 21-Robot Des., Dyn. Control
569 (2016) 93–101.
[20] R.M. EI Zahry, Optimum design of preloaded bolted joint under harmonic excitation, J. Sound Vib. 108 (1986) 455–470.
[21] T.A. Duffey, Optimal bolt preload for dynamic loading, Int. J. Mech. Sci. 35 (1993) 257–265.
[22] J.A. Swansona, D.S. Kokanb, R.T. Leonc, Advanced finite element modeling of bolted T-stub connection components, J. Constr. Steel Res. 25 (2002)
1015–1031.
[23] Y. Wang, L. Zong, Y. Shi, Bending behavior and design model of bolted flange-plate connection, J. Constr. Steel Res. 84 (2013) 1–16.
[24] G. Scarselli, E. Castorini, F.W. Panella, R. Nobile, A. Maffezzoli, Structural behaviour modelling of bolted joints in composite laminates subjected to
cyclic loading, Aerosp. Sci. Technol. 43 (2015) 89–95.
[25] Y. Luan, Z. Guan, G. Cheng, S. Liu, A simplified nonlinear dynamic model for the analysis of pipe structures with bolted flange joints, J. Sound Vib. 331
(2012) 325–344.
[26] M. Oldfield, H. Ouyang, J.E. Mottershead, Simplified models of bolted joints under harmonic loading, Comput. Struct. 84 (2005) 25–33.
[27] H. Ouyang, M.J. Oldfield, J.E. Mottershead, Experimental and theoretical studies of a bolted joint excited by a torsional dynamic load, Int. J. Mech. Sci. 48
(2006) 1447–1455.
[28] J. Liu, H. Ouyang, J. Peng, C. Zhang, P. Zhou, L. Ma, M. Zhu, Experimental and numerical studies of bolted joints subjected to axial excitation, Wear 346-
347 (2016) 66–77.
[29] X. Yang, S.A. Nassar, Z. Wu, A. Meng, Nonlinear behavior of preloaded bolted joints under a cyclic separating load, J. Press. Vessel Technol. 134 (2012)
897–904.
[30] D.S. Somasundaram, M.B. Trabia, B.J. O'Toole, A methodology for predicting high impact shock propagation within bolted-joint structures, Int. J. Impact
Eng. 73 (2014) 30–42.
[31] Q. Fu, B. Yang, Y. Hu, G. Xiong, S. Nie, W. Zhang, G. Dai, Dynamic analyses of bolted-angle steel joints against progressive collapse based on component-
based model, J. Constr. Steel Res. 117 (2016) 161–174.
[32] T. Shibue, M. Yamamoto, A Numerical Simulation of Stress Loosening of a Bolt Connection Under the Impact Load, in: Proceedings of the Fifteenth
International Offshore and Polar Engineering Conference June, 2005.
[33] S. Xiao, B. Gu, Shock response analysis for a bolted connection system with disk spring, J. Vib. Shock 32 (2013) 101–104.
[34] S.W. Shaw, P.J. Holmes, A periodically forced piecewise linear oscillator, J. Sound Vib. 90 (1983) 129–155.
[35] J. Yang, M. Sanjuán, H. Liu, Vibrational subharmonic and superharmonic resonances, Commun. Nonlinear Sci. Numer. Simul. 30 (2016) 362–372.
[36] X. Lu, P. Wang, J. Wang, J. Chang, Y. Zhang, Z. Guan, Nonlinear behavior of bolted flange connections containing shear pin(cone), Chin. J. Comput. Mech.
32 (2015) 503–511.
[37] M. Chowdhury, M. Helali, The effect of frequency of vibration and humidity on the coefficient of friction, Tribology Int. 39 (2006) 958–962.
[38] M.A. Chowdhury, M. Helali, The effect of amplitude of vibration on the coefficient of friction different materials, Tribology Int. 41 (2008) 307–314.
[39] K. Iyer, Solution for contact in pinned connections, Int. J. Solids Struct. 38 (2001) 9133–9148.

Shear Pin Vibration

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Shear Pin Vibration

Uploaded by

Copyright:

Available Formats

Journal of Sound and Vibration ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Journal of Sound and Vibration

Transient response characteristics of a bolted flange

5.1. Maximum tension force of 3DOF system under harmonic excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2. Non-linear dynamic models

2.1. Simplified spring models of substructures

Based on these assumptions, the following approximations are obtained:

(1) The longitudinal force-deformation curve of bolted flange substructure is bilinear;

Fig. 3. The schematic of shear pin.

Fig. 4. Simplified spring system of shear pin.

2.2. Mass-spring model for bolted flange connection structure

2.2.1. 2DOF Mass–spring model

Fig. 6. 2DOF Mass-spring model.

2.2.2. 3DOF mass–spring model

ωv: ωθ m (kg) J (kg⋅m2) k̄ (N m 1)

0.5:1 5500 1300 2563496

⎨ k22 = k *p1 + k *p2 sin2 α + kb*1 + kb*2

3. Transient response of 2DOF mass-spring system

3.1. Transient response of 2DOF system under harmonic excitation

system is loaded by a harmonic excitation of bending moment M (N. m), where

Fig. 9. Triangular impact with same peak value.

3.2. Transient response of 2DOF system under impact excitation

3.2.1. Impact excitation of triangular waves with same peak value

0.5:1 2664.34 2582.33 96.92%

0.5:1 20201.93 15296.23 75.72%

4. Transient response of 3DOF mass-spring system

4.1. Transient response of 3DOF system under harmonic excitation

ωv: ωθ r (m) m (kg) J (kg⋅m2) k̄ (N m 1)

0.5:1 0.421 5500 324.94 2563496

F2L F3L F2L/F3L F2N F3N F2N/F3N

0.5:1 51945.10 50828.67 102.19% 34345.80 33667.61 102.01%

4.2. Transient analysis of 3DOF mass-spring system under impact excitation

4.2.1. Impact excitation of triangular waves with a same peak value

0.5:1 9946.86 8361.78 84.06%

5. Effects of inclination angle of shear pin on the maximum tension force

5.1. Maximum tension force of 3DOF system under harmonic excitation

with respect to different values of inclination angle as shown in Fig. 15(d).

You might also like

⎨ k22 = k p1 + k p2 sin2 α + kb1 + kb2