Fundamental Study on High Strength Bolted Tensile Joints(

Title Dissertation_ )

Author(s) Yamaguchi, Takashi

Citation Kyoto University ()

Issue Date 1996-03-23

URL http://dx.doi.org/10.11501/3110543

Right

Type Thesis or Dissertation

Textversion author

Kyoto University
 FUNDAMENTAL STUDY
ON
HIGH STRENGTH BOLTED TENSILE JOINTS

TAKASHI YAMAGUCHI

MARCH 1996
 Fundamental Study
on
High Strength Bolted Tensile Joints

A Dissertation
Submitted to
the Faculty of Engineering of Kyoto University

In Partial Fulfi1lment
of the Requirements for the Degree of
Doctor of Engineering

by

Takashi Yamaguchi

1996
 Abstract

This dissertation presents a fundamental study on high strength bolted tensile joints \'vith special attention
to the mechanical behavior such as contact/separation behavior and the joint stiffness by experimental and
analytical approach.
Firstly, the mechanical behavior of high strength bolts under static and cyclic loading ts studied
considering the bolt pre-stress force. Especially, load-dcfonnation characteristics of different cross
section such as bolt shank and bolt thread and the local stress state is investigated. In addition, the effect
of the bolt pre-stress force on fatigue life is also discussed.
Secondly, the high strength bolt and its adjacent structural element are studied with special
attention to the stifiness of such basic joint system. The contact/separation behavior of flange plates is
investigated experimentally and analytically. Furthennore, parametric study using 2-dimcnsional
axisymmetric finite element analysis is carried out and the evaluation fonnula for the stiffness of such
joints is proposed based on the multiple regression analysis.
Thirdly, the mechanical behavior on the split tee flange joints under static loading is discussed
by experimental and analytical approach. This split tee flange joint is the simplest high strength bolted
tensile joints. Here, the contact/separation behavior is investigated in detail by 3-d.imensional finite
element analysis. Furthennore, the cyclic behavior and fatigue strength which may be crucial to split tee
flange joints when applied to bridge structures, is also investigated experimentaJiy and the fatigue strength
estimation method is proposed. In addition, 2-dimensional finite element model using the effective width
coefficients is proposed as the simple analytical method useful for the practical design.
Finally, as an application of the high strength bolted tensile joints, high strength bolted tube
flange joints subjected to combined bending and tension is studied experimentally in detail. Experimental
results are assessed based on results obtained from basic studies on such as lugh strength bolts, high
strength bolt and its adjacent flange plate and split tee flange joints. Then, based on these results the
rational simple design procedure for the tube flange joints considering the effective cross sectional area is
proposed.
 Acknowledgments Table of Contents

Abstract
Tius dissertation is synthesis of the author's research work at the Department of Civil Engineering of Acknowledgments 11

Kyoto University. First of all, the author wishes to express sincere gratitude to Professor Eiichi Watanabe Table of Contents Ill

of Kyoto University, for his excellent guidance and continuous encouragement throughout course of this List of Tables VI

study If this dissertation makes a contribunon to academjc research, most of the credit should be directed List of Figures Vlll

to him. List of Photos Xll

The author also wishes to express grateful apprectat10n to Associate Professor Kurtitomo
Sugmra of Kyoto University, for his swtable advice during the course of this study and making this
manuscript better. Acknowledgment also goes to Dr. Tomoaki Utsunomiya, Instructor of Kyoto Chapter 1 Introduction 1
University, for his useful help.
In the study, a series of experiments was carried out on the specimens carefully fabricated by 1.1 General Remarks on High Strength Bolted Tensile Joints
KOBELCO Co., Ltd.. under gutdance of tvfr Shun-icruro Kasai of Kobe Steel, Ltd. A lot of technical 1.2 Classification of High Strength Bolted Tensile Joints 3
advice and his invaluable assistance and advice is sincerely acknowledged. L.3 State of the Arts on Study of High Strength Bolted Tensile Joints 4
TI1e author is also thankful to excellent students of the Structural Mechanjcs Laboratory, Kenji 1.4 Technical Problems to be Solved for Future Application 6
FuJttani, Tetsuya Matsumura Kazutoshi Nagata, Takeruro Takasuka, et al. for their assistance in writing 1.5 Objectives and Scopes 7
the manuscript. Especially, the author is thankful to Kenji Fujitanj and Tetsutya Matumuro for the References 8
collaboration in experiments.
Finally, the author expresses his deep appreciation to his parents, Mr. Katsumi Yamaguchi and
Mrs. Akiko Yan1aguchi, for thetr infinite affection and complete devotion for the completion of this Chapter 2 Mechanical Behavior of High Strength Bolts 15
dissertation. This dissertation is dedicated to them.
2.1 Introduction 15
2.2 Mechanjcal Behavior under Monotonic Loading 16
2.2.1 Experimental approach 16
2.2.2 Analytical approach 17
2.2.3 Assessment of the effective cross sectional area of the high strength bolt 18
2.2.4 Simple analysis of high strength bolt 19
2.3 Fatigue Strength 21
2.3.1 General remarks 21
2.3.2 Outline of fatigue test 21
2.3.3 Results of fatigue test and discussions 22
2.3.4 Outline of stress concentration analysis 22
2.3.5 Analytical results and discussions 23
2.4 Conclusions 24
References 25

11 Ill
Chapter 3 Mechanical Behavior of High Strength Bolts 5 2. 1 General procedure for calibration of effective \vidth coefficients 142
and Its Adjacent Structural Elements 59 5.2.2 Analytical model 142
5.3 Results and Discussions 143
3 I Introduction 59 5 3. 1 Effective width on mechanical behavior 143
3 2 Expenment on BAF model 59 5.3 .2 Calibration of effective width coefficients for the flange plate and the bolt 144
3 2. 1 Outline of the experiment 59 5.3.3 Deformation and stress verification by 2-dimensional analysis
3.2.2 Experimental results and discussions 61 using effective coefficients 144
3.3 Finite Element Analysis on BAF model 63 5.4 Conclusions 145
3.3. 1 Outline of finite element analysis 63 References 146
3.3 .2 Numerical analysis method 63
3.3.3 Parametric study for BAF model 64
3.3.4 Results and discussions 65 Chapter 6 The Application of the High Strength Bolted Tensile Flange Joint
3.3.5 Evaluation of the stiffness on BAF model 67 -High Strength Bolted Tube Flange Joint- 156
3 4 Conclusions 69
References 70 6. 1 Introduction 156
6.2 The Mechanical Behavior of the Tube Flange Joints 157
6.2. 1 Outline of the experiments 157
Chapter 4 Mechanical Behavior of Split Tee Flange Joints 92 6.2.2 Experimental results and discussions 160
6.3 Simple Design Procedure for the Tube Flange Joints 166
4. 1 Introduction 92 6.3.1 Current design procedure 166
4.2 Mechanical Behavior under Static Loading 93 6.3.2 Simple design procedure 167
4.2 . 1 Experimental approach 93 6.3.3 Results and discussions 168
4.2.2 Analytical approach 94 6.4 Conclusions and Future Needs 169
4.3 Fatigue Behavior of the Split Tee Flange Joints 100 References 170
4.3. 1 Outline of fatigue test 100
4.3 .2 Results of fatigue test 102
4.3 3 Simple estimation of fatigue strength 106 Chapter 7 Conclusions 204
4.4 Conclusions 108
References 109 Reserch Activities 207

Chapter 5 Simple Analysis on Split Tee Flange Joints

using 2-dimensional Finite Element Method 141

5. I Introduction 141
5.2 Quasi-2 dimensional Analysis on the Split Tee Flange Joints 142

I\' v
 List of Tables
Table 5. 1 Material Properties used m 2-dimensional Analysis 147

Table 5.2 Comparison of Maximum Stress

Table 2. 1 Material Properties of High Strength Bolts by specified in JIS 26 between 2-dimensional and 3-dimensional Analysis 147

Table 2.2 Dimensions of Specimens 26

Table 2.3 Slope of Load-Strain Curves at Elastic Area 27 Table 6. l Geometrical Configurations of the Specimens 171

Table 2.4 List of Ultimate Strength 27 Table 6.2 Results of Material Tests 171

Table 2.5 Numbers of Elements and Nodal Points 28 Table 6.3 Pure Bending Test Results 171

Table 2.6 Material Properties used in the Analysis 28 Table 6.4 Deformation Characteristics 172

Table 2.7 Comparison of the Stiffness between Table 6.5 Combined Loading Test Results 172

the Experimental and Analytical Results 28 Table 6.6 List of Strain at the Initial Yielding 173

Table 2.8 Comparison of the Total Stiffness (EAo) between Table 6.7 Results by Proposed Simple Design Procedure 174

the Experiment and Proposed Model 28

Table 2.9 Material Properties for Simple Analytical Model 29
Table 2. 10 Comparison ofNumbers of Elements and Nodal Points
between the Exact Model and the Simple Model 29
Table 2.11 Maximwn Stress and its Location 29
Table2. 12 L1st of Specimens for Fatigue Test 30
Table 2.13 List of Stress Ranges applied to the Specimens 31
Table 2. 14 Result of Fatigue Test 31
Table 2. 15 Number of Elements and Nodal Points of Stress Concentration Analysis 31
Table 2.16 Material Properties of Stress Concentration Analysis 31

Table 3. l Dimensions of the Specimens 71

Table 3 2 Material Propert1cs of the Specimens 71
Table 3.3 List of Analytical Cases 72
Table 3.4 Material Properties used in the Analysis 72
Table 3.5 Results of Regression 72

Table 4. 1 L1st of the Specimens 110

Table 4.2 List of Analytical Cases 110
Table 4.3 Material Properties used in the Analysis 110
Table 4.4 List of Stiffness per a Unit Length at Each Section of Each Section 110
Table 4.5 List of Loading Pattern 111
Table 4.6 Results of Monotomc Tensile Loading Test Ill
Table 4.7 Results of the Fat1gue Test 111

\I VII
 List of Figures Fig. 3.4 Schematic View of the Strain Gage for Measurement of the Bolt Force
and its Location 74

Fig. 3.5 Displacement Transducer for Measuring the Gap 75

Fig. I I Load Transferring Mechanism of the Connection using Fasteners II Fig. 3.6 Location of the Strain Gages glued on the Circular Plate and
Fig. 1.2 Typical High Strength Bolted Tensile Joints 12 Displacement Transducer 75

Fig. 1.3 Mechanism of Prying Force 13 Fig. 3.7 Load-Separation Curves of BAF Model 76

Fig. 1.4 Typical Examples of Short Connection Type 14 Fig. 3.8 Load-Bolt Force Curves of BAF Model 77

Fig. 3.9 Strain Distribution on the Circular Plate 78

Fig. 2.1 High Strength Bolt and Nut Set 32 Fig. 3. 10 Schematic Figure of Procedure of considering Boundary Non-Linearity 80
Fig. 2.2 Geometrical Configurations of High Strength Bolts 32 Fig.3.11 Example of Finite Element Discretization of BAF Model(AS-2) 81
Fig. 2.3 Loading System 33 Fig. 3.12 Boundary Conditions of BAF Model 82
Fig. 2.4 Testing Setup 33 Fig. 3. 13 Loading Procedure of the Analysis 82
Fig. 2 5 Schematic View of Measuring 34 Fig. 3. 14 Load-Deformation Curves (Analysis) 83
F1g. 2.6 Location ofFailure of the Bolt 34 Fig. 3.15 Progress of Yielding Area 85

Fig 2.7 Load-Strain Curves (Experiment) 35 Fig. 3. 16 Stiffens-Load Curves (Analysis) 86

Fig. 2.8 Analytical Model for High Strength Bolts 37 Fig. 3. 17 Pattern of Stiffness-Load Curve 88
Fig 2 9 Ftnite Element Discretization by Triangle Elements 37 Fig. 3.18 Outline of the Function 88
Fig. 2. 10 Load-Strain Curves (Analysis) 38 Fig. 3.19 Evaluation Results of the Stiffness 89
Fig. 2.11 Load-Stiffness Curves (Analysis) 39 Fig. 3.20 Comparison between the Experimental and Analytical Results 91
Fig. 2.12 Evaluation Model for the Total Stiffness of the Bolt 40
Fig. 2.13 Evaluation Results 41 Fig. 4.1 Overview of the Split Tee Flange Joint 112
Fig. 2.14 Proposed Models for Simple Analysis 42 Fig. 4.2 Geometrical Configurations of the Specimens 112
Fig. 2.15 Finite Element Discretization and Boundary Conditions of the Simple Models 42 Fig. 4.3 Testing Setup 113
Fig. 2. 16 Load-Strain Curves (Simple Analysis) 43 Fig. 4.4 Displacement Transducer for Measuring the Gap 113
Fig. 2.17 Fatigue Test Setup 44 Fig. 4.5 Strain Gage buried into the Bolt 114
Fig. 2. 18 S-N Diagram specified by Guideline of Fatigue Design 44 Fig. 4.6 Location of Measurement of Gap 114
Fig. 2 19 S-N Diagram obtained from the Fatigue Test 45 Fig. 4.7 Load-Separation Curves (Experiment) 115
Fig. 2.20 Location of the Fatigue Failure of the Bolt 45 Fig. 4.8 Load-Bolt Force Curves (Experiment) 116
Fig. 2.21 Models of Stress Concentration Analysis 46 Fig. 4.9 Overview of the Anal)'1ical Model (ST-Al) 117
Fig. 2.22 Change of Maximum Stress Concentration Factor 47 Fig. 4. 10 Example of Finite Element Discretization of tl1e Analytical Model 117
Fig. 223 Stress Concentration Factor at Each Section of the Bolt 48 Fig. 4.11 Boundary Conditions of Anal)1ical Model 118

Fig. 4. 12 Loading Procedure of Analysis 118

Fig. 3. 1 Model of High Strength Bolt and its Adjacent Flange Plate (BAF Model) 73 Fig. 4.13 Load-Separation Curves (Analysis) 119
Fig. 3.2 Testing Setup 73 Fig. 4. 14 List of Yield Strength of All Cases 120

Fig. 3.3 Schematic View of Loading 74 Fig. 4. 15 Estimated Model for Yield Strengtl1 121

Fig. 4. 16 Load-Bolt Force Curves (Analysis) 121

VIII IX
 Fig. 4.17 Deformation of the Contact Surface 122 Fig. 6.9 Measuring Points of Local Strain(CL Test) 180
Fig. 4.18 Deformation at the Center of the Model in Longitudinal Direction 124 Fig. 6.10 Dimensions of the Specimens 181
Fig.4.19 Distribution of Nodal Force on the Contact/Separation Surface 126 Fig. 6.11 Bending Moment-Curvature Curves(BL Test) 182
Fig. 4.20 Stiffness-Load Curves Obtained from the Analysis 128 Fig. 6.12 Strain Distribution at the Axial Direction(BL Test) 182
Fig. 4.21 Model for Estimation of the Stiffness of the Flange Plate 128 Fig. 6.13 Strains at the Radial Direction and at the Tangential Line Direction
Fig. 4.22 Dimensions of the Specimens for the Fatigue Test 129 at the Flange Plate(BL Test) 183
Fig. 4.23 Time History of Applied Tensile Load 129 Fig. 6.14 Principal Strains and their Directions at the Flange Plate(BL Test) 183
Fig. 4.24 Location of Strain Gages glued on the Specimen 130 Fig. 6.15 Bending Moment-Curvature Curves(CL Test) 184
Fig. 4.25 Location of the Fatigue Failure 130 Fig. 6.16 Strain Distribution of the Flange Plate(CL Test) 185
Fig. 4.26 Time History of Bolt Force and Applied Tensile Load under One Cycle 131 Fig. 6.17 Stress induced by Change of Cross Section of the Tube 168
Fig. 4.27 Change of the Maximum and Minimum Bolt Force 133 Fig. 6.18 Axial Strain Distribution of the Tube and Strain Distribution of the Bolt(CL Test) 189
Fig. 4.28 Time History of the Strain at the Bolt Shank 134 Fig. 6.19 Load-Bolt Strain Curves(CL Test) 193
Fig. 4.29 Strain Distribution on the Flange Plate 136 Fig. 6.20 Load-Bending Strain of the Bolt Curves(CL Test) 194
Fig. 4.30 Time History of the Stress on the Flange Plate 138 Fig. 6.21 Hypothesis used in the Current Design 197
Fig. 4.3 1 S-N Diagram (I) 139 Fig. 6.22 Model for Computation of Working Stress at the Flange Plate(Current Design) 197
Fig. 4.32 S-N Diagram (2) 139 Fig. 6.23 Assumption of Proposed Design Procedure 198
Fig. 4.33 Estimated Model for the Fatigue Strength of the Flange Plate 140 Fig. 6.24 Model for Computation of Working Stress at the Flange Plate
Fig. 4.34 Model of Working Stress Verification on the Flange Plate 140 (Axial Force is only applied) 200

Fig. 5.1 3-dimensional Analytical Model 148

:
Fig. 5.2 List of Analytical Cases 149
Fig. 5.3 2-dimensional Analytical Model 150
Fig. 5.4 Boundary Conditions of 2-dimensional Analytical Model 150
Fig. 5.5 Load-Separation Curves changing the Coefficient of the Flange Plate 151
Fig. 5.6 Load-Separation Curves changing the Coefficient of the Bolt 154
Fig. 5.7 Deformation of the Bolt and the Flange Plate 155
Fig. 5.8 Bending Stress-Load Curves(Bolt) 155

Fig. 6.1 Typical Types of High Strength Bolted Tube Flange Joints 175
Fig. 6.2 Test Setup for the Pure Bending Test 176
Fig. 6.3 Applied Moment Distribution Diagram 176
Fig. 6.4 Setup of the Displacement Transducer(BL Test) 177
Fig. 6.5 Measuring Points of Local Strain(BL Test) 177
Fig. 6.6 Test Setup for the Combined Loading Test 178
Fig. 6.7 Loading Procedure(CL Test) 179
Fig. 6.8 Strain Gages Glued on the Bolt Shank(CL Test) 180
 List of Photos Chapter 1
Photo 6.1 Example of High Strength Bolted Tube Flange Joint
Introduction
-Steel Erosion Control Darn- 201
Photo 6.2 Loading Controller System 202
Photo 6.3 Local Defonnat10n after Loading 203
1.1 General Remarks on High Strength Bolted Tensile Joints

Generally speaking, a structural system is assembled from individual parts or members by fastening them
together by some means. Welding is one of such means and the other is to use fasteners such as pins or
rivets and bolts. Although the fonncr has been often utilized at present time due to its recent remarkable
developments, the quality control is still very difficult to be accomplished. On the other hand, the fastener
method is a traditional connection method and can be thought to be much simpler compared with the
welding because of its relatively easy quality control. Accordingly, the fasteners have been the versatile
connection method in the past and these arc still playing an important role in the current construction
environments.
The connection using fasteners may be classified into 2 types; that is, use of pins/rivets and
bolts. For many years rivets were the means of connecting members adopted by many engineers because
of easiness of design using simple elementary mechanics in which load transferring mechanism sh0\\11 in
Fig. 1.1 (a) is assumed. However, it has become virtually obsolete because the highly skilled crews are
required and the danger of fire exists and th~ high level noise is caused in driving rivets. l11erefore, from
the view point of the labor cost and rational design of the joints, bolts superior to pins and rivets have
taken them over.
TI1e bolted connection can be further classified into 3 types such as the bearing type, the friction
type and the tension type. The load transfer mechanism of the bearing type is similar to that of pins and
rivets fasteners. The difference may be the existence of the friction between plates which is not considered
in the design because of insignificant friction resistance. If the bolt pre-stress force is introduced, the
friction resistance is found to increase, so that the friction type is proposed. The load transferring
mechanism of friction type is shown in Fig. 1.1(b). The friction type is expected to transfer the load by the
friction force resulting between two plates as shown in Fig. 1.1 (b). Because of this load transferring
mechanism, the friction type is better than the bearing type from the view point of fatigue behavior and
the stifTncss. However, the reduction of stress concentration and the increase of friction resistance
depends on the magnitude of the compressive force between two plates; that is, the high pre-stress force
have to be introduced to the bolts. This made the engineers develop the high strength bolt allowing the
higher bolt pre-stress force in order to improve better friction resistance. 1l1crcfore, the friction type has
become the most popular connection method of primary members in bridge and building structures in the
recent days.

XJI
 On the other hand, lhc tension type is also utilizing lhe contact compressiYc force between two
plates given by the bolts; it is considered to be a joint more efficiently to usc high strength bolts than the
friction type. The friction force m the friction type acts perpendicular to the contact compressive force and
its given by multiplying the friction coefficient between plates to the contact compressive force; whereas;
for the tension type the contact compressive force will directly tend to cancel out with the applied external
load. The schematic view of lhe load transferring mechanism for the tension type is shomt in Fig. 1.1 (c).
In case that the flange plate is thick enough, the tensile load applied to joints may not cause an
additional increase in the bolt force, if the applied external tensile load is smaller than the bolt pre-stress
force, the bolt force is kept constant due to the cancellation of the applied external loads by release of the
compressive force between two plates. Because of difference in the load transferring mechanism, the load
carrying capacity per a bolt of the tension type is higher than that of other types; therefore, the tension
type so called tlte high strength bolted tcnsi lc joints is considered to be tlte most superior and desired to be
used in practice.
Recently, the constructor/fabricator is being threatened by the social trend that the experienced
workers in tl1e fabrication factory and the experts in erection arc predicted to decrease. Furthennore,
construction of structures arc expected to be undertaken at lhe locations where the erection is more
dtfficult, such as mountainous district, deep underground, offshore environment because of shortage of the
land in Japan. Titcreforc, the easiness of erection is tlte most important factor for the future construction.
At the present time, the best and simples procedure may be such that structural members and complex
joint details are pre-fabricated by welding in lhe automated factory and tltcn assembled by bolts at the
construction site. The bolted connection may be preferred at the construction site to the welding.
Furthcrntorc, ductile structures are desirable to have higher energy absorption capacity fonn the
view point of the cartltquakc resistance. Because of lhe redundancy of bolted connections, the flexibility
and the energy absorption capacity can be made very high and in particular those of the tension type can
be controlled very casi ly by varying the thickness of the flange plate or the bolt pre-s tress force. Therefore,
lhc bolted connection, particularly the tension type can be implemented in the structures as energy
absorption devices in tltc future. In addition, tl1e tension type is desired from an aesthetic design point of
view. In case of friction type or bearing type of connection, nun1erous nwnbcr of heads and nuts of the
high strength bolts appears on the outer surface of the members at the connection; whereas, in case of
tensile joints, the bolts can be hidden inside the section of members, so that the surface of the connected
section is smoothen and good appearance can be achieved.
AltJ1ough the tension type has many advantages o' cr the other types of bolted connections and
welding connections, it hos not been used as the joints of primary members of bridge structures in civil
engineering field . It has been mamly used at the connection of structures constructed in the mountain area
such as the steel erosion control dom, the tmnsmission tower, the joints of the temporary facilities for
construction ,the joints of the pipe lines and auxiliary fuel tank for rockets. As for building structures,
2 3
construction procedure is based on assembly on site, therefore, such jomts have been frequently used at
the connection of beams and colunms. The main reason why it has not been used as the connection of
primary members of civil engineering structures is that the design procedure is not specified in Japanese
specifications for highway bridges(JSHB)[ I]. TI1ereforc, considering the circumstances mentioned above,
the reliable and rational design procedure for the tension type should be established as urgently as
possible and furtller study to understand the mechanical behavior of the tension type must be carried out.
1.2 Classification of High Strength Bolted Tensile Joints
In general, tensile joints are classified into two types according to the length of bolts. One type is called as
short connection type witlt short bolts as shown in Fig. 1.2(a). Two plates are connected by short bolts, so
t11at these plates arc usually stiffened by triangular rib plates. TI1e otl1cr type is called as long connection
type with long bolts and additional ring plates, namely, double flanges stiffened by rectangular rib plates.
Typical example of tltis type of joints is shown in Fig. 1.2(b).
As sh0\\11 in Fig. 1.2(a), the applied load is transferred through flange plates with bolts, where
the premature failure of flange plates are prevented by longitudinal stiffeners. Main teclutical problems of
these joints are sW1m1arized as follows: Firstly, tlte most remarkable demerit is the occurrence of the
prying force. Tite mechanism of tl1e prying force is shown in Fig. 1.3. "R" in this figure designates the
prying force. It is mainly caused by tlte defonnation of the flange plate. Titis prying force occurs as an
increase of contact compressive force at tlte outside of the bolt. Titc occurrence of tl1c prying force is
considered to be critical for the high strength bolted tensile joints because lit1le increase in bolt force can
be allowed due to the high pre-stress force initially given e.g. up to 80% of yield force of bolts. However,
in case that the flange plate is thick enough, the defonnation of tlte flange plate is negligible, so that the
prying force can be limited at very low level. In order to prevent the occurrence of such a prying force, tl1c
bending stiffness of lhe flange plates; thus, the tl1ickncss of the flange plates is the key. lltereforc, the
flange plates are usually stiffened by longitudinal rib plates in practice. As mentioned here, previous
studies on this type of joints have focused on only this prying force. Secondary, it is recognized that not
only the structural details arc very simple but also its erection procedure is very easy; tl1ercfore, it is
possible that structural members are fabricated from steel plates and pipes by cutting and welding in
automated factory and their assembly will be carried out on site. Two typical examples of the short
connection type arc shown in Fig. 1.4. One type is split tee type connection and another type is end plate
connection. Split tee type connection is the simplest among short connection types and it consists ofT-
shape members and high strength bolts. High strength bolts are located at both sides of the tee web plate.
On the other hand, end plate type connection consists of cndplatcs welded at the end of the beam and high
strength bolts. TI1e fonncr is often utilized in the United States, the latter is often used in Gennany.
 On the other hand, the long connection type is proposed because the prying force is very hard to
occur due to the structural dcta1l as shomt in Fig. I 2(b). However, the structural complexity has forced
designers not to use this type of connection. Recently it is found that the dcfonnation capacity of the joint
as well as its energy absorption capacity is considered to be high, so that this type of joints is considered
to be suitable for structures subjected to the loading with large intense. Very little study on this type of
connection has been made in the past. However, due to aforementioned better characteristics, a lot of
research activities arc on going recently. Particularly, this long connection type is used at the connection
of the suspension bridge tower, namely, Kurushima Strait Bridge under construction, and further
utilization of long connection type is expected in the future.
1.3 State of the Arts on Study of High Strength Bolted Tensile Joints
The past studies arc classified into 4 groups such as the study on short connection type, botJ1 split tee
flange joints and end plate connections, the study on long connection type and t11e study on tube flange
joints. Almost all tJ1e past studies are for short type connection and those are basically based on the
experimental results. These arc carried out for the application to building structures. However very little
study for other connections can be found.
The studies on split tee flange joints have been carried out extensively in 1960-1970's. The
main objectives of these studies were to estimate the prying force which is necessary for estimation of the
maximun1 strength of tJte joints. In Japan, B. Kato and A. Tanaka who have proposed the design formula
considering the prying force in 1968 are the first investigators on tensile joints(2]-(4]. The prying force is
estimated by the cantilever model in which the dencction of the beam and the elongation of tJ1e bolt are
taken into consideration. The amount of the prying force is varied as the thickness of the flange plate.
Moreover, in 1972, T. Hashimoto and M. Fujimoto have also proposed tJte fonnula to estimate the prying
force based on the anal)1ical results for t11c bolt and plates adjacent to tJ1e bolt[5]-l8]. Consequently, in
1975, H. Tanaka and T. Tanaka have proposed the design fonnula which is used in the current design
code for buildings in Japan[9]( 10]. In this fonnula, the estimation of the prying force is carried out by
plastic analysis on cantilever model. On the other hand, in the United States, R. T. Douty and W. McGire
proposed design fom10la using the cantilever model for estimation of the pl)ing force[ I I]. In this fonnula,
the prying force is assumed to be applied at the edge of the nange plate. Basically, these were the
experimental studies in order to investigate the prying force vs. structural parameters such as the thickness
of the flange plate, the position of the bolt, the bolt pre-stress force applied to the bolt. Although the many
modified methods to estimate pl)ing force have been proposed and various design fonnulas have been
established, there exists many ambiguity in the applicable range for structural parameters. Furthcnnore,
the various design fonnulas have been implemented with rigorous safety margin. Therefore, the general
4 5
and rational estimation method for general tensile joints should be established.
Almost all the studies on end plate connections are also based on the experimental results.
Many bending tests for the beams which have end plate connections were carried out and mechanical
behavior and the estimation of the maximum strength for this type of connections were invcstigated[12]-
[15). For example, A.N . Sherbourne carried out the bending test for bolted beam to colunm connections
and proposed the design method based on fully plastic bending moment, considering the behavior of high
strength bolt and that of end-plate in 1961. In I973 , M. Fujimoto and T. Hashimoto also carried out
bending tests for end-plate connections and investigated applicability of tJtc design fonuula for split tee
connections to end plate connection experimentally.
As for the tube flange joint, the studies were based on the experimental approach as well as
analytical approach. In 1966, K. Washio and K. Wakiyama et al. carried out the tensile loading test and
investigated the mechanical behavior considering the thickness of the flange plate and the number of the
bolt[ 16). In 1972, K. W akiyanta and S. K.ik-ukawa also carried out tJ1e tensile loading test in which
focused are the prying action, tJtat is, the effect of the prying force on tensile strength of tube joints was
investigated[ 17]. In 1979, B. Kato and R. Hirose carried out the tensile loading test and investigated load
carrying capacity under tension and proposed was the estimation method of load carrying capacity using
yielding line theory[ 18]. In addition, there is a research work on limit state design method on tube flange
joints with rib plate and without rib plates by S. Igarashi, K. Wakiyama and I. Inoue et al. [ 19](20].
Moreover, in 1992, E. Watanabe and co-workers carried out the monotonic bending test for the tube
flange joints with rib plates and "~thout rib plates and simple design method for general loading condition
was proposed(21].
On the otJter hand, as for long connection type, the key studies have not been done yet as
compared with lots of progress for short type connection. But fundamental studies have been carried out
in 1960's in the architectural engineering field[l3]. Recently, the research activity is becoming more
active in civil engineering field since 1984. T. Nishiwaki and N. Masuda ct al. have conducted a series of
the tensile loading test for long connection type and investigated its mechanical behavior in which focused
are the bolt pre-stress force, tJtiekness of the base plate and change of bolt force(22]-(31]. As a result, t11e
applicability of long type connection was found to be very high. In addition, T. Nisltiwaki and Kuroda et
al. investigated the effect of the contact surface on the mechanical behavior by tJte loading test as well as
2-dimensional finite element analyses, where the dimensions of the joints such as the position of the bolt,
t11e nuntbcr of the bolt, the lengtJt of the bolt were investigated in detail. Particularly, in 1995, the big
movement occurred in Japan; that is, the connection of the tower for suspension bridge, namely,
Kurushima Bridge was designed and being construction by tensile joints. Before this application, fmite
clement analysis was carried out in order to assess the applicability of the joints and to in\CStigatc the
mechanical behavior[32]. This construction is the first application to usc tensile flange joints as tJte
connection of primary members of bridge structures in Japan.

1.4 Technical Problems to be Solved for Future Application
According to the literature survey, the following problems arc still considered to be solved in order to
develop the rational design procedure of high strength bolted tensile flange joints.
( 1) The rational estimation of the maximun1 load carrying capacity of tensile joints considering the
prying action:
The applicable ranges of structural dimensions of the connection, such as the thickness of the
flange plate, the width of the flange plate, the distance between the loading edge and the center of
the bolt arc made not clear in the estimation of the maximwn strength in current design
specification. Because the estimation is based on tl1c experimental results and limited only for
joints with thick flange plates. Therefore, the effect of each parameter on strength estimation is
investigated in detail and estimation method applicable to wide range of parameters should be
deYeloped.
(2) The.estimation of rigidity of the connection:
ln the past studies, only an attention is paid to the load carrying capacity, not to the deformation or
rigidity of tl1c joints. However, from the view point of limit state design or energy absorption, the
rigidity and dcfonnation are important factors. For example, for accurate stability analysis on
frame structures, to know the joint rigidity is very important. In addition, flexible structures wluch
can absorb more energy at joints are desired for earthquake resistant design.
(3) Failure mode and its related strength:
It is desirable to choose the favorable failure mode by varying the geometrical configuration of the
connection. For example, if tl1e joint has thick flange plate, it will fail at the bolt whose strengtl1
and stiffness are high. On the other hand, if it has very tlunner flange plate, it will be fail at the
flange plate which is considered to be ductile. It is very important to control the failure mode from
the view point of the demand required to joints. If the high earthquake resistance is required, the
bending failure in the flange plate is desired to occur because energy absorption of this mode is
high. On the other hand, if the high load carrying capacity is required, the bolt failure should be
chosen. Therefore, it is necessary to understand the relation of structural dimensions and each
failure mode.
( 4) Investigation of mechanical behavior under cyclic loading:
In tlte application of tensile joints for civil engineering structures, the fatigue strcngtlt is very
important factor, different fom1 building structures. That is, the live load applied to the civil
engineering structures is rclati\'cly large and d)1lamically applied as comparison with those to
building structures. Ho\\c,cr, there is little infonnation available about tensile joint behavior under
6 7
cyclic and dynamic loading and there is only a report by Hirokawa ct al.[34) in civil engineering.
Therefore, the mechanical behavior under cyclic loading of this type of joints and the fatigue
strength should be investigated.
1.5 Objectives and Scopes
Tit is study focuses on the most popular tensile joints, that is, the short connection type. The objectives of
t]1is study are an investigation of the mechanical behavior of tensile joints under botl1 static and cyclic
loading in detail as well as feasibility study of its application to civil engineering structures. First of all,
t]1e mechanical behavior of the high strengtl1 bolt, which is the basic component of the joints, is
investigated considering the effective cross sectional area, in other words, the stiffness and bolt pre-stress
force. Secondly, the mechanical behavior of the split tee flange joint, which is the typical type of short
connection type, is investigated where attention is paid to the contact/separate behavior, the stiffness of
the joint. In addition, mechanical behavior of short connection type under cyclic loading is also an
objective in tlus study. Lastly, the mechanical behavior and the simple design method for the tube flange
joints are investigated as tlte application of tlte high strength bolted tensile flange joints.
Following Chapter 1 which describes the current status of tlte joint design and the applicability
of the joints, in Chapter 2, the mechanical behavior of tltc high strength bolt is discussed. llte lugh
strength bolts arc the most important component for bolted joints. Especially, the mechanical behavior
under both monotonic and cyclic loading is investigated. The loading test and finite element analysis arc
carried out for static behavior, then the effective cross sectional area and the stiffness of the bolt is
assessed by taking into consideration the bolt thread. [n addition, the fatigue test for the bolt and stress
concentration analysis by finite clement analysis arc carried out for understanding of the cyclic behavior,
then tl1e fatigue strength is assessed considering the bolt pre-stress force.
In Chapter 3, the mechanical behavior of the high strength bolt and adjacent joint clements is
discussed. The tensile loading test as well as finite clement analysis for such structural models arc carried
out. Paying attention to load transferring mechanism at joints, a contribution of bolts and adjacent plate
elements is assessed. Finally, the stifincss and its effective cross sectional area arc discussed.
In Chapter 4, the mechanical behavior of the split tee flange joint is discussed in detail as one of
the basic assembly. Both static behavior and cyclic behavior of it is investigated. The tensile loading test
and 3-dimensional finite element analysis arc carried out for static behavior, namely, the mechanical
behavior focusing on the contact surface, its stiiTncss and effective cross sectional area. In addition, as for
the cyclic behavior, the cyclic loading test is carried out by paying attention to not only tl1e overall
dcfonnation behavior of tlte split tee flange joints but also tl1e local defonnation behavior of the bolts and
the flange plates. Furthcnnore, stress concentration analysis for the split tee flange joint is also carried out,
then the fatigue strength is assessed considering the failure mode and local stresses in the joints.
It is not suitable to carry out three dimensional analysis for designing the joints of structures all
the time. Proper parameter limitation in the design specification or simple anal) sis method should be
provided for saving time in the tedious design procedure for joints. Therefore, the simple analysis of the
split tee flange joint, that is, 2-dimcnsional finite clement analysis using the effective width coefficient is
proposed and its applicability is investigated in Chapter 5.
In Chapter 6, the application of the high strength bolted tensile flange joint is discussed. For the
tube flange joint used for steel erosion control dams, the loading test is carried out subjected to bending
and tension. The mechanical behavior of such joints is discussed and compared to those of basic unit of
tensile joints such as split tee joints. In addition, the simple design method for it is proposed based on
these results.
In Chapter 7, the conclusions obtained from this study are sunm1arized and the future research
needs arc also discussed.
References
I) Japan Road Association : Specifications of Highway Bridges(JSHB), Maruzcn, 1991 (in Japanese).
2) B.Kato, A. Tanaka : High Strength Bolted Tensile Joints -lnflucncc of Bolt Pre-stress Force-, Journal
of Structural and Construction Engineering, No. 146, Apr. 1968, pp. 21-27, Architectural Institute of
Japan(in Japanese).
3) B.Kato, A. Tanaka : Experimental Study on High Strength Bolted Tensile Joints -Mechanical Behavior
under monotonic tensile loading-, Journal of Structural and Construction Engineering, No. 147, May
1968, pp. 33-4 1, Architectural Institute of Japan( in Japanese).
4) B.Kato, A.Tanaka : Experimental Study on High Strength Bolted Tensile Joints -Connection Behavior
of beam-colw1m-, Journal of Structural and Construction Engineering, No. 151 , Sep. 1968, pp. 31-38,
Architccturallnstitute of Japan(in Japanese).
5) M.Fujimoto, T.hashimoto : High Strength Bolted Tensile Joints -Part l Axisynunetric clastic analysis
for identification of initial stiffness of the plate( I)-, Journal of Structural and Construction
Engineering, No. 164, Oct. 1969, pp. 27-33, Architectural Institute of Japan(in Japanese).
6) M Fujimoto, T .Hashimoto : High Strength Bolted Tensile Joints -Part 1 Axisynunetric clastic analysis
for identification of initial stiffness of the platc(2)-, Journal of Structural and Construction
Engineering, No. 165, Nov. 1969, pp. 67-76, Architectural Institute of Japan(in Japanese).
7) M.FuJimoto, T. llashimoto : High Strength Bolted Tensile Joints -Part 2 Analysis on Split Tee
Joints( I)-, Joumal of Structural and Construction Engineering, No. 190, Dec. 1971 , pp. 59-67,
Architecturallnstitutc of Japan(in Japanese).
8) M.FuJimoto, T.Hashimoto : High Strength Bolted Tensile Joints -Part 2 Analysis on Split Tee
8 9
Joints(2)-, Journal of Structural and Construction Engineering, No. 191, Jan. 1972, pp. 7-18,
Architectural Institute of Japan(in Japanese).
9) T.Tanaka, A.Tanaka: Design Fommlas for the Strength ofT-stub Connection, JSSC Vol. 11 No. 120,
Dec. 1975, pp. 5-lO(in Japanese).
lO)Architecturallnstitute of Japan: Reconunendation for the Design Fabrication of High Strength Bolted
Joints, Maruzen, Mar. 1993(in Japanese).
11 )R.T .Douty, W .McGire : High Strength Bolted Moment Connections, Journal of the Structural
Division, Proceedings of ASCE, Vol. 91, No. ST2, Apr. 1965, pp. 101-128.
12)A.N.Shcrbourne : Bolted Beam to Column Connections, The Structural Engineer, June 1961, pp. 203-
210.
13)T.Naka, B.Kato, S.Yoshimoto : High Strength Bolted Connections -Bcam-Colw1m Connections by
High Strength Bolts-, Journal of Structural and Construction Engineering, No. 60, Oct. I 958, pp. 54 1-
544, Architectural Institute of Japan(in Japanese).
14)K.Tanaka, Mukushiro, Analysis on High Strength Bolted Tensile Joints using End-Plate Connection,
Swnrnaries of Technical Papers of Annual Meeting(Tohoku Branch), AIJ, Oct. 1973, pp. 1177-
ll78(in Japanese).
15)T.Hashimoto : Study on End-Plate Connections -Part 2 Mechanical Behavior of End-Plate
Connections, Swnrnaries ofTechnical Papers of Annual Mceting(Tohoku Branch), AlJ, Oct. 1973, pp.
ll79-1180(in Japanese).
16)K.Washio, K.Wakiyama, T. fawanari : Study on Tube Flange Joints, Swmnaries ofTcclmical Papers
of Annual Mecting(Kinki Branch), AIJ, May 1966, Structure Material Construction, pp. 185- I 88(in
Japanese).
17)K.Wakiyama, S.Kikukawa, Experimental Study on Mechanical Behavior of High Strength Bolted
Tube Flange Joints, Sununaries of Teclmical Papers of Annual Mecting(Kinki Branch), AlJ, June
1972, Structure II, pp. 157-160(in Japanese).
18)B.Kato, R.Hiorose : Bolted Tension Flange Joints of Circular Hollow Section Tubes, Journal of
Structural and Construction Engineering, No. 339, May 1984, pp. 73-83, Architectural Institute of
Japan(in Japanese).
19)S.lgarashi, K.Wakiyama, K.lnoue, T.Matswnoto, Y.Murase : Limit design of high strength bolted
tube flange joints Part I. Joint without rib-plates and ring-stiffeners, Journal of Structural and
Construction Engineering, No. 354, Aug. 1985, pp. 52-66, Architectural Institute of Japan(in
Japanese).
20)S.lgarashi, K.Wakiyama, K.Inoue, T.Matsumoto, Y.Murase : Limit design of high strength bolted
tube flange joints Part 2. Joint with rib-plates and ring-stiffeners, Journal of Structural and
Construction Engineering, No. 358, Dec. 1985, pp. 71-82, Architectural Institute of Japan(in
Japanese).
 Rivet

2l)E.Watanabc, K.Sugiura, T.Yamaguchi, S.Kasai : Design Method of High Strength Bolted Tube
Flange Joints, Journal of Structural Engineering, JSCE, Vol. 38A, Mar. 1992, pp. l-12(in Japanese).
22)K.Horie, T.Nishiwaki, N.Masuda, ct al. : High Strength Bolted Tensile Joint -Long Connection Type- Free Body of Plate A

Part I, Proc. of the 39th Annual Conference of JSCE, I, JSCE, Oct. 1984, pp. 309-31 O(in Japanese).
~----------~~=E~.~~ I
23)K.Horie, T.Nishiwaki, N.Masuda, et al. :High Strength Bolted Tensile Joint -Long Connection Type- I
Free Bodies
Part 2 Proc. of the 40th Annual Conference of JSCE, I, JSCE, Oct. 1985, pp. 929-930(in Japanese).
24)T.Nishiwaki, N.Masuda, M.Minagawa ct al. : High Strength Bolted Tensile Joint -Long Connection
Free Body of
Rivet Shank
PJU=p of Portions of Pin
Showing ShC3r Transfa-
I
I
Type- Part 3 Proc. of the 41th Annual Conference of JSCE, I, JSCE, Nov. 1986, pp. 563-564(in I p
;:r:l ~
Japanese). ~~~~-----------
Free Body of Plate B
25)T.Nishiwaki N.Masuda, M.Minagawa : High Strength Bolted Tensile Joint -Long Connection Type-
Part 4 Proc. of the 42th Annual Conference of JSCE, 1, JSCE, Sep. 1987, pp. 514-515(in Japanese).
(a) Pin I Rivet
26)M . Kuroda, T.Nishiwaki, N.Masuda ct at. : Stress Behavior on High Strength Bolted Tensile Joint -
High Strength Bolt
Long Connection Type- Proc. of the 43th Annual Conference of JSCE, 1, JSCE, Oct. 1988, pp. 560- Plate A
561(in Japanese).
27)T.Nishiwaki, N.Masuda, M.Minagawa ct al.: Bolt Axial Force of Tension-Type Connection by Long Plate B
Bolts, Journal of Structural Engineering, JSCE, Vol. 35A, Mar. 1989, pp. 991-999(in Japanese). I

28)T.Nishiwaki, N.Masuda, M.Minagawa et al. : Bolt Axial Force of Long Bolt Tension Type
Connections using Spring Model, Proc. of JSCE, No. 416/1-13, Apr. 1990, pp. 403-41 O(in Japanese).
~ ,T
1
Free Body of Plate A
I ! I
29)T.Nishiwaki, M.Kuroda, N.Masuda, Y.Suzuki : Load Transferring Mechanism of Long Bolt Tension
T Tensile Foree
Type Connections considering Contact Surface, Proc. of JSCE, No. 428/1-15, Apr. 1991, pp. 87-96(in iJT '" Frictional Resistance
11 Coefficient of Friction
Japanese).
p pT
30)T.Nislliwaki, M.Kuroda, N.Masuda : Bolt Length and Hole Alignment in Long Bolt Tension Type
Connections, Proc. of JSCE, No. 437/1-17, Oct. 1991, pp. 115-123(in Japanese).

L-rl~~-;r-------------
- p
3 l)M. Kuroda, T.Nishiwaki, N.Masuda ct al. : Experimental Study on Tension Type Connections by
Long Bolts Located at the One Side only, Proc. of the 45th Annual Conference of JSCE, 1, JSCE, Sep. 1
F rce Body of Plate A

~
1990, pp. 298-299(in Japanese).
32)H.Ohashi, Y.Yanaka, Y.Mi.wkawa, A.Umeda : Mechanical Behavior of Long Tension Type Bolted
Connections for Towers of a Suspension Bridge, Journal of Structural Engineering, JSCE, Vol. 41A, (b) Friction Type
Mar. 1995, pp. 991-100 1. Contact Compressive Force
33)K.Ohi, H.Kondo, K.Takanashi et at. : Earthquake Response Tests on Steel Frames with Semi-Rigid
Connections, Journal of Structural Engineering, AIJ, Vol. 39B, Mar. 1993, pp. 155-164(in Japanese). Bolt Pre-Stress
34)H.Yamanari, K.Ogawa, Kurobane: Inelastic Behavior of Semi-Rigid Corner Com1cetions with RHS Applied Load

Colunms and Wide Flange Beams, Journal of Structural Engineering, AIJ, Vol. 40B, Mar. 1993, pp.
703-71 O(in Japanese).
35)Y.Miki, K.Horikawa : Fatigue Behavior on Split Tee Flange Joints, Proc. of the 46th Annual
Conference of JSCE, I, JSCE, Scp. 1991, pp. 606-607(in Japanese). (c) Tension Type

Fig. 1. 1 Load Transferring Mechanism of the Connection using Fastener

10 II
 (a)
2F

Tee Web Plate

(b)

2F
(a) Short Connection Type

--
....
...
~
/
1 /~l
/
//
-......
-
c( l
l (I
.....
.....
(

~
(
--
.....

-
..... / L
I ...
~-
F F

~-
l l --...
.... (c)

~- c( (

~
(

I
{0 - 2F

.....
.....
L-V
/

~
L
'-----
v
-
.....

R : Prying Force

(b) Long Connection Type

R R
F+R F+R
Fig. 1.2 Typical High Strength Bolted Tensile Joints

Fig. 1.3 Mechanism of Prying Force

12 13
 Chapter 2

Mechanical Behavior of High Strength Bolts

2.1 Introduction

(a) Split Tee Type Connection
CD
m Among these M22 bolt is the most popular for bridge structures. The geometrical configuration of bolts
CD
(b) End Plate Connection
Fig. 1.4 Typical Examples of Short Connection Type
The high strength bolt is very important structural element for tensile flange joints. Therefore, to
understand the mechanical behavior is very important not only for tensile joints but also for general bolted
joints. The high strength bolt consists of the hexagonal head, bolt shank, bolt thread, nut, and washer as
shown in Fig. 2.1. Since it is made of high strength steel, the strength of bolt is higher up to l 00
2
(kgflmm ) compared to ordinary structural steels. It is one of the features that high strength bolts are
manufactured keeping the torque coefficient constant. The manufacturing process of the high strength bolt
is summarized as follows: At first, the bolt is fanned using low carbon alloy steel or special steel such as
chromiwn steel, chromiwu-molybdenwn steel. After that, the bolt is heated up to about 800 C for
hardening and then the bolt is further heated at about 400 C for tempering at last.
The material properties and geometrical configurations of high strength bolts and nuts are
prescribed by the specification of JIS B 1186-/979(Japanese Industrial Specification)[!]. TI1e high
strength bolt and nut set is classified by the material property and the nominal diameter. TI1e nominal
diameter is equivalent to that ofthe bolt shank. It has 3 strength grades such as F8T, FIOT and FliT. The
material property for all three grades is tabulated in Table 2.1. In Japan, FlOT is generally used for bridge
structures and strength grade F ll T is prohibited to be used due to the delayed fracture frequently occurred
in l970's[2]. On the other hand, it has 7 nominal diameters, such as Ml2, Ml 6, M20, M22, M24, M27
and M30. TI1e nwuber which follows character " M" denotes the diameter of the bolt in millimeters.
specified in JIS is shown in Fig. 2.2. As understood from shown in this figure, the shape of the bolt is very
complicated and the geometry is specified in detail. The portion between bolt shank and bolt threads
named incomplete bolt thread is only a section not specified clearly.
Since the shape of the cross section is suddenly changed at incomplete bolt thread section, the
stress concentration occur here and this has been is thought to be a cause of delayed fracture. As
mentioned above, the complexity in geometry of bolts indicates the possibility of fai lure at other high
stress concentrated part not only at incomplete bolt thread. In the past, nwncrous studies on bolts have
been carried out[3][4]. Although the effective cross sectional area is very important to understand joint
behavior, its physical basis is not clear. On the other hand, to understand the defom1ation and the stiffness
of bolts is very important to understand load transferring mechanism at joints and to design joints
considering the defonnation capacity or joint flexibility from the view point of earthquake resistance of

15
14
 slructures. Therefore, in this chapter; the mechanical bcha\'ior of bolts is investigated in detail by means
of the experimental and analytical approaches. shown in Fig. 2.6. It is observed that the fracture surface of each specimen is fine enough to conclude that
the failure is brittle. The load-average strain curves obtained from the experiment are also shown in Fig.
2.7. As the elongation depends on the stiffness of bolts, the elongation is nonnalized by a certain gage
2.2 Mechanical Behavior under Monotonic Loading length; namely, the average strain is used for estimation of the load-displacement behavior of each cross
section. In this figure, the horizontal axis shows the average strain and the vertical axis shows the tensile
2.2.1 Experimental approach load. It is understood that the strain at the bolt thread is larger than that of the bolt shank, and that the
(a) Outline of tensile loading test total strain of the bolt is larger than that of the bolt shank and smaller than U1at of the bolt thread.

The high strength bolt consists of difTcrcnt cross sectional area along bolt axis, so that the evaluation for Therefore, load-displacement curve of bolts is considered to depend on relative lengUt of Ute bolt thread.

overall bolt stiffness is very difficult. In this study, in order to estimate the overall stiffness of the high The initial tangential slopes of each load-average strain curve in Fig. 2. 7 are calculated by the

strength bolt accurately, the stiffness of the bolt shank and the bolt UU'ead are evaluated separately by a least square method and elastic properties of bolt materials are summarized in Table 2.3. In addition, the

tensile loading test. Therefore, the objective of this experiment is to evaluate total stiffness of the bolt maximum loads of each specimen are tabulated in Table 2.4. It is found Utat maximun1 load is higher Utan

taking into account the bolt tlU'cad; that is, effective cross sectional area based on the ratio of bolt UU'ead the nominal value given in the specification of JlS, so that the quality of bolt specified in JIS is satisfied.

to total bolt length between head and nut. The infonnation obtained tlll'ough this experiment is considered As shown in Fig. 2. 7, after the tensile load reached to the peak load, Ute load began to decrease aiU10ugh

to be utilized for the design of bolted joints such as tensile joints, friction type joints. U1e overall elongation increase. Therefore, it is understood U1at the elastic unloading occurs at the bolt

In this section, monotonic loading test for two types of Ute bolt which has different nominal shank; and that the severe plastic defom1ation takes place only in bolt thread.

diameter, namely, M 12 and M20 arc described. The dimensions of all the specimens are sununarized in
Table 2.2. Five specimens for each nominal diameter are prepared in tlus experiment. Since U1e bolt shank 2.2.2 Analytical approach

and bolt tlU'cad should be prepared long enough to measure the average displacement(elongation) wiU1 a (a)Outline of analytical method

certain gage length, the bolt specimens are specially manufactured with enough length for both the bolt The finite element analysis is carried out in order to investigate the mechanical behavior, especially, local

shank and the bolt tlU'ead. load-deformation relationship at different cross section such as bolt shank and bolt thread. The effect of

The tensile loading is carried out by the testing machine by SHIMADZU as shown in Fig. 2.3, various parameters on the mechanical behavior could not be made clear by only experiments due to tin1e

whose loading capacity is +/- 30 (toni) under static loading(+/- 20 (toni) under dynamic loading) and Ute and cost limitation. Furthem1ore, it is also difficult to investigate the local average stress/strain states by

displacement stroke is +/- 50 (nun). Titis is a electrically controlled closed loop hydraulic actuator. Tile only experiments due to structural complexity and size of bolts.

schematic view of the setup used in this experiment is shown in Fig. 2.4. As shown in tltis figure, hinge In this section, 2 dimensional finite element analysis is carried out in order to understand the

connections are used in order to avoid the eccentric loading. As for the measuring system used in tJ1e local stress/strain state for three types of bolts, M12, M20 and M22. Although the material is same, it is

experiment, the average elongation along the bolt shank, the average elongation within the bolt tlU'ead and considered that the difference in nominal diameter may cause different relation in inelastic state. Here, the

the total elongation between bolt head and nut arc measured separately as sho\m in Fig. 2.5. The strain bolt is modeled as the axisynuuetric problem. Dimensions of the analytical models are referred to those of

gage type displacement transducer with resolution of 500 (u/mm) such as shown in this figure is used for specimens used in tlte experiment. Namely, geometrical configurations of analytical models are

the measurement of the total displacement, and cxtcnsomctcr with gage length of 30 (mm) is used for the reproduced to be same with test specimens as much as possible. Tite shape of the bolt tlU'ead and tl1e

measurement of the elongation at the bolt tlll'cad and the elongation of the bolt shank is measured by transition area between bolt shank and bolt tlll'cad, in other words, incomplete bolt t!U'ead are also

another extcnsomctcr with gage length of I 2.5 (nun). The loading is operated by the micro computer and referred to the specific value of JlS. Boundary condition is also assumed as sho"'n in Fig. 2.8. Tensile
load is applied at the end of the bolt as an unifonn load. Furthcm10re, the displacement in the axial
control command from the computer is send to analog-controller of actuator through GP-18. Jn addition
on line measuring arc made by the same computer. direction is fixed at the bolt head. Finite element discretization by triangle elements with constant strain

(b) Experimental results and discussions for each bolt are sho\\n in Fig. 2.9. Particularly, the location considered to have a high stress
concentration such as the bolt tlll'ead and t11e bolt shank near the bolt head, is discretized by fine mesh.
All the test specimens arc failed at the bolt thread within about 10 (nun) away from the edge of the nut as
The number of elements and nodal points for each analytical model is listed in Table 2.5. Material

16
17
properties used in this analys1s is also gi,cn in Table 2.6. These values are referred to the nominal values
of high strength steel specified by JlS. Moreover, the stress-strain relationship of the material is assumed
to be clastic perfectly-plastic.
(b) Analytical results
Load-average strain curves obtained by the finite clement analysis arc shown in Fig. 2.1 0. The
comparison of the initial slope of load-average strain curves obtained by the finite element analysis to that
obtained by the experiment is shO\m in Table 2.7. The slope based on the effective cross sectional area
specified in JlS is also shown by a solid line for reference. It can be seen that anal)1ical results arc in
good agreement with the experimental results; so that it is considered that the modeling such as
discretization, boundary condition are applicable. Next, the load-stiffness curves arc shown in Fig. 2.11 .
Stiffness in these figures means the slope of the load-average strain curve and it is calculated by dividing
the increment of load by corresponding increment of average strain. It is found from these figures that the
stiffness of the bolt shank is almost same as that given by the cross sectional area of the bolt shank for all
the cases and that the stiffness of the bolt thread is lower than that obtained based on the cfTective cross
section specified by JIS. However, it is found that the overall stiffness of the bolt is little higher than that
based on the effective cross sectional area specified in J IS, and the stiffness estimation by JIS is
considered to be a little conservative. But the estimation of the effective cross sectional area by JIS may
happen to be smaller if the bolt has much longer bolt thread within clamped length, namely, distance
between the bolt head and the nul,. Furthennorc, it is fow1d that the stiffness of the bolt is kept high even
the tensile load reached to the bolt pre-stress force.
It is concluded that the efTectivc cross sectional area of the high strength bolt including the bolt
shank and the bolt thread significantly depends on the length of the bolt thread, and that cfTective cross
sectional area specified in JIS may be a little conservative if the bolt and the nut set is properly used.
2.2.3 Assessment of the effective cross sectional area of the high strength bolt
As the high strength bolt has different cross section along the longitudinal axis of bolts such as bolt shank,
bolt tltread, not like PC tendons with the unifonn cross section, it is difficult to estimate the effective cross
sectional area which represents tl1c overall stiffness of the bolt. This efTectivc cross sectional area of bolts
is very important for design of the joints from the view point of joint-defonnability and working stress
check. The accurate estimation of the stiffness of the bolt is needed for rational design of the joints. In JIS,
effective cross sectional area has been dctcnnincd based on only the strength. However, in case that the
various limit states arc considered, the estimation of the cffecti,e cross sectional area should be made
based on strength and the dcfomwtion ofjoint clement.
The model that bolt shank and bolt thread arc connected in series as shom1 in Fig. 2.12 is
utiliL.Cd for this assessment. The stiffness of each section is dctcnnined by results of the tensile loading
test, "hose values arc already listed in Table 2. 7. TI1c total stiffness of the model can be obtained from
18
the following equation.
(2.1)
in which ko. k 1, and k1 are total stiffness, stiffness of the bolt shank and stiffness of the bolt thread
respectively. Total stiffness obtained using above equation is shown in Table 2.8. It is understood that the
stiffness obtained by the proposed model and tl1at obtained from experiments are in good agreement;
therefore, it is concluded that tl1e proposed model is accurate enough to estimate the stiffness of the bolts
with various length. Comparison of tl1e efTective cross sectional area specified in JIS to the effective cross
sectional area obtained by the proposed model arc shO\m in Fig. 2.13. The effective cross sectional area
of the bolt, ~ is defined by tl1e following equation.
x+s
Ao= .....-------..
_!_ +-s- E (2.2)
(E4). (E4)2
in which, (EA)J, (EAh are Young's modulus of bolt material, slope of the load-strain curve at the bolt
shank, the slope of the load-strain curve at the bolt tl1read respectively. And x, s are the length of tl1e bolt
shank and the length of bolt t11read respectively
The horizontal axis shows tl1e length of the bolt tl1rcads or the bolt shank, tl1c vertical axis
shows tl1e effective cross sectional area which relates to the stiffness. It is found from these figures that if
the length of bolt tltreads is short compared witl1 the length of the bolt shank, the cfTective cross sectional
area is larger than t11at specification in JJS. On the other hand, it is understood that if the lengtl1 of tl1e bolt
tltreads becomes much larger, the efTective cross sectional area by JIS is smaller tl1an the actual cross
sectional area. Tilcrefore, it is concluded that if tlle bolt tltreads length is less than the 0.4 times of bolt
shank length, tl1e effective cross sectional area by JIS becomes conservative. In case of M 12 and M20,
the length of bolt thread suggested by JIS is less than 25 (nun) and 30 (nun) respectively, so that tl1e
length of bolt shank should be more than 62.5 (nun) and 70 (nun) respectively.
2.2.4 Simple analysis of high strength bolt
In general, tl1e analytical approach is considered to be very powerful for the parametric study on
engineering problems including many structural parameters, particularly for 3-dimcnsional structures. 3-
dimcnsional analysis may be required in order to investigate the structural behavior of bolted connections
in detail. But geometry of the high strength bolts is very complex such as the incomplete bolt thread and
the bolt tltread, so that the bolts have to be discrctizcd with fine mesh. TI1crcforc simple model of high
strength bolts is needed in tcnns of load-dcfonnation relation \\here tl1e shape of the bolt tl1rcad must be
taken into consideration. In this section, based on load-strain relationship obtained from the
19
 aforementioned tensile loading test, the simple modeling of the bolt is proposed.
In this study, two simple modeling arc proposed and those arc schematically shown in Fig. 2. 14. further modification such as implementation of modification factor should be made. However it is
TI1csc modeling are described in detail as follows: considered to be good enough to use these proposed modeling in order to obtain the global load-
defomtation relationship of tl1e bolt for structural analysis.

Model A : Asswue the cross sectional area of the bolt tl1rcad is same as that of the bolt
shank,
Model 8 : Assume the Youngs modulus of the bolt tl1read is same as tl1at of tlte bolt 2.3 Fatigue Strength of High Strength Bolts
shank.
2.3.1 General remarks on fatigue strength

TI1e cross section of Model A is constant along the bolt axis, so that the finite clement discretization is In tJ1is section, fatigue strength of the high strengtlt bolts is discussed. To investigate the fatigue strcngtlt

very easy. This is suitable for the analysis of the large joint system. However, clfcctivc material of tJ1e high strength bolts is important for the design of tensile joints. Because durability assessment for

properties for the bolt thread should be used. On the other hand, Model 8 is more physical because the the traffic load is very important in civil engineering structures due to large live load and long service life.

same material properties arc used for the bolt thread. In order to take into consideration the effect of stress Accordingly, tl1e fatigue strength of the bolt, which is basic component of the joint, is a necessary

concentration in bolt thread, effective cross section should be defined. TI1e material properties used in infomtation to establish a rational design procedure of the tensile joints. However, few research on the

these modeling are summarized in Table 2.9. These values arc dctcnnined by the experimental results fatigue strength of the high strength bolt can be found in civil engineering as well as architectural
described in 2.2. J. cngineering[5)[6]. In addition, especially, little study on the fatigue strcngtl1 of the bolt considering the

TI1c Yalidity of these modeling is checked by an axisymmetric model of high strength bolts. TI1e bolt pre-stress force can be found. Fatigue strength of the high strength bolt is considered to be

finite clement discretization of these models and boundary conditions are sh0\\11 in Fig. 2. 15. TI1e significantly affected by the bolt pre-stress force because the bolt pre-stress force make a certain section

boundary condition used in this analysis is same as tl1at of exact modeling used in previous section. In tl1e such as the bolt tl1read or tlte incomplete bolt tl1rcad and the bolt thread be locally yielded. Titcrcforc, the

exact modeling a shape of the bolt tl1rcad and transition area arc exactly modeled. TI1e comparison of the fatigue test for the high strength bolts is carried out and stress analysis by FEM is also done.

number of the clements and nodes between proposed simple modeling and exact modeling is shown in
Table 2. I 0. From the fact that the nwnbcr of clements and nodes of proposed simple modeling is much 2.3.2 Outline of fatigue test

fewer than that of exact modeling. Tilcrcforc, the proposed modeling for the high strength bolts is The fatigue test is carried out varying the bolt pre-stress force in order to investigate the effect of the bolt
considered to be very effective. pre-stress force on fatigue strength. TI1e specimens listed in Table 2.12(a) arc prepared for the fatigue test.

TI1e load-strain curves by the simple modeling arc sho\\11 in Fig. 2.16. In tl1cse figures, the Same as tl1e monotonic tensile loading test, tlte specimens witl1 different nominal diameter i.e. M I 2, M20

cuncs obtained by tl1c tensile loading test arc also sho\m for comparison. It is understood tl1at tl1e cunc and M22 arc considered. Dimensions of test specimens are sh0\\11 in Table 2.12(b). Material properties of

obtained by the simple modeling is in good agreement with the curve obtained by the experiment. bolts specified in JIS are also sununarizcd in this table. Test setup used in tltis fatigue test is shown in Fig.

TI1ercforc, the proposed simple modeling should be able to be applied for the analysis of structures with 2.17, which is the same as that used in the tensile monotonic loading test. Stress ranges applied in this

the high strength bolts. Next, in order to assess validity of stress distribution predicted by the simple fatigue test arc tabulated in Table 2.13. TI1cy arc determined based on tl1c guideline for fatigue design of

modeling of the bolt, the stress concentration ratios are compared with that of exact modeling. Maximum steel structurcs[7]. Tile S-N cunc specified in the guideline is shown in Fig. 2.18. Magnitude of tl1e

stress and its location arc shom1 in Table 2.1I . The maximum stress is nomJalizcd by the average stress applied stress range is constant for specimens with different nominal diameter. As mentioned above, in

of the bolt at unifonn cross scction(bolt shank). Compared with the results of exact modeling, Model 8 is order to investigate the efTect of the bolt pre-stress force on fatigue strength, the stress cycles arc applied

more appropriate than Model A as for the prediction of the location where the maximum stress occurs; by setting mean stress to the bolt pre-stress force. Here, the bolt pre-stress force arc detcnnined to be 6 I .3
(kN) according to the specification of JIS[ 1]. In addition, 30.7 (kN) which is 0.5 times of the bolt pre-
however, it is due to the geometrical discontinuity at bolt shank and bolt threads. Moreover, both
modeling arc not sufficient for estimation of the magnitude of the maximunt stress. As a result, it is stress force is also considered. Titis fatigue test is carried out by load control. Before tlte fatigue test, the

thought that the estimation of stress distribution using proposed simple modeling is not satisfactory and test loading is carried out in order to obtain tJ1e maximum frequency in which the load-time curve can be
reproduced by the sinusoidal cune, and then, its frequency is dctennined to be 2.0 (Hz) according to the

20
21
 loading capac1ty of the testing machine. However, it is dctcmuned to be 0 5 Hz in case of the large stress
range. Loading procedure is sununarizcd as follows: At first, the prcscnbed bolt pre-stress force is given The geometry of tl1e anal)1ical model is the same as that of the specimens use in t11e fatigue test in

to the high strength bolt; then, the stress cycles applied in the sinusoidal wave fom1 with specified stress previous sections. TI1e shape of the bolt t11rcad m t11c anal)1ical model is referred to the specification of
range. JIS. As for the discretization, the location considered to ha\e a high stress concentration, such as tl1c bolt
tllfead and bolt shank near tl1e bolt head, is discretized by fine mesh. TI1e nwnber of tl1e elements and
2.3.3 Results of fatigue test and discussions degrees of freedom for each cases are swmnarized in Table 2. 15 respectively. Material properties used in

The fatigue life for each stress range obtained from the test is tabulated 111 Table 2.14 and plotted in Fig. this analysis are also given in Table 2.16, which arc dctcnnincd based on t11e specification of JIS and

2.19. The design S-N curve specified in the guideline is also shown for a reference. All the specimens arc stress-strain relationship of t11e material is asswned to be clastic perfectly plastic.

failed at the bolt thread ncar the nut as shown in Fig. 2.20. Comparing with the result of monotonic
loading test, the location of failure in the fatigue test is much closer to the nut. It is observed that the 2.3.5 Analytical results and discussions

failure surface is very fine, so that tl1e brittle failure is considered to occur. It is found from this figure tJ1at Maximwn stress concentration factor of the bolt and its location obtained from the analysis are shown in

the fatigue life of all the cases are longer than that given by the guideline for fatigue design. TI1ereforc, Fig. 2.21. Here, stress concentration factor K is defined to be the ratio of the actual stress to the nominal

fatigue life estimated by tl1e guideline is considered to have enough safety margin, namely, fatigue life stress as follows;

given in the guideline is considered to be very conservative. Moreover, compared witJ1 tJ1e bolt which has
different nominal diameter, there exist little difference on fatigue life. Therefore, it is concluded that (2.3)

fatigue life is not significantly affected by the difference in geometrical configurations of bolts as long as
if the bolt configuration satisfies JIS. in which K, <7, ,<7eq are the stress concentration factor, the nominal stress and the equivalent stress

Comparing M12-A witl1 M 12-D, M 12-B with M12-E, M20-A with M20-D, M20-B witJ1 respectively. TI1e nominal stress <7, in this consideration is the stress by dividing the applied tensile load

M20-E, it can be seen that tl1c fatigue life of the high strength bolt is moderately affected by tJ1e bolt pre- by the effective cross sectional area of the bolt specified in JlS.

stress force. In case that 50 % of bolt pre-stress force dctcm1ined by JIS is given, it is found that the For all the cases, it is found that the maximum stress concentration is about 3.0, and that it

fatigue life is 1.2 times longer than that in case that I 00 % of bolt pre-stress force is given. It may be occurs at the transition area between the bolt shank and the bolt thread; namely, incomplete bolt tl1read. It

thought that the higher bolt pre-stress force makes the fatigue life shorter because the severe stress is understood that higher stress compared with the nominal stress exists, therefore, the bolt pre-stress

concentration at the bolt thread and the bolt shank ncar the bolt head make bolts yielded locally. In force may cause a local yielding. Next, the stress concentration factor at the bolt shank ncar the bolt head,

addition, it is found from this figure that the slope of design S-N curve is different from that obtained in incomplete bolt thread and the bolt tltrcad are compared in Fig. 2.23. In this figure, the order of the stress

this fatigue test. This also concludes tl1at the slope of S-N curve is also alTected by the level of the pre- concentration factor of the each location is also shown in Fig. 2.23. It is found that there is almost no

tension force. In the future, farther fatigue test is required to set the reliable design S-N curve considering dilTercnce on stress concentration at the bolt tl1rcad and at the incomplete bolt tl1read among the bolts
bolt pre-stress force. which have different nominal diameters. But the difference of stress concentration exists at tlte bolt shank
ncar the bolt head; namely, stress concentration factor of M22 is 30 % smaller than that of M 12 at the

2.3.4 Outline of stress concentration analysis bolt shank near tl1e head. As for tl1e stress state of the bolt when tl1e prescribed pre-stress force( 61.3 kN)

As mentioned in the previous section, the local stress concentration might affect the fatigue strengtJ1 of by JIS is given, incomplete bolt thread and the bolt thread is locally yielded for all the cases, but the bolt

bolts. Since the fatigue test carmot provide actual stress distribution or maximum stress, stress analysis is shank ncar the head is not yielded for M20, M22, but yielded for M 12.

carried out by axiS)lnmctrie finite clement modeling for bolts in order to investigate the physical aspect of From tl1csc anal)1ical results, it is found that the location where the shape of cross section is

the fatigue strength based on local stress concentration. The finite clement model is shomt in Fig. 2.2 1 changed, such as bolt shank ncar the bolt head, incomplete bolt tltrcad and the bolt tltread is yielded. It is

where the washer is omitted for simplicity. In addition, boundary conditions used in this analytical model concluded that the fatigue strength of the pre-tensioned high strength bolt is considered to be alTccted by

and finite clement discretization by triangle clements with constant strain arc also shown in Fig. 2.21 . the bolt pre-stress force significantly.
In the future, compared with the fatigue test results, the failure occurred at the bolt tltrcad ncar
Three cases taken into consideration for dilTcrent nominal diameter such as M 12, M20, M22 are prepared.
the nut. But the stress concentration at the bolt tltrcad obtained from the stress concentration analysis is

22
23
not so high due to the lack of nut modeling. Titercforc, in order to assess the actual phenomena, the
fundamental infonnation of the joints with high strength bolts.
modeling including the nut section in the finite clement analysis may be required.

References
2.4 Conclusions
I) Japanese Industrial Standard Committee : Sets of High Strength Hexagon Bolt, Hexagon Nut and
Plain Washers for Friction Grip Joints(B 1186), 1979.
In this chapter, the mechanical behavior of the high strength bolts under both static and cyclic loading arc
2) Japanese Society of Steel Construction, Joint Subconmtittec, Strength of bolts Working Group
investigated experimentally and anal)1ically. In addition, the simple modeling of high strength bolts
Delayed Fracture of High Strength Bolt, JSSC, Vol. 15, No. 158, Mar. 1979.
considering the cfTective local-elongation relation of each section such as the bolt shank and the bolt
3) A. Hashimoto : Mechanical Properties of Fl OT High strength Bolts Subjected to Direct Tension,
thread is proposed, and its applicability is investigated. TI1e following conclusions and future research
Journal of Structural and Construction Engineering, No. 309, Nov. 1981, Architectural Institute of
needs are obtained:
Japan.
4) K.Wakiyama, K.Hirai : A Study on Fatigue of High Strength Bolt, Journal of Structural and
I) When the tensile load is applied to the high strength bolts, strain of the bolt thread is larger than
Construction Engineering, No. 288, Feb. 1980, Architectural Institute of Japan(in Japanese).
that of the bolt shank, and total strain of the bolt is larger than that of the bolt shank and smaller
5) Y.Miki, K.Horikawa : Fatigue Behavior on Split Tee Flange Joints, Proc. of the 46th Annual
than that of the bolt thread. Namely, it is understood that severe plastic defonnation take place
Conference of JSCE, 1, JSCE, Sop. 1991 , pp. 606-607(in Japanese).
in bolt thread and that load-displacement curve of the bolt depends on relative length of the bolt
6) K.Wakiyama, K.Hirai : A Study on Fatigue of High Strength Bolt, Journal of Structural and
thread.
Construction Engineering, No. 288, Feb. 1980, pp. 21-27, Architectural Institute of Japan(in
2) TI1c effective cross sectional area specified in JIS has considerably high safety margin in
Japanese).
general usc, but in case of the bolt \\~th long bolt thread, the effective cross sectional area
7) Japanese Society of Steel Construction : Recommendation for Fatigue Design, Apr. 1993(in
specified in JIS may become critical from the view point of the stiffness of the bolts; however
Japanese).
the ductility could be improved.
3) TI1e simple finite clement modeling of the high strength bolts useful for the joint system analysis
is proposed based on the load-dcfonnation behavior of the bolt shank and bolt thread, and it is
verified that this modeling is effective for investigation of load-dcfom1ation behavior of bolted
joints. In the future, for the stress verification, this simple modeling should be modified.
4) Fatigue failure of high strength bolts occurs at the bolt thread near the nut. Fatigue strength of
high strength bolts giYcn by the guideline for fatigue design of steel structures is considerably
conservative, and it is significantly affected by bolt pre-stress force. When bolt pre-stress force
is applied to the bolt, the bolt has already yielded locally at the bolt thread and incomplete bolt
thread.

In the future, in order to estimate the fatigue strength quantitatively, further fatigue test and parametric
stress analysis varying the shape of the bolt and bolt pre-stress force should be required. Moreover, the
analysis including the nut section is expected for accurate consideration of failure mode of the bolts.
Especially, based on these results, relationship between stress concentration factor and fatigue strength
should be made clear. Furthcnnorc, the results obtained in this chapter arc expected to be used as the

24 25
 Tb
a le 21M a ten' a IP roperttes o fH !&I
' h S trenf!tI1 BoI t spec1'lied
1 In. JIS
Classification Yield Stress Ultimate Stress Elongation Contraction of
(kgf/nun 2) (kgf/mm 2) (%) Area(%) Table 2.3 Slope of Load-Strain Curve at Elastic Area
F8T more than 64 80-100 16J;).J: more than 45
FIOT more than 90 100-120 l4J;J. J: more than 40 (a) M20
FliT more than 95 110-130 14btJ: more than 40 Specimen Bolt Shank Bolt Tiliead Total Young's Modulus
(toni) (toni) (toni) (Bolt Shank)
(~flnuu )
2

Table 2.2 Dimensions of Specimens M20-0l - - - -

(a)Design Dimensions M20-02 6200 5014 5799 19730
(unit nun) M20-03 6900 4531 4462 21960
Type of Bolt Diameter of Length of Bolt shank Length of M20-04 6395 5934
4904 20350
Bolt Shank and Bolt TI1read Bolt TI1Iead M20-05 7013 4710 7783 22320
(dl) (I) (s) Mean 6627 4790 5995 21090
MI2 12
M20 20 Coefficient of 0.05115 0.03859 0.1972 0.05122
210 115
M22 22 Variation

(b) T a bleofMeasurmg s
JmenSIOnS 0 fth e >pec1mens (b)Ml2
Specimen B c dll dl2 dl3 diO dl d2
- d2 d3 do H I s
Specimen Bolt Shank Bolt Tiliead Total Young's Modulus
Ml2-l 26.80 23.40 12.00 11 .95 12.00 11.98 (tonf) (toni) (toni) (Bolt Shank)
12.00 - 12.00 11.95 11.98 I 0.10 199.80 116.15
(k_g_f/nun2)
Ml2-2 26.75 23.35 11.95 11.95 12.05 11.98 12.00 - 12.05 12.00 12.02 10.20 199.60 115.65
M12-3 26.75 23.35 11.95 12.00 12.05 12.00 12.00 - 12.00 11.95 11.98 10.25 199.60 115.15
M12-01 - - - -
M12-4 26.75 23.35 12.00 11.95 12.00 11.98 12.00 - 12.00 11.95 11.98 10.00 199.95 115.90
M12-02 - - - -
M12-5 26.80 23.35 11.90 11.95 12.00 11.95 M12-03 2402 1814 1746 21240
12.00 - 12.05 11.95 12.00 10.20 199.90 115.00
M12-04 2458 1820 1858 21730
M20-l 33.90 29.60 19.90 19.85 19.80 19.85 19.90 19.90 19.35 19.90 19.90 13.10 199.70 114.50
M20-2 M12-05 2534 . 1600 2226 22400
33.65 29.50 19.90 19.80 19.80 19.83 19.85 19.85 19.15 19.85 19.85 13.40 199.25 115.25
M20-3 Mean 2465 1745 1943 21790
34.00 29.55 19.95 19.85 19.80 19.87 19.80 19.85 19.15 19.80 19.82 13.05 199.90 115.60 Coefficient of 0.02195 0.05864 0.1055 0.02182
M20-4 33.75 29.45 19.85 19.80 19.80 19.82 19.85 19.85 19.20 19.85 19.85 13.50 199.55 114.60 Variation
M20-5 34.20 29.50 20.00 19.90 19.85 19.92 19.80 19.80 19.15 19.85 19.82 13.25 199.95 115.25
M22-l 35.90 31.35 22.60 22.05 22.00 22.22 22.05 22.00 21.05 21.85 21.97 15.15 199.40 114.15
M22-2 35.80 31.30 22.60 22.20 22.10 22.30 22.10 22.00 21.10 21.80 21.97 Table 2.4 L.1St 0 fUI.lunate stref!.g_ll
tl
M22-3 35.85 3 1.35 22.40 21.10 21.95 22.15 21.95 22.10 21.15 21.80 21.95 Specimen Ultimate Strength Specimen Ultimate Strength
M22-4 35.90 3 1.30 22.40 21.95 21.95 22.10 21.95 22.00 2l.l5 21.80 21.92 (tonf) (toni)
M22-5 36.10 31.30 22.55 21.95 22.00 21.17 21.95 22.00 21.05 21.80 21.92 M12-0I (9.6480) M20-0I (27.9720)
Note: diO=(d11+d12+d13)/J, do=(d1+d2+d3)/3 Ml2-02 - M20-02 27.6180
Bolt Head Bolt Shank Bolt Thread Ml2-03 9.7560 M20-03 27.7860
M12-04 9.6720 M20-04 27.8640
Ml2-05 9.6560 M20-05 27.8640
Mean 9.695 Mean 27.783
Coefficient of 0.004524 Coefficient of 0.003615
Variation Variation
~k-------------~~----------------~
H Minimwn Tensile 8.43 Minimwn Tensile 24.5
Stre~h Strength
Transition Area

26
27
 . fior stmp Ie An al)'
Table 2.9 Material Properties I r( ICa I Mode I
Table 2 5 Numbers of Elements and Nodal Points Bolt Yielding Stress Young's Modulus at Cross Sectional Area
Analytical Model Nwnber of Number of Number of Type Bolt Thread at Bolt Thread
Elements Nodal Points Degrees of a c n 2
cry (kgf/nun ) (kgf/nun 2) A (mm2)
Freedom MI2 Method I O.OOI56 100.6 0.04I6 76.8 15,422 Il3.1
Ml2 4,418 2,687 5,374 Method 2 0.00115 139.6 0.0383 109.9 21,790 80.2
M20 3,867 2,387 4,774 M20 Method I O.OOI98 I06.9 0.0476 79.5 15,245 314. 1
M22 3,867 2,387 4,774 Method 2 0.00116 I42.5 0.0392 109.3 21,090 228.1

Table 2.10 Comparison of Numbers of Elements and Nodal Points

Tabl e 26M at en.aI Propert1es us ed.Ill 111e Analysis
between the Exact ModeI an d tI1e sunp Ie Mode I
Young's Poisson's Ratio Yielding Stress
Modulus Ml2 M20
(kgflnun 2) (kgf/nuu 2) Number of Number of Number of Number of
21,000 0.3 95.0 Nodal Points Elements Nodal Points Elements
Exact Model 2,699 4,418 2,387 3,868
Method 1 Model 253 402 345 582
T abl e 2 7 Companson of t11e sItaness betwcen tI1e E Method 2 Model 278 427 370 607
xpcnmental Results and Analytical Results
Bolt Type Bolt Shank Bolt 11uead Total Young's Modulus
(toni) (toni) (toni) (kgflnuu 2)
MI2 Experiment 2,465 1,745 I,943 2,I790 Table 2 II Maximum Stress and its Location
Analysis 2,367 1,582 2,043 2,0927 Exact Model Method I Metl1od 2
M20 Ex peri men t 6,627 4,790 5,995 2,1090 Stress Concentration Ml2 2.6705 0.8613 1.0829
Analysis 6,578 4,905 5,599 2,0938 Factor M20 2.9246 0.7888 1.2284
Location where Maximum TA BH TA
Stress is Arisen
Table 2.8 Comparison of the Total StifTness(EAo) between tl1e Experiment and Proposed Model
[NOTE]
Bolt Type Experiment Proposed Model Relative Error
TA: Transition Area between Bolt Shank and Bolt Tiuead
Ml2 5,995 5,561 7.2% BH: Bolt Shank near Bolt Head
M20 1,943 2,033 4.6%

unit : tonf

28 29
 Table 2.13 List of Stress Ranges AppliI cd 10 I he S>J>CCimens
Specimen Bolt Pre-stress Load Range Stress Range Bolt Pre-stress aJab Loadmg
Force a, (jp Period
Table 2. 12 List of Specimens for Fatigue Test 1
(kgf) (kgf) (kgf/mnh (kgf/mm ) (Hz)
( a) D es1gn
. d.llllCnSIOnS 74.3 0.41 2
M12-A 6,260 2,520 30
Bolt Diameter Length of Length of Yielding Effective Minimum Standard One Half of 6,260 1,680 20 74.3 0.27 2
M12-B
Type of Bolt Bolt Shank Threaded Stress Cross Tensile Bolt PI 74.3 0.14 2
M12-C 6,260 840 10
d, and Portion Sectional Strength Pre-stress P2=0.5P 1 0.41 2
M12-D 3,130 2,520 30 37.1
Bolt Thread Area Force
M12-E 3,130 1,680 20 37.1 0.27 2
I PI
(mm) s(mm) (kgflmm 2) (mm 2) (kgf) (kgf) M20-A 18,200 7,350 30 74.3 0.41 2
(mm) (kgf)
M20-B 18,200 4,900 20 74.3 0.27 2
Ml2 12 84 8,430 6,260 3,130 10 74.3 0.14 2
M20-C 18,200 2,450
M20 20 200 I 15 90.0 245 24,500 18,200 9,100 9,100 7,350 30 37.1 0.41 2
M20-D
M22 22 303 30,300 22,600 11,300 M20-E 9,100 4,900 20 37.1 0.27 2
M22-A 22,600 9,090 30 74.3 0.4 1 0.5
M22-B 22,600 6,060 20 74.3 0.27 0.5
(b)Mcasuring Dimensions M22-C 22,600 3,030 10 74 .3 0.14 0.5
(unit nun) M22-D 11,300 9,090 30 37.1 0.41 0.5
Specimen B c dll dn dl3 diO dl d2
- d2 dl do H I s M22-E 11 ,300 6,060 20 37. 1 0.27 0.5

M12-A 26.75 23.35 11.95 11 .95 12.00 11.97 12.00 - 12.00 12.00 12.00 10.25 199.70 116.20
M12-B 26.80 23.35 11 .95 12.00 12.00 11 .98 12.00 - 12.00 12.00 12.00 10.00 200.00 114.95
M I2-C 26.75 23.40 12.00 12.00 12.00 12.00 12.05 - 12.00 12.00 12.02 10.25 200.00 115.65 Table 2 14 Result of Fatigue Test (Number of Cycle to Failure)
M12-D 26.75 23.35 11 .95 12.00 12.05 12.00 12.00 - 12.05 11 .95 12.00 10.20 199.70 114.70 Specimen A B c 0 E
M12-E 26.80 23.45 11 .95 11.95 12.00 11 .97 12.00 - 12.00 11.95 11 .98 10.20 200.00 114.10 M12 23,913 68,688 not failed 35,589 123,580
M20-A 33.90 29.60 19.90 19.80 19.80 19.83 19.80 19.80 19. 10 19.80 19.80 12.85 200.00 115.65 M20 26,516 89,5 13 676,074 32, 132 122,459
M20-B 33.80 29.50 19.90 19.80 19.85 19.85 9.85 19.85 19.20 19.80 19.83 13.00 200.00 118.60 M22 25,975 59,562 - - -
M20-C 34.00 29.60 19.90 19.80 19.80 19.83 19.80 19.80 19.15 19.80 19.82 13.00 199.75 114.65 [NOTE] - : Expenment was not earned out.
M20-D 33.90 29.60 19.85 19.80 19.85 19.83 19.85 19.85 19.10 19.80 19.83 13.00 200. 10 116.45
M20-E 33.90 29.40 19.90 19.80 19.85 19.85 19.85 19.85 19.10 19.85 19.85 13.00 199.75 11 6.85
M22-A 35.90 3 1.30 22.40 22.00 22.00 22 13 22.05 22.00 21.15 21 .80 21.95 15.45 200.60 117.95 c
Table 2.15 Number of Elements and Nodal Points ofS tress oncentral10n AJ1a I>sis
M22-B 35.90 31.40 22.50 22. 10 22.05 22.22 2 1.95 22.00 21.10 21.85 21.93 14.60 200.50 116.60 Analytical Model Nwubcr Number Number
M22-C 35.85 31.35 22.40 22.00 22.00 22.13 2 1.95 21.95 21.20 21.85 21.92 of of of Degrees
M22-D 35.90 31.35 22.45 22.05 22.05 22.18 22.05 22.00 21.10 21.90 21.98 Elements Nodal Points of Freedom
MI2 4,418 2,687 5,374
M22-F 36.00 31.35 22.50 22.00 22.00 22.17 22.00 22.00 2l.IO 21.85 21.95
M20 3,867 2,387 4,774
Note: dw=(d,, +d, 2+d 13)/3, do=(d,+d2+d3)/3
M22 3,867 2,387 4,774
Bolt Head Bolt Shank Bolt Thread

c
. I propcrt1es ofS tress oncen t ra t 1011 Analvs Is
T able 2. 16 Malena
Young's Modulus Poisson's Ratio Yielding Stress
2
(kgf/nun ) (kgf/nun2)

21,000 0.3 95.0

Transition Area

31
30
 Washer Washer
Bolt Thread

Hexagon Head Bolt Slmak Hexagon Nut

Fig. 2. 1 High Strength Bolt and Nut Set

Bolt Head

c"O
Sf

Transition Area
Dimensions of High Strength Bolt Specified by JIS(JIS B 0205)
dl H B c s
Q) Hydraulic Pump Testing Machine
Bolt Type Standard Tolerance Standard Tolerance Standard Tolerance About Standard Tolerance
@Display @) Micro Computer
(d) Length Length Length Length
Ml2 12 +0.7 8 +0.8 22 0 25.4 25 +5 @ Printer Controler

-0.2 -0.8 -0.8 0

M20 20 +0.8 13 +0.9 32 0 37 35 +6 Fig. 2.3 Loading System
M22 22 -0.4 14 -0.9 36 -1 41.6 40 0
(unit: mm)
Tensile Load
Actuator

om1enSJOI1S 0 fB otI Th rcad S>PCCI'tied

I b>y JIS (JIS B 020-)
)

Pitch Diameter Diamter Diameter

(P) (01) (D) (d)
Ml2 I. 75 10.106 12.000 4.91 Tensile Load
Loading Frame
M20 2.5 17.294 20.000 8.467
M22 2.5 19.294 22.000 9.467 Fig. 2.4 Testing Setup
(unit : nun)
Fig. 2.2 Geometrical Configurations of High Strength Bolt 33
32
 12 12

10 10

8 8
cc cc
g 6
.B
6
"0
111
:0
111
0 0
...J ...J

4 4 - - Bolt Shank
- - Bolt Shank
-o- Bolt Thread -o- Bolt Thread
_._Total _ .....,_ Total
2 2

0 W---~--~--~--~~--~--~ 0 ----~----~r-~---.---.--
0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035
Strain Strain

(a) Ml2-03 (b) Ml2-04

12

Displacement 10
Displacement Transducer
Transducer for Threaded Portion 8
for Total Displacement Bolt Shank Bolt Thread
cc
.B
:0 6
111
0
...J

Fig. 2.5 Schematic View of Measuring 4

- - Bolt Shank
-o- Bolt Thread
2 _._Total

o.---.---.--.---.---.--~~
Bolt Head 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035
Strain

(c) Ml2-05
Bolt Shnak

Fig. 2.7 Load-Strain Curves (Experiment) (M I 2) (continued)

Fig. 2.6 Location of Failure of the Bolt

35
34
 r
Bolt Head
30 30 Appli ed
Tensile Load
25 25
F

c 20 20 Bolt Shank
c cc
.8
:0 0
co 15
0
~ 15
_J (1l
0
...J
Fig. 2.8 Analytical Model for High Strength Bolts
10 10
- . - Bolt Shank - BoltShank
---<:r Bolt Thread ---<:r Bolt Thread
5 _...._Total 5
__._ Total
!Onun
I<E-7i
0--------~----~--~---.--~ Oir-----.----,-----.----.--------,---~

! I ~~~~ t j
0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.000 0.005 0.010 0.015 0.020 0.025 0.030
Strain Strain

(a) M20-02 (b) M20-03

k :1 I( 115mm

200mm

30 30
(a) Ml2
25 25
13nun

u
I<E---7!

u
20 20
cc cc
0 0
.; 15 .; 15
(1l (1l

~
0 0
_J ...J

K 115nun
10
- BoltShank
10
- BoltShank I( 200nun
---<:r Bolt Thread ---<:r Bolt Thread
5 5 _...._ Total
- .t.- Total (b) M20
15mm
0 -------,-----.----~--~---.--~ 0 @---~---.--~----~--~--~ I( ) I
0.000 0.005 0.010 0.015 0.020 0.025 0.030

U
0.000 0.005 0.010 0.015 0.020 0.025 0.030

r~
Strain Strain

l
(c) M20-04 (d) M20-05

Ilolt Shank Bolt Thread

I( Jl5mm J
200mm
(c) M22

Fig. 2.7 Load-Strain Curves (Experiment) (M20)

Fig. 2.9 Finite Element Discretization by Triangle Elements

36
37
 X 103 X 103
3.0 -,----------....,..1----r--, 8 .----------------------,--------..,...,
9 25

-- -----..-~e~t~ W'" H-J...wi-

I
8
I 7
7 20
2.5
--4..----~-...-.~- - 6
6 ................ ~
I --~~~ I
2.0
Is
.,
'E
0
~
-o
15
C'
c "''"T'"
-o
4 g 1.5
0
.3 10 _._ Bolt Shank

~
_J
_._ Bolt Shank -o- Bolt Thread
3 -o- Bolt Thread - .a.- Total
3
- BoltShank - BoltShank 1.0 -"'- Total EA. 1
2 -o- Bolt Thread 5 -o- Bolt Thread ........... EA. 1 - - EA,.,..
__.__ Total __.__ Total -- EA.,., 2 - - EA...,
0.5 _ .. _ EA..., - - Pre-stress Force
o.----.------,---,.-----.-----.------l o.---.--.--.r--~----.--~ - - Pre-stress Force - Minimum Tensile Strength
0.000 0.000 0.002 0.003 0.004 0.005 0.006 0.000 0.000 0.002 0.003 0.004 0.005 0.006 - Minimum Tensile Strength

Strain Strain
0 1 2 3 4 5 6 7 8 9 0 5 10 15 20 25
Load (tonf) Load (tonf)
(a)Ml2 (b)M20
(a) Ml2 (b) M20

X 103
25 1 0 .------------~~----~--~

.. ....-.....
I
20 Bolt Shank I
~- ---
Bolt Thread
cc
.2
8 ---
:;- 15 .. ~..~.A "'IIrA
. ..........,I!+~...
:g .......................~~t ~ ... ~ :::.~~
_J

10
C' 6 f-.. - ....

~~
c
0
- BoltShank ~ .-------~1
-o- Bolt Thread _._ Bolt Shank Bolt Shank
5 __.__Total -o- Bolt Thread I
4
-"'- Total I (
Threaded Portion
........... EA.w
0~-------.---.---.---.--~ I
0.000 0.000 0.002 0.003 0.004 0.005 0.006
-- EA,., I
2 - .. - EA..., I
Strain - - Pre-stress Force I
- Minimum Tensile Strength 1

(c)M22 0 +---.---.---r--.__J,~--~------l
0 5 10 15 20 25 30 35
Load (tonf)
Fig. 2.10 Load-Strain Curves (Analysis)
(c) M22

Fig. 2.11 Load-Stiffness Curves (Analysis)

38 39
Bolt Head Nut Bolt Head Nut

Bolt Shnak Bolt Shnak

115 320
.... N'
E 110 x 2 = 15.0
E
s s 300 x2 =15.0
"'~
<(
105
"'
~
<(
;;; 100
1 1 1 c: ;;; 280
c:
-=-+- u
0 0
n
ko kt k2 Q)
(f)
95 Ql
260

--- ___l______ _
(f)
C/)
C/)
e 90 la
e
k0 : Effective Stiffness of the High Strength Bolt
(.)
Ql
>
o
85 ______ l _______ _ (.)
Q)
>
240

.21 ~ 2
k 1 : Stiffness of the Bolt Shank Specification (A" = 245 mm
w
..
80 Specification (Aen = 84.3 mm
2
)
~ 220
w
)

k2 : Stiffness of the Bolt Thread ..

<(
0

oT x 2 : Length of Boll Thread (mm)

' I ' I
J <(
0

OJ x2 : Length of Bolt Thread (mm) J

0 20 40 60 80
Fig. 2.12 Evaluation Model for the Total Sti1Iness of the Bolt 0 20 40 60 80
x1 =Length of Boll Sh~nk (mm) x 1 : Length of Bolt Shank (mm)

(i) MI2 (ii) M20

(a) Length of Bolt Shank vs. Effective Cross Sectional Area

115 320
N' x 1 : Length of Bolt Shank (mm) N'
E E
s 110
s 300

.;c"' 105 "'~

Q)

<(
;;; iii 280
c: 100 c:
0 0
'-"'
0
Ql 95
uQl
(f)
(f) 260
la la
e
(.)
90 e
(.) 240

--------r-------
Ql
Ql
.2: 85 >
0
~
u
Q)

w 80 2 :z:: 220
Specificalion (84.3mm ) w
..
<(
0

oT J 0
<(
OJ x, : Length or Bolt Shank (mm) J
i I
0 20 40 60 80 0 20 40 60 80 100
x 2 : Length or Bolt Thread (mm) x 2 : Length of Bolt Thread (mm)

(i) Ml2 (ii) M20

(b) Length of Bolt Thread vs. Effective Cross Sectional Area
Fig. 2.13 Evaluation Results

40 41
 Bolt Head 10 10
& -'{J----
~ . . c-.

8 I 8
f .
c- I . c-
c Lines: Experiment c Lines: Experiment
~p 0 6
Method I ~
0 6 I : Dots : Method 1 ~ Dots : Method 2
Bolt Head -o
(13
i . - - Bolt Shank
"0
(13
- - Bolt Shank
0
....J
i .. ... .. Bolt Thread
0
....J 4 Bolt Thread
Bolt Shank Bolt Thread 4
- - Total - - Total
0 Bolt Shank Bolt Shank
2 0 Bolt Thread 2 Bolt Thread
1:;. Total Total
Method 2 0
0
0.000 0.005 0.010 0.015 0.000 0.005 0.010 0.015
Fig.2.14 Proposed Models for Simple Analysis
Strain Strain

(a) M 12 -Method 1 (b)M20-Method 2

(a) M 12 Method I

L:-~E(l~~~ 30 . , . . . . - - - - - - - - - - - - - - - - - , 30
..c a-
vr:Ja; :coo
a-iD -o- cc-
(b) Ml2 Method 2 25 25
fO
20 20 I .
c- Lines ; Experiment
c Dots ; Method 1
c-
c I Lines ; Experiment
0 I Dots ; Method 2
:;; 15 - - Bolt Shank ~ 15
(13
.. ..... Bolt Thread
"0
(13
I - - Bolt Shank
0
(c) M20 Method I ....J
- - Total
0
....J I./ Bolt Thread
10
o Bolt Shank
10 j/ - - Total
I 0 Bolt Shank
o Bolt Thread
.z
5 0 Bolt Thread
t:.. Total 5
"' Total
~
0 o-~~~-r~--~~-~~~ 0 ~~---~--~--~~~~~
(d) M20 Method 2 0.000 0.005 0.010 0.015 0.000 0.005 0.010 0.015
r Strain Strain
(c) M12 -Method 1 (d)M 20-Method 2

C -JI Applied

~--
Tensile Load
o-.. F Fig. 2. 16 Load-Strain Curves (Simple Analysis)
E
Bolt Shank
Axisof A r---~--~--~--~~~--~~~~~~~~---4)
Symmetry z
(c) Boundary Condition
Fig. 2. 15 Finite Element Discretization and Boundary Conditions of the Simple Models

43
42
 8000
Actuator Tensile Load 7000 o Pre-tension Force :~P
6000
5000 o Pre-tension Force : P2
N' 4000 Specification

~ 3000 oo M12
;;:::: M20
Hinge
~ 2000 M22
a>
0>
c Not Failed
co
~@ 1~~~
700
~

U5 600
P1=6.26 tonf
500
400 P =3.13 tonf
300
Load Cell 1~ 1~ 1~
Number of Cycle to Failure
Loading Frame
Tensile Load
Fig. 2.19 S-N Diagram obtained from the Fatigue Test

Fig. 2.17 Fatigue Test Setup

1000 .---------------------------, Bolt Head

Bolt Shnak
~b
100
~
Cl)
0)
c:
fr. Fig. 2.20 Location of the Fatigue Failure of the Bolt
::l
Cl)
10
l:s
(/)

1 (MPa) = 1 (kgffcm 2 )

106 107 108

Cycle

Fig. 2.18 S-N Diagram Specified by Guideline of Fatigue Design

44 45
 -----------
............ -Transition Ar~"'a--- ..........

,I

I
't---~-___;,~~~ r~~-F-K T-~____.:~

(a) Ml2

IOmm
!<E-71 ---------------

n I.e~'''- ~ I I 5mm
t
j 5
u
~
c
~
3.5 . , - - - - - - - - - - - - - .

3.0

2.5
Pre-stress Force
{18200kgl)
................Transition Area----..............
. ....
..
''

~ ~
2.0
200mm

(a) M12 ~ 1.0 +---------:--~~

u;
0.5

E 0.0 +----.-----.---~-......---~
E 5000 10000 15000 20000 25000
0
00 Load {kgl)
-
V'l

I< 115mm
(b) M20

200nun ------------- ...

3.5 , - - - - - - - - - - - - - , .............. Transition Area---.......... ......
(b) M20
3.0
Pre-stress Force
(22600kgl)
......
.
~
~ 25 \ .
''

12.0 \
c

~
u
1.5 ' '--

~~ 1.0 ~------------...:!.....--j -~
~
(/)

0 .5

0.0 +--..,---,---,.---..--.--~
0 5000 10000 15000 20000 25000 30000
Load (kgl)
(c) M22
(c) M22

Fig. 2.21 Models of Stress Concentration Analysis Fig. 2.22 Change of Maximwn Stress Conccnlralion Factor

47
46
 Bolt Shank ncar the Bolt Head Incomplete Threaded Portion Tensile Load= 5840 kgf
(a) (b)
Bolt Shank near Bolt Head Incomplete 1lu-caded Portion 1lu-cadcd Portion

j Threaded Portion (c)

j
Nodal
Nwnlx:r
Stress
Concentration
Factor
Nodal
Nwnlx:r
Stress
Concentration
Factor
Nodal
Nwnlx:r
Stress
Concentration
Factor
10 1.5700 3 2.6799 4 2.0641
7 1.2770 10 2.4517 10 2.0641
Bolt Shank Bolt Thread 8 1.1382 1 1.6790 2 1.3216
9 1.0834 13 1.5675 13 1.3216
High Strength Bolt 5 1.0414 2 1.5435 3 1.2137
6 0.8334 4 1.4607 12 1.2133
1 0.8080 7 1.4272 6 1.2060
4 0.7237 5 1.4239 8 1.2060
2 0.6755 8 1.3967 7 l.l666
3 0.6188 12 1.3801 9 l.l666
9 1.3308 14 1.1419
14 1.2376 I 1.1415
11 0.9718 5 0.9317
6 0.9074 II 0.9317

Tensile Load= 6260 kgf

Bolt Shank ncar Bolt Head Incomplete 1lrrcaded Portion Threaded Portion
Nodal Stress Nodal Stress Nodal Stress
Nwnber Concentration Nwnlx:r Concentration Nwnber Concentration
(b)1lu-cadcd Portion Factor Factor Factor
(a) Incomplete 11\I'Cadcd Portion
lO 1.2793 8 1.2828 4 1.2794
7 1.2544 9 1.2827 10 1.2794
9 l.l531 10 1.2827 6 1.2792
8 l.l059 3 1.2824 8 1.2792
5 1.0225 4 1.2820 7 1.2785
6 0.8147 5 1.2809 9 1.2785
1 0.8109 I 1.2800 2 1.2747
4 0.7253 2 1.2799 13 1.2746
2 0.7079 7 1.2796 3 1.2722
3 0.6550 13 1.2796 12 1.2720
12 1.2794 14 1.0315
14 1.0854 I 1.0313
II 1.0416 5 0.9433
6 1.0372 II 0.9433

(c) Bolt Shank ncar the Bolt Head

Fig. 2.23 Stress Concentration Factor of Each S<:tiion of the Bolt (M 12-2) (continued)
Fig. 2.23 Stress Concentration Factor of Each Section of the Bolt (M12-I)(continucd)

48 49
 Tensile Load= 7520 kgf Bolt Shank near the Bolt Head Incomplete Threaded Portion
(a) (b)

I
Bolt Shank ncar Bolt Head Incomplete Threaded Portion lltreadcd Portion Threaded Portion (c)
Nodal
Nwnbcr
Stress
Concentration
Factor
Nodal
Nwnbcr
Stress
Concentration
Factor
Nodal
Nwnbcr
Stress
Concentration
Factor
I
10 1.0651 8 1.0685 4 1.0652
7 1.0650 9 1.0685 6 1.0652
9 1.0650 10 1.0679 8 1.0652 Bolt Shank Blot Thread
8 1.0597 3 1.0675 10 1.0652 High Strength Bolt
5 1.0035 4 1.0674 7 1.0646
I 0.8228 5 1.0665 9 1.0646
6 0.7820 I 1.0656 3 l.O<H5
2 0.7808 2 1.0655 12 1.0644
4 0.7469 13 1.0655 2 1.0611
3 0.7350 7 1.0653 l3 1.0610
12 1.0653 14 1.0022
6 1.0627 l 1. 0018
14 1.0571 5 0.9607
11 1.0560 11 0.9607

Fig. 2.23 Stress Concentration Factor of Each ~tion of the Bolt (Ml2-3) (continli<Xi)

(a) Incomplete Threaded Portion (b)Tirreaded Portion

(c) Bolt Shank ncar the Bolt Head

Fig. 2.23 Stress Concentration Factor of Each Section of the Bolt (M20-1) (continued)

50 51
 Tensile Load = 16975kgf Tesile Load = 18200 kgf

Bolt Shank ncar Bolt Head Incomplete lll!cadcd Portion Threaded Portion Bolt Shank ncar Bolt Head Incomplete Threaded Portion lltreadcd Portion
Nodal Stress Nodal Stress Nodal Stress Nodal Stress Nodal Stress Nodal Stress
Nwnbcr Concentration Number Concentration Nwnber Concentration Nwnber Concentration Nwnber Concentration Nwnbcr Concentration
Factor Factor Factor Factor Factor Factor
10 0.5906 9 2.9238 7 2.0488 10 0.5910 9 1.2839 19 1.2784
8 0.2844 8 2.9064 18 2.0485 8 0.2845 8 1.2837 8 1.2779
5 0.2026 14 2.7709 5 1.3838 5 0.2027 16 1.2818 20 1.2778
4 0.1999 16 2.4507 21 1.3834 4 0.1999 3 1.2809 9 1.2776
9 0. 1978 3 2.0599 9 1.3230 9 0.1979 12 1.2807 7 1.2772
6 0.1735 I 1.7592 20 1.3230 6 0.1735 14 1.2807 18 1.2770
7 0.1670 18 1.6314 19 1.2364 7 0.1669 10 1.2805 4 1.2367
3 0. 1474 19 1.5775 8 1.2361 3 0. 1474 I 1.2798 22 1.2082
2 0.1420 6 1.5702 4 1.2184 2 0.1421 2 1.2798 6 1.2070
1 0.1263 10 1.5337 22 1.1488 1 0.1264 5 1.2791 5 1.1329
II 1.5153 6 1.1476 18 1.2791 21 1.1324
5 1.5116 2 1.0858 17 1.2790 2 1.1236
4 1.4962 3 1.0718 20 1.2790 11 1.0312
12 1.480 1 11 1.0224 6 1.2787 17 1.03 12
21 1.4589 17 1.0224 4 1.2768 10 1.0308
17 1.4579 10 1.0212 13 1.2750 16 1.0308
13 1.4487 16 1.02 12 21 1.2748 12 1.0270
2 1.3933 12 1.0 187 II 1.2738 13 1.0270
20 1.3267 13 1.0187 19 1.1779 14 ] .0270
22 1.1834 14 1.0187 15 1.1339 15 1.0270
15 1.0231 15 1.0187 7 1.1252 3 0.92 13
7 0.9722 I 0.8746 22 0.9586 1 0.8354

Fig. 2.23 Stress Concentration Factor of Each Section of the Bolt (M20-2) (continued) Fig. 2.23 Stress Concentration of Each Section of the Bolt (M20-3) (continued)

52 53
 Bolt Shank near the Bolt Head Incomplete Threaded Portion
(a) (b)

j Threaded Portion (c)

j

Bolt Shank Bolt Thread

Tensile Load = 21875 kgf High Strength Bolt

Bolt Shank near Bolt Head Incomplete Threaded Portion Threaded Portion
Nodal Stress Nodal Stress Nodal Stress
Number Concentration Number Concentration Number Concentration
Factor Factor Factor
10 0.5910 9 1.0679 4 1.0635
8 0.2844 8 1.0678 19 1.0635
5 0 2028 16 1.0662 22 1.0632
4 0.2000 3 1.0656 8 1.0631
9 0.1979 12 1.0656 20 1.0629
6 0.1735 14 1.0655 9 1.0627
7 0.1670 10 1.0652 7 1.0626
3 0.1475 1 1.0649 18 1.0623
2 0.1421 2 1.0649 6 1.0621
1 0.1263 18 1.0642 2 1.0596
5 1.0641 5 1.0575
(a) Incomplete Threaded Portion (b )Threaded Portion
17 1.0641 21 1.0563
20 1.0640 10 1.0489
15 1.0638 16 1.0489
6 1.0638 11 1.0457
19 1.0630 17 1.0457
4 1.0622 13 1.0423
13 1.0609 14 1.0423
21 1.0606 15 1.0423
11 1.0597 12 1.0422
7 1.0592 3 0.8724
22 0.9207 I 0.8406

(c) Bolt Shank near the Bolt Head

Fig. 2.23 Stress Concentration of Each Section of the Bolt (M22-l) (continued)
Fig. 2.23 Stress Concentration of Each Section of the Bolt (M20-4) (continued)

55
54
 Tensile Load = I 8055 kgf Tcsile Load = 22600 kgf

Bolt Shank ncar Bolt Head Incomplete Threaded Portion Threaded Portion Bolt Shank near Bolt Head Incomplete Threaded Portion 1l1readed Portion
Nodal Stress Nodal Stress Nodal Stress Nodal Stress Nodal Stress Nodal Stress
Number Concentration Number Concentration Nwnbcr Concentration Nwnbcr Concentration Nwnber Concentration Number Concentration
Factor Factor Factor Factor Factor Factor
10 0.5979 9 2.9347 7 2.0358 IO 0.5910 9 1.2786 7 1.2753
8 0.3032 8 2.9168 18 2.0355 8 0.2845 8 1.2784 18 1.2747
5 0.2157 14 2.78 12 5 1.3763 5 0.2027 16 1.2765 19 1.2728
4 0.2 139 16 2.4595 21 1.3759 4 0.1999 3 1.2757 8 1.2726
9 0.2086 3 2.0690 9 1.3160 9 0.1979 12 1.2755 20 1.2712
6 0.1853 I 1.7667 20 1.3160 6 0. 1735 13 1.2755 9 1.2700
7 0. 1789 18 1.6385 19 1.2299 7 0.1669 14 1.2755 4 1.2297
3 0.1582 19 1.5846 8 1.2296 3 0.1474 10 1.2753 22 1.2002
2 0.1511 6 1.5777 4 1.2124 2 0.1421 1 1.2747 6 1.1990
1 0. 1359 10 1.5406 22 1.1418 1 0.1264 2 1.2747 5 1.1284
11 1.5223 6 1.1406 11 1.2743 21 1.1279
5 1.5194 2 1.0808 18 1.2741 2 1.1 I 66
4 1.5025 3 1.0658 5 1.2739 11 1.0262
12 1.4908 1I 1.0184 17 1.2738 17 1.0262
21 1.4652 17 1.0184 20 1.2738 10 1.0258
17 1.4641 10 1.0172 21 1.2738 16 1.0258
13 1.4556 16 1.0172 6 1.2722 12 1.0220
2 1.4007 12 1.0147 4 1. 2715 13 1.0220
20 1.3318 13 1.0147 19 1.1808 14 1.0220
22 1.1889 14 1.0147 15 1.1388 15 1.0220
15 1.0292 15 1.0147 7 1.1326 3 0.9173
7 0.9790 1 0.8696 22 0.9613 1 0.8314

Fig. 2.23 Stress Concentration of Each Section of the Bolt (M22-2) (continued) Fig. 2.23 Stress Concentration of Each Section of the Bolt (M22-3) (continued)

56 57
 Chapter 3

Mechanical Behavior of High Strength Bolts and Its Adjacent

Structural Elements

3.1 Introduction
Tens1lc Load = 24115 kgf

Bolt Shank ncar Bolt Head Incomplete Threaded Portion Threaded Portion In order to fully understand the mechanical behavior of high strength bolted tensile joints, fundamental
Nodal Stress Nodal Stress Nodal Stress characteristics such as load transferring mechanism by bolts and the flange plates should be studied; so
Number Concentration Nwllbcr Concentration Nwnbcr Concentration that the mechanical behavior of the high strength bolts and its adjacent structural clements as shown in
Factor Factor Factor
Fig. 3.1 is discussed in this chapter. The mechanical behavior of the bolt and its adjacent structural
10 0.5983 9 1.0644 7 1.0616
elements is very basic for understanding of load-bolt force relation and eiTcct of the bolt pre-stress force
8 0.3033 8 1.0641 18 1.06 12
5 0.2158 16 1.0626 19 1.0600 on mechanical behavior. Moreover, these understandings arc considered to be applicable for other
4 0.2139 13 1.0622 2 1.0599 connections using high strength bolts. Here, structural elements adjacent to the high strength bolts as
9 0.2087 3 1.0621 4 1.0598 shown in Fig. 3.1 is the axisymmetric model for simplicity in order to understand the interaction of bolt
6 0.1854 12 1.062 1 8 1.0596
elongation and flange plate bending. In this study, this is named BAF modcl(high strength fiolt and its
7 0.1790 14 1.0618 20 1.0577
3 0.1583 10 1.0617 9 1.0575 Adjacent flange plate). Using this model, mechanical behavior in detail under tensile loading varying
2 0.1512 I 1.0614 22 1.0572 thickness of the circular plate diameter of it, and bolt pre-stress force is investigated experimentally and
I 0.1359 2 1.0614 6 1.0561 analytically. As shown in Fig. 3. 1, this model consists of two circular plates jointed by a high strength
II 1.0610 5 1.0532
bolt at the center of the plates and it is simple enough to identify Ute contribution of the bolt and Ute flange
18 1.0608 21 1.0521
5 1.0606 1.0419 plate to joint behavior. It is also very easy to analyze this model using finite clement analysis because of
10
17 1.0606 16 1.04 19 structural axis)1nmetry. In addition, pl)ing force according to the mechanism discussed in the previous
21 1.0606 11 1.0389 studies [I ]-[3] does not occur due to Ute shape of the model, so that it is able to study the eiTect of Ute
20 1.0604 17 1.0389 following parameters only: pre-stress force of the bolt, thickness of the circular plate and diameter of
6 1.0591 12 1.0354
circular plate(distance between the center ofU1e bolt and loading point)
4 1.0585 13 1.0354
7 1.0577 14 1.0354
15 1.0568 15 1.0354
19 1.0566 3 0.8684 3.2 Experiment on BAF Model
22 0.9284 1 0.8350

3.2.1 Outline of the experiment

Fig. 2.23 Stress Concentration of Each Section of the Bolt (M22-4)
Monotonic tensile loading test for BAF model is carried out in order to study its mechanical behavior.
Dimensions of all the specimens are tabulated in Table 3.1. In this experiment, varied are the thickness
and the diameter of the circular plate, the bolt pre-stress force as the parameter. SS400 steel is used for
the circular platc(fl ange plate), and either M16(FIOT) or MJ2(F lOT) bolt is used for the high strengUt
bolt. The diameter of the circular flange plate is detennined by considering the loading capacity of Ute

58 59
 tcstmg machine. Material properties obtained from the matenal test are swmnarized in Table 3.2. separation of two circular plates and bolt force are measured. Each measurement is descnbed m the

l11e characteristics of each specimen arc described as follows: BAF-1 and BAF-2 consists of following:

circular flange plates with thickness of 25 (rmn) and diameter of 52 (mm) and high strength bolt of
M 16(F IOT), but bolt pre-stress force given to the bolt is different. The objective to test these two models (l) Tensile applied load

is to investigate the effect of the bolt pre-stress force. In case of BAF-1, the bolt pre-stress force is 11 .7 Load is measured by the load cell installed in the actuator.

{ton!), which is standard bolt pre-stress force for M16 defined by JIS[4]. On the other hand, in case of (2) Separation between two circular plates

BAF-2, pre-stress force of bolt is 5.85 (ton!), which is 1/2 of standard bolt pre-stress force mentioned in 11tis is measured by tl1e strain gage type displacement transducer as shown in Fig. 3.5. The

the above. The thickness of the circular plate, 25 (mm) is dctcnnincd as follows: Asswning the simply location to measure this separation is also sh0\\11 in Fig 3.6.

supported circular plate subjected to the concentrated load at the center of the plate, it is dctem1ined such (3) Bending dcfonnation of the circular pate

that the maximum bending stress of circular plate reaches to yield stress (30 kgf/nun 2). Among BAF-3, ln order to measure the bending deformation of the circular plate, the strain gage for the

BAF-4 and BAF-5 models, diameter of the circular plate (the loading length between the center of the stress concentration measurement is used. l11e location to measure bending strain is also

bolt and the loading point) and the bolt pre-stress force arc the same; but only the thickness of the circular shown in Fig 3.6.

plate is drffcrcnt each other. The thickness of the circular plate of BAF-3, BAF-4 and BAF-5 is 1.0 times, (4) Bolt force

0 5 times and 0.25 times of the standard thickness(BAF-3) respectively, which is determined by the same Bolt force is obtained by tl1e strain gage installed into tlte center of the bolt shank as sh0\\11

procedure used for BAF-1 and BAF-2 and tl1is standard thickness of the circular plate is 19 (mm). 111e in Fig. 3.4. 11tc bolt force is calculated by using data-sheet of tl1e calibration test for each

high strength bolt of M 12(Fl on is used for all the specimens. The objective to compare these models is bolt.

to imcstigatc the effect of tl1e tl1ickncss of tl1e circular plate on contact/separation behavior. It is
considered tl1at the failure mode is diiTerent by the tl1ickncss of the circular flange plate. In case of the 3.2.2 Experimental results and discussions

thinner circular plate, the bending failure is asswned to occur at the circular plate. On the other hand, in Load vs. separation curves obtained from the experiment are sh0\\11 in Fig. 3. 7. In tltis figure, the

case of the thicker circular plate, the tensile failure of the bolt is asswned to occur. The bolt pre-stress horizontal axis shows the average separation which is given by averaging tlte reading of 4 displacement

force used into these models is 6.26 (ton!), which is specified in JIS for M12(FlOT). Finally, BAF-6 is transducers in 90 degree pitch around the circular plate. In addition, load vs. bolt force curves and the
for investigation of the effect of the distance between the center of the bolt and loading edge by comparing strain distributions on the circular plate are shown in Fig. 3.8 and Fig. 3.9, respectively. Results and
to BAF-4. This distance of BAF-6 is 1.5 times of that of BAF-4, namely, 78 (nun). discussions are written in the following for each factor to affect the mechanical behavior.

In the experiment, attention is paid to the local dcfonnation of tl1e circular plate, t11e gap (a) Effect of the bolt pre-stress force

developed between two circular flange plates(scparation) and the variation of the bolt force in order to In order to investigate tlte effect of the bolt pre-stress force, BAF-1 and BAF-2 are compared. Different
study the contact/separate behavior and the stiffness of tl1e joint in detail. bolt pre-stress force is given to BAF-1 and BAF-2; bolt pre-stress force of BAF-2 is half of BAF-1 . As

Test setup is sh0\\11 in Fig. 3.2. The testing machine used in this experiment is SHIMADZU shown in Fig 3. 7, it is found that two circular plates arc not separated in a large scale until tltc tensile load

electrically servo-controlled hydraulic actuator whose loading capacity is +/-30 (ton!) statically, +/-20 reaches to tlte bolt pre-stress force for both cases. Furtl1ennorc, it is also shown that tl1e higher tl1e bolt

(ton!) d)11amically and maximum stroke is +/-50 (nun). In this system, the loading is controlled by the pre-stress force is, the higher the joint stiffness is obtained because of less separation. Here, the stiffness
analog controller operated by micro-computer through GP-IB interface, where the displacement of tile of the joint is defined as the tangential slope of tl1c load-separation curve. In addition, it is found that the
actuator is monotonically increased. The tensile load is applred to the test specimen unifonnly around the stiiTncss of BAF- 1 is same as that of BAF-2 after the tensile load reaches to the bolt pre-stress force.
edge of the circular plate. The loading is continued until the maximum load point is obtained. Schematic Therefore, it is concluded that the bolt pre-stress force affects only the load level at separation of the two

view of this loading is shom1 in Fig. 3.3. l11c bolt pre-stress force is checked by the reading of the strain circular plates and initial slope of load-separation curve.

gages installed in the bolt shank as shown in Fig. 3.4, which is already calibrated to bolt force. In order to Furt11cnnore, it is understood from Fig. 3.8 that the bolt force is kept constant before the tensile

avoid the eccentric loading, hinge joints arc utilized in the loading apparatus as shown in Fig. 3.2. load reached to the bolt pre-stress force, and that after the tensile load reaches to the bolt pre-stress force,

In the experiment, tensile applied load, stroke of the actuator, strain on the circular plates, bolt force becomes larger as the tensile load increases. At the early stage of loading, the applied load is

60 61
 considered to be balanced to the release of the compressive force bet\\cen two circular plates given by the
bolt pre-stress force before the tensile load reaches to the bolt pre-stress force, and after that, the applied
load is carried only by the bolt. Therefore, it is concluded that prying force docs not exist in this model as
expected. Moreover, it is shO\m that circular plate is not dcfonncd significantly even if large separation
took place as shown in Fig 3.9. Therefore, it is understood that the bending strength of the circular plate is
higher than that of high strength bolt to make only tltc bolt elongate.
(b) Effect of the thickness of the circular plate
Results of BAF-3, BAF-4 and BAF-5 arc compared in order to investigate the effect of the thickness of
the circular plate. Among these specimens, only a difference is the thickness of the circular plate and other
conditions arc the same. TI1c U1ickness of the circular plate of BAF-3, BAF-4 and BAF-5 arc 19 (nun), 10
(mm) and 5 (mm) respectively. From the load-separation curves( Fig. 3. 7), it is fow1d that Ute stiffness of
the mode l becomes higher as the U1ickness of the circular plate becomes larger. Namely, the stiffness of
BAF-3 is Ute highest among those of these specimens, and that of BAF-5 is the lowest. Furthern10re, in
the case of the BAF-3, which has the thickest circular plate, after Ute applied load reached to the
maximum load, applied tensile load decreases gradually. On the other hand, in case of the BAF-4 and
BAF-5, \\hic h has the thinner circular plate than BAF-3, the applied load increases monotonically even if
the separation becomes quite large. It is considered that this difference in load-separation curves is caused
by the difference of the failure mode. Namely, in case of the thick circular flange plate, the limit state is
defined by the failure of Ute high str ength bolt under the tension ; on the other hand, in case of U1e thinner
circular flange plate, it is the failure of the circular plate under bending. Therefore, the load-separation
curve of BAF-3 significantly depends on load-elongation curve of only the high strength bolt; and in
similar way, that of BAF-5 significantly depends on load-bending deflection curve of the simply
supported circular plate subjected to point load at the center of the circular plate.
Moreover, from Fig. 3. 9 of the load-strain curves at the circular plate, in case of BAF-3, it is
found that the strain due to plate deflection is at most 200 (micro). Titis also concludes that the circular
plate is not dcfonned significantly and that the separation of BAF-3 depends o n elongation of the bolt. On
the other hand, in case of BAF-5, the circular plate is considered to be bent significantly according to the
strain distribution on the circular plate: where Ute maximum strain is beyond the yield strain. As for BAF-
4, which has intcnnediatc thickness of the circular plate, it is understood that the behavior discussed in the
above for thin and thick flange plate is somehow combined. As for the load-bolt force relationship, in case
of BAF-3 , the bolt force is kept constant until Ute applied load reaches to the bolt pre-stress force, and
after that the bolt force increases as U1c applied load increases where the linear relation exists. On the
other hand, in case of BAF-5, the bolt force increases gradually before the applied load reaches to the bolt
pre-stress force as compared with BAF-3 . Particularly, the bolt force increases even if the applied load is
very small. The increase of the bolt force at the early stage of loading is similar to the prying force effect
in the previous study[ I l-[3]. However bolted circular plate tested in this study cannot have prying force
62
action. This increase of bolt force is caused by the phenomena that the local deformation of flange plate
pull the bolt head; nan1ely, at the edge of the bolt hole, prying force action may occur due to large
deflection of Ute circular plate.
(c) Effect of the loading length
In order to investigate the effect of the loading length between the center of the bolt and the loading edge,
results of BAF-4 and BAF-6 are compared. TI1c loading length of BAF-4 and BAF-6 arc 52 (mm ) and 78
(mm) respectively. From the load-separation curve, the tangential slope of U1e BAF-4 is higher Ulan that
of BAF-6. Moreover, it is understood from the load-bolt force curve that the bolt force of BAF-6
increases at U1e early stage of loading compared with BAF-4. It is also concluded that the less stiffness of
the circular plate with the large loading lcngth(diamctcr of the circular plate) cause large deflection so
that this local defom1ation cause an additional increase of bolt force.
3.3 Finite Element Analysis on BAF Model
3.3.1 Outline of finite element analysis
In the previous section, the loading test for BAF model is described. 1l1e loading test is very useful to
understand the mechanical behavior for particular model; however it is not suitable for the problem with
many parameters to be considered because of the limitation of the time and cost for tl1e experiment. On
U1e other hand, the nwnerical analysis only needs the computer which has high computation capacity; in
addition, it is very easy to analyze the problem for many combinations of parameters, so tltat it is
considered to be suitable for quantitative evaluation. In this section, finite e lement analysis for BAF
model is carried out by varying Ute parameter, such as thickness of the circular plate, the diameter of the
circular plate, in order to supplement the result of the loading test and furtJter to investigate tltc
mechanical behavior of BAF model quantitatively.
3.3.2 Numerical analysis method
As Ute tensile load is applied to BAF model, the gap between two circular plates becomes large gradually.
1l1at is, a boundary condition on the contact surface should be changed from the state of contact to tltat of
separation during the loading. lltcrcforc, in the analysis, this bow1dary non-linearity must be taken into
consideration. Accordingly, in this study, 2-dimcnsional axiS)11Unctric finite clement analysis is carried
out considering both material non-linearity and boundary non-linearity.
The program developed in this study is the finjte clement analysis program using triangular
element with constant strain, whose core utilizes EPIC IV. Tile detail of EPIC IV can be referred to the
refcrence[5]; here, Ute function added to EPIC IV is described in the following. As mentioned above, it is
required to consider boundary non-linearity for contact/separation behavior on BAF model. In this
63
program, the change of boundary condition is executed according to a certain condition on the nodal force
or nodal displacement along the contact surface(line in 2-0 problem). Contact or separation surface is
located on x-axis as sho\\n in Fig. 3.1 0. Tile nodal points in contact condition is constrained to be zero
displacement in y-dircction at first, and the nodal point in separation condition is constrained to be zero
force in y-direction. If the nodal force becomes negative on the contact surface, the separation occurs and
boundary condition is changed from that for contact condition to that of separation condition. On the other
hand, 1f the nodal displacement in the separation condition becomes negative, the contact occurs and
boundary condition is changed from that of separation condition to that of contact condition.
3.3.3 Parametric study for BAF model
Here, parametric study is carried out in order to estimate t11e stiiTness of BAF model, where the thickness
of t11c circular plate and the loading length between t11e center of the bolt and loading edge arc considered
to be parameters. In this analysis, II cases arc taken into consideration. All the cases arc listed in Table
3.3. In the name of analytical cases considered, t11e character following "A" denotes the half of t11e
diameter of the circular plate. Tile characters "S" ,"M" and "L'' denotes 52 (mm), 78 (nuu) and I 04 (mm)
respecti,cly. The nominal diameter of t11e bolt and the bolt pre-stress force for each case are t11e same as
those used for test specimens, namely, M 12(F I OT) and 6.26 (ton f) respectively.
Finite element model for BAF model and its discretization by triangular elements are shown in
Fig. 3. 11 . 1l1c location considered to have high stress concentration is discretized by fine mesh such as
the bolt shank ncar the head and the loading edge. Boundary conditions are also shown in Fig. 3 .12.
Locating t11c surface to be judged either in contact or separation condition on x-axis, nodal points on x-
axis is fixed in y-dircction at the initial state. Furthcnnore, in this model, it is assumed that the bolt head
and the circular plate are kept in contact all the time; that is, both arc treated to be a continuous body.
DiiTercncc between the circular plate and the bolt is just made by changing material properties. Material
properties used in this analysis are swnmarized in Table 3.4. Tilcse are detennined based on t11e
specification of JJS. The material is assumed to be clastic perfectly plastic. As shown in Table 3.4, the
properties of the bolt thread arc different from those of the bolt shank in order to consider the shape of the
bolt thread. 1l1c yielding stress and the stiiTncss of the bolt thread is 75% of t11osc of t11c bolt shank. This
ratio is defined based on the ratio of cross sectional area of bolt shank to that of tllC bolt thread. Moreover,
in tl1is model, the washer is omitted for simplicity. Because the objective of this analysis is mainly to
understand the global mechanica l behavior of bolted joints, that is, the separation of the 11ange plates.
Furthcnnorc, the difference between the diameter of the bolt head and that of the washer is a little, i.e.,
0.6 mm and tltc thickness of the washer is much thinner than that of the flange plate.
In case of the thickest and largest circular plate, total number of the nodal points and clements
is 390 and 684 respectively. ll1c procedure of this analysis is summarized in the followings(referred to
Fig. 3.13):
64
(Step I) Bolt pre-stress force is given to the bolt, by applying the tensile load at t11e end
of t11e bolt as unifonuly distnbuted load.
(Step 2) After bolt pre-stress force prescribed in JIS is given, the bolt end is fixed at the
displaced position in vertical dircction(y-direction).
(Step 3) Unifonnly distributed tensile load is applied along t11e edge of circular plate;
namely, ring load
3.3.4 Results and discussions
(a)Load-deformation curve
Load vs. defom1ation curves obtained by the analysis are shO\m in Fig. 3.14. Tile vertical axis shows the
applied tensile load and tl1e horizontal axis shows the displacement of t11c loading point as sho\\n in Fig.
3.11.
In case of small diameter-to-thickness ratio of circular plate, large displacement begin to occur
when the applied tensile load becomes about 8.0 (ton(), so that t11e curves become bi-linear, which is very
similar to that of the bolts. It is considered that mechanical behaviors of these cases depend on only load-
elongation curve of the high strength bolt, not t11e thicker circular plate. In addition, it is found that in the
case of small diameter-to-thickness ratio of circular plate, the slope of the load-displacement curve is
steeper. On the other hand, in case of large diameter-to-thickness ratio of circular plate, large deformation
begin to occur even before the tensile applied load reaches to 8.0 (ton() which is a minimwn yielding load
of the bolt specified in JIS. And, it is found that the load-defonuation curve is significantly non-linear.
This is considered to be caused by the yielding of the circular plate. The magnitude of tl1e load when the
large defonnation occurs is different and depend on the geometrical configuration of the circular plate.
Tilerefore, the mechanical behavior of t11ese cases is considered to depend on not the high strength bolt
but the circular plate.
Next, t11e process to spread the yielding region are sho\\n in Fig. 3. 15. It is understood from
these figures that in case of small diameter-to-thickness ratio, t11e bolt yields. On the other hand, in case of
large diameter-to-thickness ratio, the circular plate is found to yield. This characteristics is in good
agreement witl1 t11e results observed in the experiment.
(b)Joint stiffness of the BAF model
Stiffness vs. load curves obtained by the analysis are sho\\n in Fig. 3.16. TI1e vertical axis shows the
stiffness, and the horizontal axis shows the applied tensile load. In these figures , the spread of the contact
surface to be separated, the stiffness of the high strength bolt under tension and the bending stiffness of
t11e circular plate and t11e total stiffness of the joint arc also sho\m for reference. The joint stiffness is
defined by the slope of load-separation curve. Tile stiffness of tl1e circular plate is defined by the slope of
t11e load-deflection curve for a simply supported circular plate subjected a concentrated load at the center
of the plate(6]. ll1e stiffness of the bolt is the tensile stiffness of the bar whjch has the same cross
65
sectional area as the bolt. joint stiffness K, stiffness of the !Ugh strength bolt Ks, and bending stiffness of
the circular plate Kp is derived by the following equation.
flP (3.1)
K =-
flu
(3.2)
(3.3)
(3.4)
in which 61', 6.11, 8, I. R. I, v, and Ep, is the increment of the tensile load at each loading step, the
increment of the separation at each loading, Young's modulus of the bolt, clamped length of the bolt,
radius of the circular plate, thickness of the circular plate, the Poisson's ratio of the circular plate and
Young's modulus of the circular plate respectively.
On the other hand, the total stiffness of the joint for a simple model assumed that tlte high
strength bolt and the circular plate arc connected in series is derived. Total stiffness of the joint Kr is
given in the following based on the stiffness of tlte bolt and tlte bending stiffness of the circular plate
defmed by equation (3.2) and (3.3).
(3 .5)
It can be seen from Fig. 3.16 that afier the nodal point in contact condition turns to be separated,
the stiffness of the joint decreases gradually. It means that the stiffness decreases as the more separation
takes place. Furthennore, the sudden decrease of tltc stiffness as the tensile load increases is considered to
be due to a coarse discretization on the contact surface. It can be concluded that the compressive force
between flange plates given by the bolt pre-stress force is contributed to the high stiffness of joints. In
addition, in case of small diamcter-to-thiclmess ratio, it is found that the circular plates are separated
completely without any significant deflection of circular plate; as a result, only the bolt resist against
applied tensile load. Finally the stiffness becomes to zero suddenly afler the tensile load reaches to the
yielding load of the bolt. On the other hand, in case of large diameter-to-thickness ratio, it is shown that
two circular plates arc not separated completely, and it is considered tltat the applied tensile load causes
the significant dcfonnation on the circular plate and only tlte flange plates compressed by the bolt head
are in contact condition. The failure mode of all the cases can be summarized in the followings : in case of
AS-I, AS-2, AM- I, AM-2, AL-l, a decrease of tlte sllffness is caused by the yielding of the bolt thread,
in case of AS-4, AS-5, AM-4, it is caused by tltc yielding of the circular plate, in case of AS-3, AM-3,
66
AL-2, it is caused by the yielding of both the bolt and the circular plate.
Based on a result of observations mentioned in tlte aboYe, stiffness-load cune ts classified into
tltree patterns by the thickness of the circular plate and tlte diameter of the circular plate (tlte loading
length behveen the center of bolt and loading edge). Schematic 'iew of tltree patterns for stiffness-load
curve is shown in Fig. 3.17. In case of small diameter-to-thickness ratio, initial stiffness is very high. As
the applied tensile load increases, the stiffness decreases significantly at the early stage of loading; and
then the stiffness is kept almost constant. Titis constant stiffness is considered to be equivalent to that of
the circular plate. Afier the tensile load reaches to the bolt yielding load, finally the stiffness decreases
again and becomes almost .lCro. On tlte other hand, in case of large diameter-to-thickness ratio, initial
stilTness is not so high as that of small diameter-to-thickness ratio, and the stiffness decreases gradually as
the applied tensile load increases and then it becomes almost zero very quickly before the applied tensile
load reaches to the bolt yielding load. Finally, in tlte case of tltat medium diameter-to-thickness ratio, tlte
decrease of the stiffness is in behveen the above two cases.
Next, in order to investigate the effect of the bolt pre-stress force on initial stiffness, tltc
stiffness obtained by ilie analysis is compared witlt total stiffness Kr of the model. lf the total stiffness is
rugher than that obtained from the analysis, the stiffness of ilie model increases by the bolt pre-stress force.
As already shown in Fig. 3.16, total stiffness calculated by equation (3.5) is compared to tltat obtained
from ilie analysis. Here, ilie effect of tlte bolt pre-stress force is observed for all cases. It is found that in
case of small diameter-to-thickness ratio, the sti!Tncss increases by 100 % at the initial state due to the
bolt pre-stress force; but less for case of large diamcter-to-tltickness ratio. Titcrefore, it is concluded that
ilie thicker the circular plate is and the smaller tlte diameter is, the higher the stiffness is obtained by the
bolt pre-stress force.
3.3.5 Evaluation of the stiffness on BAF model
From the view point of the design of the joint considering the energy absorption or rational economical
design, accurate evaluation of the stiffness of the joint is very important. Accordingly, in this section, in
order to establish the quantitative evaluation fonnula for the joint stiffness of the BAF model which
should be applicable in general use, multiple regression analysis is carried out based on the results of the
fmite element analysis for the BAF model.
In this evaluation, used are the following norntalL~ed parameters considered to be most
influential on the stiffness of the joint. These are t!R, P P0 , PIPu. t/R is tltc ratio of radius to thickness for
circular plate. P!P0, and PIPu arc nonnalized load, " here Po, Pu is bolt pre-stress force given into the bolt
and the ultimate load of joints respectively. This ultimate load is small force of either the bolt yielding
force or the plastic collapse load of tlte circular plate subjected to ring load at ilie edge of the circular
plate. Titis plastic collapse load of the circular plate is gi,en by the following equation .
67

1
p =;r 6R a,t (3.6)
3R- 2a 4
in which R, t, 0) and a is the radius of the circular plate, the thickness of the circular plate, the yielding
stress and the radius of the loading area respectively.
Accordingly, the equation for evaluation of the stiiTncss on BAF model is proposed by using the above
parameters as follows:
(3.7)
in which K and Ko arc the stitTness of the joint and the stiffness of the circular plate defmed by the
equation (3.3). Furthennorc, Ct-C7 are coefficients to be dctcnnined.
This equation is dctcnnined based on 3 pat1cms of the decrease of the stifTness as increase of
the applied tensile load. Tite first and the second tcnns are the functions to represent a decrease of the
stiilitess exponentially, which is concerned with the ratio oftJte radius to the thickness oftlte circular plate.
The third temt is a function to have characteristics shown in Fig. 3.18, and it represents the rapid decrease
of the stiffness when the applied tensile load reaches to tlte ultimate load. All the tenns are concerned with
tltc ultimate load of the joint, so that the stiffness becomes almost zero if the applied tensile load reaches
to the ultimate load.
If tlte linear least square method is used in multiple regression analysis, in case of this equation
(3.7) proper lincariLation must be needed. Here, the different weight for each data may be taken if this
linearization is carried out; therefore, it is difficult to carry out regression analysis for tlus type of
equation accurately. Accordingly, in tltis study non-linear least square method is utilized to obtain tlte
coefficients dircctly[7Jl 8]. The coefficients in this equation obtained by the multiple regression analysis
are sh0\\11 in Table 3.5. The applied tensile load P and the stiiTness Kat the each loading step obtained by
the finite element analysis arc used as the input data. But the data that a stiffness is very small is
neglected for accurate estimation because of a\'oiding that the small stiffness is improperly weighted in
the analysis. Here, the data that a load is larger than 1. 1 times of the yielding load of tlte bolt arc
neglected.
Tite regression curves of tltc stiffness of BAF model for each analytical case are compared and
shown in Fig. 3. 19. In these figures, the error in the regression analysis is also shown. Tit is error is
nonnalizcd such as the relative error di ... idcd by the maximum stiffness of each case obta ined by the finite
68
element analysis. From these figures, the evaluation of the stiffness using the equation (3. 7) is considered
to be applicable and in case of thicker circular plate, it is understood that tlte evaluation by this equation
is very accurate. On tlte other hand, it is observed tltat tlte error tends to be big as the tensile load reaches
to the ultimate load, where the stiffness is VCI)' small. However, fomt the view point of the order in the
stiffness, the range that tlte stiffiless is almost zero is not so important. In addition, Tite comparison \\itJ1
the stitiness obtained by the evaluation and that obtained by finite clement analysis for all tlte cases is
also shown in Fig. 3.20. ft is also understood from tltis figure that the regression results for all the data
using the equation seems quite good except for the data that a stiffness is very large. Titcrcfore, the
evaluation for tlte stiffness of BAF model using tltis equation could be applicable.
3.4 Conclusions
In this chapter, mechanical behavior of the high strength bolt and its adjacent structural elemcnt(BAF
model) is investigated experimentally and analytically. In particular, the contactl~eparation behavior and
its stitrncss arc taken into consideration. Tite stiffness is an important infonnation for design of the joint
considering the energy absorption capacity or eartltquake resistance of structures. Therefore, the simple
evaluation method for the stiffness such a basic element are proposed. Titc conclusions and future needs
are obtained as follows:
1) Tite mechanical behavior of BAF model depends on the diamcter-to-tJtickness ratio of circular
flange plates and bolt pre-stress force, and failure mode of this model is aficcted by diameter-
to-thickness ratio. In case of smaller diameter-to-thickness ratio, the initial stiffness is higher
and the failure is caused by the bolt yielding, which is brittle. On the other hand, in case of
larger dianteter-to-thickness ratio, tlte initial stiiincss is not so high and tlte failure is caused by
the circular plate yielding. Furthcmtore, energy absorption capacity is considered to be higher.
As for the effect of tlte bolt pre-stress force, the higher the bolt pre-stress force is, the higher the
stiffness is obtained.
2) In general, the bolt force is kept constant until the tensile load reaches to the bolt pre-stress
force given to the bolt, the applied load is carried by the release of tltc compressive force given
by the bolt pre-stress force. However, in case of thinner circular plate, that is, the bending
strength of the circular plate is low, it is observed that the increase of the bolt force occurs at
the early stage of loading. It is considered to be similar to the pl)ing force effect where the bolt
is pulled by local defonnation of Ute circular plate.
3) Titc evaluation method for the stiiTness of the BAF model is proposed based on tlte results of
multiple regression analysis using non-linear least square method. It is concluded that the
69
 proposed equation is applicable for the evaluation of the stiffness. Tabte 3. 1 D.tmenStOnS 0 f tJle spcctmens
Specimen l11ickness of the Radius of the Bolt Pre-Stress High Strength Bolt
In the future, the equation should be further modified in order to evaluate the stiffness more accurately at Circular Plate Circular Plate
the wide range and large intense of stiffness. And the utilization of the results obtained in this chapter is (I) (R)
expected to be implemented in the design for the various bolted connections. NB-1 25 78 11.7 M16(F10T)
NB-2 25 52 5.85 M 16(F10T)
References NB-3 19 52 6.26 MI2(F10T)
I) B.Kato, A.Tanaka : High Strength Bolted Tensile Joints -Influence of Bolt Pre-stress Force-, Journal NB-4 10 52 6.26 MI2(FIOT)
of Structural and Construction Engineering, No. 146, Apr. 1968, pp. 21-27, Architectural Institute of NB-5 5 52 6.26 Ml2(FIOT)
Japan(in Japanese). NB-6 10 52 6.26 M 12(FIOT)
2) B.Kato, A. Tanaka : Experimental Study on High Strength Bolted Tensile Joints -Mechanical Behavior wtit: mm
under monotonic tensile loading-, Journal of Structural and Construction Engineering, No. 147, May
1968, pp. 33-41, Architectural Institute of Japan(in Japanese).
3) R.T.Douty, W.McGire : High Strength Bolted Moment Connections, Journal of the Structural
Division, Proceedings of ASCE, Vol. 91, No. ST2, Apr. 1965, pp. 101-128.
4) Japanese Industrial Standard Conunittec : Sets of High Strength Hexagon Bolt, Hexagon Nut and
Side View
Plain Washers for Friction Grip Joints(B 1186), 1979(in Japanese).
5) Y.Yan1ada, Y.Yokouchi : The Progranuning of Elasto-Piastic Analysis by Finite Element Method-
EPIC IV Manual-, Baihukan, 1971.3(in Japanese).
6) S.Timoshenko, S.WoinowsJ,:y-Krieger : l11cory of Plates and Shells, Second Edition, McGraw-hill,
1970. Grand View
7) ASNOP Research Group : The Progranuning of Non-Linear Optimization, Nikkan Kougyo
Shinbunsha, Apr. 199I(in Japanese). Specimen
8) T.Watanabe, M.Natori, T.Oguni : Nwnerical Analysis Software by FORTRAN 77, Maruzen, Dec.
1989(in Japanese). Half Part of Specimen

. IPropcrt1cs
Table 3.2 Malena
Circular Plate Yung's Modurus Yielding Stress
(less than I0 nun) 21000 28.5
(more than 10 mm) 21000 26.9
unit kgf/cm2

70 71
 Table 3.3 L.lSI 0 fAn ai)1Jca
I . I Cases
Case Titickness Radius Radius
of of to
the Circular Plate the Circular Plate Titickness Ratio
AS-1 22 52 2.36
AS-2
AS-3
19
15
52
52
2.74
3.47
.. ~
a
I
1 Circular Plate

AS-4 10 52 5.2
AS-5 5 52 10.4
AM-I 25 78 2.08
AM-2 19 78 4.11
AM-3 15 78 5.2
AM-4 10 78 7.8 Modeling I
I
I

$
AL-l 25 104 4. 16
AL-2 19 104 5.47
Flange Plate
(unit : nun)

Table 3 4 Matenal Properties used in the Analysis

Young's Modulus Yielding Stress
Fig. 3.1 Model of High Strength Bolt and its Adjacent Flange Plate (BAF Model)
Circular Plate 21,000 28.5
(less than I0 nun)
Circular Plate 21,000 26.9
(more than I 0 nun)
Bolt Shank 21 ,000 90.1
of
High Strength Bolt
Threaded Portion 15,700 67.1
of
High Strength Bolt

Tabl e 3 5 Resu Its 0 fRegress JOn

Loading Frame
c1 c2 c3 c4 Cs c6 c1
1.63 0.0 1 3.65 0.845 1.34 7.30 0.01

Hinge

t : Titickness of the Circular Plate R : Radius of the Circular Plate

P : Applied Tensile Load Po: Bolt Pre-stress Force
Pw : Ultimate Strength of the Joint
Fig. 3.2 Testing Setup

72 73
 Output/Input Line
r

High Strength Bolt

- - - --

Circular Plate
(Flange Plate)

1=1=1=
1--- ' - ' -

16 unit: mm 12

Fig. 3.3 Schematic View of Loading

Fig. 3.5 Displacement Transducer for Measuring the Gap

,...- - -
i i Strain Gage for Measurement
!I ~

ofthe Bolt Force
I
Circular Plate

High Strength Bolt -

D
Measuring Point
for Separation
Strain Gage for Stress
Concentration
~ I

---- - -~-ED-- _-_]_-~~- mm

0---!--
R: Radius of the Circular Plate
1. ; ~ ~
l~
f"'ll<------7;'
-......-,

R 26mm

High Strength Bolt

Fig. 3.6 Location of the Strain Gages glued on the Circular Plate and Displacement Transducer

Fig 3.4 Schematic View of the Strain Gage for Measurement of the Bolt Force and its Location

75
74
 20 ~------------------------~/ 20 -.------------------------~
/ //

,,,
20 20 / /
/
18 18 /
15 / 15
16 cc / cc
16
'E 14 ~ / g
'E
0
% 12
.3"'
14

10
.-" 0
% 12
"'
0
_j 10
"'0

.3"'
~
10
/
/
"'0

.3"'
Cl)
10

.!1 ;;; / ~
Cl
'iii 8 c / c
~ 8 c Cll Cl)
c Cl I- / I-
Cl
1- 6 1- 6 5 /
/
4 4 /
/
2 /
2
I//
0 0 0 ~-----.----~~---.----~

0 2 3 4 5 0 2 3 4 5 0 5 10 15 20 10 15 20
Scparation(mm) Separation(mm) Bolt Force(tonf) Bolt Force(tonf)

(a) BAF-1 (b) BAF-2

(a)BAF-1 (b) BAF-2

20 -.-----------------------------_, 20 -.------------------------~
// //
20 20 / /
/ /
18 18 / /
/ 15 /
16 15 / /
16 cc c /
/ c
'E 14 'E 14 9 / .B /
0 0 :u /
/ :u /
/
%12 %12
.3"' 10 / .3"' 10 /
.3"'
Cl
~
c
Cl
1-
10
8

4
.3"'
.!1
-~
1-
Cl
10
8
6
4
/.
...... - ~
;;;
c
Cl)
I-
5
/
/
/
//
/

5
/
/
//;
/
/ /
2 2 I / /
// //
0 0 0~------,.-e-----,------,,..---~ 0 ~----,._----,------,,.----f

0 2 3 4 5 0 2 3 4 5 0 5 10 15 20 0 5 10 15 20
Scparalion(mm) Separation(mm) Bolt Force(tonf)
Bolt Force(tonf)

(c) BAF-3 (d) BAF-4

(c)BAF-3 (d)BAF-4

20 -.-----------------------------_,/ 20 -.--------------""71/
/ /
20 20 / /
/ /
18 18 / /
/ 15 /
15 /
16 16 cc /
/ cc /
/
'E 14 'E 9 / .9
:u
0 0
14
:u /
/
/
/
% 12 % 12 "' 10
0 / _g"' 10 /

/I
"'
0
_j 10 _j"'
0
10
_J

~ /
/
~
.!1 ;;; / ;;;
8 .!1 c c
iii /

,,
c 'iii 8

: : ......
Cll Cll
c I- / I-
~ 6 ~ 6 5 / 5
4 /
4
2 ....,.,...,.- ~]
2
v/
////

v/
/
/

0 I 0 ~-----.e-----.------.----~ 0 ~----~----.-----.----~
0 2 3 4 5 5 0 5 10 15 20 0 5 10 15 20
Scparation(mm) Separation(mm) Bolt Force(tonf) Bolt Force(tonf)

(c) BAF-5 (f) BAF-6

(c)BAF-5 (f)BAF-6
Fig. 3.7 Load-Separation CurYes ofBAF Model
Fig. 3.8 Load-Bolt Force Curves ofBAF Model
76
77
 20 20 20 ~-----------~ 20
--<>- CH 1 -o- CH 1 18 --<>- CH 1
18 18 -<>- CH 1 18
-o- CH 2 -o- CH 2 -o- CH2
-o- CH2 16 16
16 ~ CH3 16 -<>- CH 3 -<>- CH3
-<>- CH 3
_,__ CH 4 'E CH 4
'E 14 ~ cH4 'E 14 ~ CH4 'E 14 14 --9-
-<>- CH 5
2 -<>- CH 5 2 -<>- CH 5 2 -<>- CH 5 2
:; 12 :; 12 :; 12 :; 12
C'll C'll
I'll
I'll
.3 10 0
...J 10 .3 10 .3 10
Cl) Cl) ~
~
'iii 8 ~ 8 ~ 8 'iii 8

r
cCl) c c c
6
Cl)
6 (!. 6 (!. 6

I
1- 1-

4 4 4 4

2 2 2 2

0 'l~ 0 0 ~~~~~~~~~~~~~~ 0
-1000 0 1000 2000 3000 4000 1000 2000 3000 4000 -1000 0 1000 2000 3000 4000 -1000 0 1000 2000 3000 4000
-1000 0
Strain (p} Strain (p} Strain (p) Strain (p}

(a) BAF-l (A side) (b) BAF-1 (B side) (g) BAF-4 (A side) (h) BAF-4 (B side)

20 20 ,--- -----------, 20
20
-<>- CH 1 -<>-- CH 1 18 --<>- CH 1
18 18 18 -<>-- CH 1
-o- CH 2 -o- CH2 -o- CH2
~ ~
~ cH3 16 -o- CH2 16 -<>- CH 3
16 16 -<>- CH 3
CH 4 -<>- CH 3 _,__ CH 4
'E 14
~
'E 14
_,__ CH 4 'E 14 _,__ CH 4 'E 14
' -<>- CH 5 -<>- CH 5 0 0 -<>- CH 5
2 2 ~ 12 -<>- CH 5 ~ 12
:; 12 :; 12 "0 "0
ro
I'll I'll ro 0
.310 .3 10 ..3 10 ...J 10
Cl) ~
~ 8 ~ ~ 8 'iii 8
iii 'iii 8 cCl)
c c c
t! 6
Cl)
1- 6 t! 6 1- 6

4 4 4 4

2 2 2 2
0 ~~~~~~-~~~~~~ 0 ~~~~--~~~--~~~
0 0
-1000 0 1000 2000 3000 4000 - 1000 0 1000 2000 3000 4000
-1000 0 1000 2000 3000 4000 -1000 0 1000 2000 3000 4000
Strain (}I} Strain (}I) Strain (p)
Strain (p}

(c) BAF-2 (A side) (f) BAF-2 (B side) (i) BAF-5 (A side) (j) BAF-5 (B side)

20 ..,..---- - -- -------, 20
20 ~----------~====~ 20
-o- CH 1 -<>- CH 1 --- CH 1
18 -<>- CH 1 18 18
18 -o- CH 2
-o- CH2 -o- CH2 -o- CH 2
16 16 16 -<>- CH 3
~ cH3 16 -<>- CH 3 -<>- CH 3
CH 4
'E 14 --9-- CH 4
'E0 14 _,__ CH4 'E 14 _,__ CH 4 'E 14
--9-

-<>- CH 5 -<>- CH 5 -<>- CH 5 0 -<>- CH 5

2 g 12 ~ 12
:; 12 ~ 12 "0 "0
"0 C'll C'll
C'll
.3 10
C'll
.3 10 .3 10 .3 10
JV"\'1 ~ .. n
Cl) -"" ~
Cl)
~ ~ 8 ~ ~ 8
'iii 8 8 ~

~
'iii cCl) c
c cCl)
F Cl)
6
~
t! 6 6 1- 6 1-
1-
4 4
~
4 4
2 2 2

0 ~~)~~~~~~~~~ ~
2
0
~ 0
0
- 1000 1000 2000 3000 4000 -1000 0 1000 2000 3000 -1000 0 1000 2000 3000 4000
0 -1000 0 1000 2000 3000 4000
Strain (p} Strain (p} Strain (p)
Strain (p)

(c) BAF-3 (A side) (k) BAF-6 (A side) (I) BAF-6 (B side)

(f) BAF-3 (B-side)

Fig. 3.9 Strain Distribution on the Circular Plate (continued) Fig. 3.9 Strain Distribution on the Circular Plate

78 79
y
p
Nodal Displacement
TU = O
Separation
Nodal Force
TF = O
12.5
--r-------J-
14-- -.. .,. .t -- -O: X ---
Nodal Displacement 1
Nodal Displacement 1
-r
TU=O 1U>O
Nodal Force Nodal Force
TF>O TF =0

Fig .3. 10 Schematic Figure of Procedure considering Boundary Non-Linearity

0>
,...

52

(unit : mm)

Fig. 3.1 1 Example of Finite Element Discretization of BAF Modei(AS-2)

80 81
 m of Symmetry Tensile Load 16 .------------~ 16 ,------------~

14 14
Flange Plate
~ 12
c
g 10
.a.
A' "0
nl
B' .3 8
G)

Bott Pre-stress Force Swface CO!lSidering ConlacVSeparate State 'i

c 6
~
4

A 2 2

B B' 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
zszszszszs Separation (mm) Separation (mm)

(a) AS-I (b) AS-2

16~-----------~

14
Fig. 3.12 Boundary Conditions of BAF Model
12
~
c
g 10
"0
nl
8

r
0
..J
Q)

'i
c 6
G)
1-
High Strength Bolt 4

IDr--- ---. 2

0.0 0.5 1.0 1.5 2 .0 2.5 3.0 3.5 4.0 4.5 5.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Separation (mm) Separation (mm)

(c) AS-3 (d) AS-4

Bolt Pre-stress Force 16

14
(a) Introducing Bolt Pre-stress Force
~ 12
c
g 10
"0
High Strength Bolt Tensile Load nl
0
..J 8
G)

~ 6
cQ)
1-
4

2
-
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Separation (mm)

(b) Applying Tensile Load (c) AS-5

Fig. 3. 13 Loading Procedure of the Analysis Fig. 3. 14 Load- Deformation Curves (Analysis) (continued)

82 83
 16 ~----------------------~ 16

14 14

12
~
c
g 10
"U
C'G
0 8
...J
(I)
~ 6
c
(I)
1-
4
It High Strength Bolt
2 2

0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Separation (mm)

(f) AM-I
Separation (mm)

(g) AM-2
L212121212m
16 ~----------------------~ 16
(b) AS-5 (Tensile load lOkN)

14 14

~ 12
c
g 10
"U
C'G
0 8
...J
(I)

~ 6
c
(I)
1-
4

0
0.0 0.5 1.0 1.5 2 .0 2.5 3.0 3.5 4.0 4.5 5.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Separation (mm) Separation (mm) (a) AS-2 (Tensile load 80kN) (c)AS-5 (Tensile load 20kN)

(h) AM-3 (i) AM-4

16 ~----------------------~ 16
Yielding Element
14 14

~ 12 ~ 12
c c
g 10 .10
"U "U

~ r-~r..-----'
C'G
0 8
...J
8 (I)

~ 6 ~ 6
c(I) Fig. 3.15 Progress of Yiclding Area
~ 4 1-
4

2

0
~
0.0 0.5 1.0 1.5 2 0 2 5 3 0 3.5 4.0 4.5 5.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Separation (mm) Separation (mm)

(j) AL-l (k) AL-2

Fig. 3.14 Load-Deformation Curves (Analysis)

84 85
 400 . . . - - - - - - - - - - - - - - - - - - - , 200 ~-------------, 200 . - - - - - - - - - - - - - - . 200 . - - - -- - - - - - - - - - - .
180 180
350
160 160
...... 300
E eE 140 eE 140
$c 250 1 120 1 120
0 0
Plate Stiffness Plate Stiffness
g200 ~100 ~100 Bolt Stiffness
~ Ill
:E~ Plate Stiffness Bolt Stiffness
150 -l--~~...-----__;,...;,;;,;,.:c...::..:.::...:..:..::....~
G)
80 Cl)
80
D ~
(/) 60 ~
(/) 60
(/) 100
Total Stiffness 40 40 ~~~-~~MN~~--~~~
Total Stiffness
50 20 20
Total Stiffness
0 +---.--...-----r- ........_. o +--~--.--,--~--~ o +-----.----r---..--~~~ o +---,---.----.--.....::"lllll~~
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
Tensile Load (ton f) Tensile Load (tonf) Tensile Load (tonf) Tensile Load (ton f)

(a) AS-I (b) AS-2 (f) AM-1 (g) AM-2

200~------------------------. 100 100 50

Bolt Stiffness : 176.4 (tonflmm) Bolt Stiffness : 117.6 tonf/mm Bolt Stiffness : 176.4 (tonf/mm)
90 45
80 80 40
150
eE e eE 70 e 35
;;::
c
Bolt Stiffness $c 60 ;;:: 60
E
1 30
c
0
e1oo g g 0
50 ~ 25
~ ~ ~ ~
G) G)
40 G)
40 20
...
G)
:E :E
D

(/) 50 1 - - -
Plate Stiffness (/) Sep.
...:E
(/) 30 ~
(/) 15
20 Total Stiffness 20 10 Plate Stiffness
Total Stiffness
Plate Stiffness 10 5 Total Stiffness
o +--~-~-~--~.,.~ 0 0 0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
Tensile Load (tonf) Tensile Load (tonf} Tensile Load (tonf) Tensile Load (tonf)

(c) AS-3 (d) AS-4 (h) AM-3 (i) AM-4

4.0 . . . - - - - - - - - - - - - - - . 100 50
Bolt Stiffness : 352 .8 (tonflmm)
Bolt Stiffness : 92 84 (tonf/mm)
3.5 90 45

3.0 80 40
Bolt Stiffness
e eE 70 eE 35
gc 2.5 ;;:: 60 Plate Stiffness 1 30
c
g 2.0 Plate Stiffness B 50
0
e 25 Plate Stiffness
:II l ~

..
G)
:E 1.5 Total Stiffness G)
40 ~ 20
"D
(/)
1.0
...
:E
(/) 30 (/) 15
Total Stiffness
Total Stiffness
20 10
0 .5
10 5
0 0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 2 4 6 8 10 0 2 4 6 8 10
Tensile Load (tonf} Tensile Load(tonf) Tensile Load (tonf}

(c) AS-5 (j) AL-l (k) AL-2

Fig. 3.16 StifTness-Load Curves (Analysis) (continued) Fig. 3.16 Stiffness-Load Curves (Analysis)

86 87
 400 0.6 200 .-------------------~ 0.6
--<>--- Analysis
350 - - Evaluation I~ '".,,.,.
- - Evaluation
Error 0.5 05
Error
~ 300
E
Py : Yielding Load of the Bolt
.s
c
0
250
0.4 ....
0
t:
w
0.4 ....
0
t:
w
~ 200 Plate Slitfness
Thicker Circular Plate 0.3 >Q) Q)
03 .,
en
en
Plate Stiffness
., Bolt Stiffness
>
<II
Q) Qi G)
~ 150

I
., 0.2
0::
0.2 0::
en
100
Midiwn Circular Plate

50 ~--.-~.~~.~~.~~~~--~
0.1 0.1

0 ...J.--.!!.__,.___.t.--r.:..
_~--.-:~~~~~ 0.0 0.0
0 2 4 6 8 10 0 2 4 6 8 10
'Thinner Circular Plate Tensile Load (tonf) Tensile Load (tonf)

(a) AS-1 (b) AS-2

200 .------r======;~ 0.6 100 . - - - - -- - - - - - - - , - 0.6

--<>--- Analys is Bolt Stiffness : 176.4 (tonf/mm)
- - Evaluation
Tensile Load Error 0.5 --<>--- Analysis 0.5
80 - - Evaluation
Error
0.4 5 E' 0.4 ....
Bolt Stiffness E
;;::: 0
Fig. 3.17 Pattern of Stiffness-Load Curves t: 60
w c
0
tb
~ ~
0.3 ~
0.3 .,
<II ~
.,
<II
G) Q)
40 G)
0.2 0:: .,~
en
0.2 0::

20 Total Stiffness 0.1

Plate Stiffness
0
1.2 , - -- - - - -- - - -- ----., 0 2 4 6 8 10 0 2 4 6 8 10
Tensile Load (tonf) Tensile Load (ton f)

1.0 y = exp(-1 .5x8) (c) AS-3 (d) AS-4

0.8 4.0 , - - -- -- -- - -- - -----.- 0.6

Bolt Stiffness : 352.8 (tonf/mm)
3.5
--<>--- Analysis 0.5
;::..., 0.6 ~ 3.0
- - Evaluation
E Error

0.4
.s
c
0
2.5
Plate Stiffness
0.4 ....
g
w
~ 2.0 \ 0.3 ~ .,
<II

.'\
Q)
Qi
0.2 .,~ 1.5
0.2 0::
en

0.0 -t----y----,----,--~~---1
1.0

0.5
Total Stiffness '
...,

0.0 0.3 0.6 0.9 1.2 1.5

'I
0.0
X
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Tensile Load (tonf)

Fig. 3.18 Outline of the Function (c) AS-5

Fig. 3.19 Evaluation Results ofthe Stiffness (continued)

88 89
 200 0.6 200 0.6

180
---<>- Analysis I 180 -<>- Analysis
- - Evaluation - - Evaluation 0.5
0 .5
160 Error 160
Error
e 140 0.4 ....
eE 140 0.4 ....
:c: 120
0
t: :g 120 0
....
....
w
w 0
0 Plate Stiffness Q)
~ 100 Bolt Sllffness 0.3 >
Q)
~ 100 0.3 >

., .
~
D
~ 'D
350
Q)
80
1'0
Q) Q)
80 I 1'0
Q)
e
D 0.2 0:: e
~
0.2 0::
(/) 60 (/) 60
40
40 300
0.1 0.1
20 20

0
00 0 0.0 'E 250
E
;;:: a
0 2 4 6 8 10 0 2 4 6 8 10 c
Tensile Load (tonf) Tensile Load (tonf) g
~
200
Q)
(f) AM-I (g) AM-2 c
!E
en 150
100 0.6 50 0.6 "C
Q)

90
Bolt Stiffness : 117.6 tonf/mm
45
Bolt Stiffness : 176.4 (tonf/mm) ro:J
(ij
-o-- Analysis 0.5 0.5 > 100
80 40

~
- - Evaluation UJ

eE 70
Error
0.4 .... eE 35 0.4 ....
;;:: 60 ....
0
.... :g 30
0
....
.... 50
c: w ---<>- Analysis w
g 50 0.3 ..,
Q)
0
~ 25 - - Evaluation 0.3 .Q)
~
> VI >.,
e
Q)

'D
(/)
40
30
, , s'' "\" 0.2
1'0
Q)
0::
VI
Q)

e
'D
(/)
20
15 t
Error 1'0
Q)
0.2 0:::
0
0 50 100 150 200 250 300 350

20
0.1
10
I Plate Stiffness 0.1 Stiffness (tonf/mm)

.'
10 5 Total Stiffness
0 0.0 0 0.0
0 2 4 6 8 10 0 2 4 6 8 10 Fig. 3.20 Comparison between the Experimental and Analytical Results
Tensile Load (tonf) Tensile Load (tonf)

(h) AM-3 (i) AM-4

100 0.6 50 0.6

Bolt Stiffness : 92.84 (tonf/mm)
90 ---<>- Analysis 45
- - Evaluation ---<>- Analysis
80
Error
0.5
40 - - Evaluation
0.5

eE 70
04 ....
0
eE 35 Error


0.4 ....
;;::
c: 60 t:
w :g 30 0
....
....
g 0
Plate Stiffness
w
50 ~ 25
Q)
0.3 > 0.3 >
Q)

~
Q)
D
1'0
VI
VI 'D
40 ~ 20
1'0
e
D Total Shffncu
02 0::
Q)
~
Q)
0.2 0:::
(/) 30 'D
(/) 15
20

10
0.1
'
0.1
10 5
0 0.0 I'
0 0.0
0 2 4 6 8 10 0 2 4 6 8 10
Tensile Load (tonf) Tensile Load (tonf)

(j) AL-l (k) AL-2

Fig. 3. 19 Evaluation Results of the Stiffness

90
91
 Chapter 4
Mechan ical Behavior of Split Tee Flange Joints
4.1 Introduction
In this chapter, the mechanical behavior of split tee Oangc joints under monotonic loading and cyclic
loading is studied. The split tee Oangc joint is a fundamental type of the high strength bolted tensile flange
joints and it is classified into short connection type. Overall view of the joint is shown in Fig. 4.1. As
sho\m in this figure, two flange plates (in other words, end plates) are jointed by the high strength bolts.
The mechanical behavior of this type of joints has been investigated extensively during 1970's for
architectural engineering applications. Based on the results obtained by these studies, various evaluation
methods of the maximum strength of such joints were proposed(refer to section 1.2), and results of these
studies became a part of the specification in the building design[ 1]. But proposed evaluation method for
maximum strength mainly depend on the experimental results; therefore, the applicability of it for various
joint types arc unknown. Therefore, it is thought that the general evaluation method for the maximwn
strength, which should be applicable for wide range of joints, is expected to be developed. However, the
studies in the past on such joints have not been extended to such ranges.
Recently, attention is paid to the advantages of such joints because of its structural simplicity, easiness of
erection, and superiority in the aesthetic design; so that the study on such joints is being actively carried
out in . On the other hand, the present circwnstancc that the draft design code of high strength bolted
tensile flange joints for bridge structures arc proposed by the subcommittee of establishment of design
code of JSSC based on the results obtained from the studies for both architectural and civil engineering
structures[2]. However, as discussed in that draft code, there are many teclmical problems on this type of
joints to be solved. l11erefore, in order to establish the rational design code for this type of joints, furt1 1er
studies should be suggested to be carried out.
The objective of this study is to investigate the mechanical behavior of split tee flange joints in
detail by not only the experiment but also finite clement analysis, especially for the joint which has
Yarious thickness and width of the Oange plate. Furthcnnorc, the joint stiffness, separation behavior and
cyclic behavior of such joints nrc nlso discussed.
92
4.2 Mechanical Behavior under Static Loading
4.2.1 Experimental approach
(a)Outline of tensile loading test
Extending the investigation on contact/separation behavior of high strength bolts and its adjacent flange
plate model to that of fundamental bolted joint, namely, the split tee flange joint, the tensile loading test is
carried out. Since the mechanical behavior of this type of joint is considered to significantly depend on the
thickness of the flange plate, it is dctennined to make the thickness of the Oange plate be variable, and the
effect of flange plate thickness on the mechanical behavior is investigated. The specimens prepared for
this experimental study are shown in Fig. 4.2 and dimensions of all the specin1ens are swnmarized in
Table 4.1. As shown in tl1is table, 3 types of specimens are prepared, and tltickness of the flange plate
among these specimens is different each other. 11te thickness of the flange plate of STJ , ST2 and ST3 arc
5 (nun), I 0 (nun) and 22 (nun) respectively. The widt11 of tl1e flange plate, the loading Jcngt11 between the
center of tlte bolt and the web plate and the diameter of the high strength bolt are tlte same among the
specimens. 11te high strength bolt used in the specimens are M 12 (F IOT) and the boll pre-stress force
(6.26 toni) is given to the bolt according to the specification of JJS[3]. Test setup used in tltis
experimental study is shO\m in Fig. 4.3, where tl1e electrically servo controlled hydraulic actuator is
utilized, whose loading capacity is +/-30 (toni) under static loading. 11te detail of testing machine can be
referred to section 2.2.1. As sho\m in this figure, 4 hinge connections used for loading apparatus can
eliminate tl1e ccccntric loading, so t..hat bending moment will not be subjected to tlte specimens.
Tensile load, separation between two flange plates and bolt forces are measured during the
loading, where applied tensile load is measured by the load cell built in the actuator. l11e separation
between two flange plates are measured by the strain gage type displacement transducer as shown in Fig.
4.4. On the other hand, Bolt force is obtained based on tl1e output of the strain gage installed into the
center of the bolt shank as shown in Fig. 4.5. The calibration sheet of the strain of the bolt to the bolt
force is made by the proof loading in fabrication process. The separation at the center of the specimen
where the tensile load is applied and the separation at the bolt position which is useful infonnation for
investigation of the bolt elongation are also measured as shown in Fig. 4.6.
The loading is operated by the micro computer; the loading conunand is sent to analog-
controller of actuator through GP-IB. The control of applied load is made by the displacement of the
actuator. In this experiment, tl1e loading is continued until the maximum loading point is obtained. In
addition, on line measuring arc made by the same computer.
(b) Results and discussions
The load vs. separation curves obtained from the experiment are shown in Fig. 4.7. The horizontal axis
shows the separation between two flange plates, and the vertical axis shows the applied tensile load. It is
understood that the thicker the flange plate is, the higher the initial slope of the curve is obtained.
93
 particularly, m case of thicker flange plate, it ts understood that the separation is relatively small; so that
it is considered that the dcfonnation of the flange plate is not significant. In addition, as for the failure
mode, the high strength bolt is severely pulled and failure occurs at the bolt thread. On the other hand, in
case of thinner flange plate, the defonnation of the flange plate is severe and the separation between two
flange plates occurred at the early stage of loading; namely, the flange plate is bent between the location
of the bolt and the loading point and the failure only occurs at the flange plate; whereas, the high strength
bolt is not observed to have severe dcfonnation.
The load vs. bolt force curves obtained by the experiment are shO\m in fig . 4.8. The horizontal
axis shows the bolt force and the vertical axis shows the load which is one half of the applied tensile load
to compare with a change of the bolt pre-stress force. Here, the bolt force is calculated based on the
calibration table. It is found that in case of the thinner flange plate, the bolt force increases at the early
stage of loadmg; Particularly, before the load reaches to the bolt pre-stress force. On the other hand, if the
split tee flange joint has thick flange plate, the bolt force is considered to not increase until the applied
tensile load reaches to the bolt pre-stress force. Therefore, the considerable thick flange plate should be
required in order to prevent the increase of the bolt force.
4.2 .2 Analytical approach
(a) 3 -dimensional analysis for split tee flange joints
In this section, 3-dimcnsional analysis for split tee flange joints is discussed. The analysis is carried out in
order to investigate the mechanical behavior in detail, especially, the contact/separate behavior, the
stiffness of the joints and the defonnation of the bolts. Generally speaking, the numerical analysis is
expected to be very powerful for investigation of mechanical behavior quantitati,cly because of its
versatility and its advanced technology. From the view point of design for the tensile joints in practice, 2-
dimcnsional analysis is desired due to the simplicity compared with 3-dimensional analysis. Since 2-
dimensional analysis for this model may have some difficulties due to non-unifonnity of the shape in
width direction and the local deformation characteristic is focused in this study, 3-dimensional analysis is
carried out.
The analytical model for the split tee flange joint is sh0\\11 in Fig. 4.9. This model is one eighth
of overall joint by considering axisymmctry of structures. Since the gap between two flange plates
increases as the tcnstle load is applied, not only the material non-linearity but also boundary non-linearity
must be considered in this analysis. That is, the boundary condition on the contact surface is changed
based on the nodal force and displacement at each loading step. The procedure to change the bow1dary
condition is the same as the analysis used for the BAF model(refer to section 3.3.2). This 3-dimensional
finite clement analysis program is coded by extending the 2-dimcnsional analysis program described in
Chapter 2.
94
All the cases of the analysis arc listed in Table 4.2. The thickness and the width of the flange
plate are varied. Because it is considered that tl1e mechanical behavior of split tee flange joints are
significantly affected by tl1e geometrical dimensions of tl1e flange plate. The thickness of the flange plate
is detennined to be 10 (nun), 15 (mm) and 22 (mm), and the width of the flange plate is detcnnined to be
63 (mm), 78.75 (nun) and 94.5 (mm), where the width, 63 (nm1) is tl1e same as that of test specimen and
standard among analytical cases. Other dimensions of the models for all the cases arc dctennined based
on the specimens used in the experiment. Here, attention is paid to the capability to model with various
shapes of tl1e flange plates and the finite clement discretization is made by the tetrahedral elements sh0\\11
in Fig. 4.1 0. As for the discretization by finite clement in this study, the anal)1ieal model is discreti.r.ed by
the hexahedrons at first; and then, each hexahedron is further discretized into 24 tetrahedrons. ln tlus
figure, discretization by tl1e hexahedrons is only sh0\\11. TI1e discretization of the bolt and its adjacent
structural elements where high stress concentration is predicted to occur is made by the fine mesh. The
number of tl1c nodal points and the elements in the case tl1at a specimen has the tl1ickcst and the widest
flange plate(ST-A8) is 4,531 and 18,540 respectively.
Boundary conditions of tl1is model are also shown in Fig. 4. 11 . The surface that two flange
plates contact each other is tl1e plane to consider the boundary non-linearity, namely, separation or contact
state. At tlus plane considering contact/separation conditions, the friction is asswned to be infinity;
tl1erefore, tl1e slip of contact surface is not assumed to occur. On the otl1cr hand, on the surface where tl1e
bolt and tl1e flange plate contact each other, such a boundary non-linearity is not considered and these arc
asswned to be continuous. The material constants for the bolt and the flange plate are provided
independently as sho\\n in Table 4.3. In addition, the washer is neglected in this model. Because it is
considered that there exist only little effect on the overall mechanical behavior of the joint. Moreover, the
bolt U\fead is also not exactly modeled but modeled by unifonn cross section same as bolt shank.
However, different material properties are used for the bolt tlU"cad. Material properties for the bolt shank
and tl1c bolt tl\fead used in the analysis arc also shown in Table 4.3. Referring the specifications of JIS[3),
the material properties of tl1e bolt tlU"ead is dctennined based on tl1at of the bolt shank by taking into
account the difference of effective cross sectional area. Furthcnnore, the stress-strain relation used in this
analysis is assumed to be clastic perfectly plastic.
The procedure of tl1is analysis is briefly sununarized in tl1e followings: at first, bolt pre-stress
force (6.26 tonf specified by JIS) is applied at the end of the bolt as unifonn load. Secondary, after this
bolt pre-stress force is obtained, the bolt end is fixed in vertical direction(z-dircction). finally, unifonn
tensile load is applied along the edge of the tee web plate Tile flow of the procedure of this analysis is
shown in Fig. 4.12.
(b) Results and discussions
1) Load-separation relation and yield strength
TI1e load vs. separation curves of all the cases obtained from the analysis arc sho"n in Fig. 4.13. In these
95
 figures, the honzontal axis shows the separation at the center of the flange plates. TI1e vertical axis shows
the applied tensile load diYided by the width of t11c flange plate; that is, the applied load per unit width. In
addition, the yield strength for each case are sununariLcd in Fig. 4.14(a). Horizontal axis is widt11 of the
flange plate nonnalized by the standard width 63 (mm) and t11c vertical axis is the yield load divided by
the width of the flange plate, namely, yield load per unit width. Here, the yield load is defined by the point
"Y" on the load-separation cun'e obtained from this analysis as indicated in Fig. 4.14(b ). TI1is point 'Y"
is defined as the intersection of tJJC tangential lines in the clastic area and plastic area.
From these figures, in case of the thinnest flange plate that is I0 (mm), it is understood that the
shape of curves is almost same in spite of the difference of flange plate width. On the other hand, in case
of the thid.cst flange plate that is 22 (nun), it is found that the post yielding behavior is di1Terent, and tl1e
yielding strength decreased gradually as the width of the tee web plate becomes large. Moreover, in case
of the medium thickness among anal)1ical cases, little difference of the load-separation cune can be
found. TI1ercforc, it is concluded that if the tensile yield strength is evaluated by tensile applied load per
un1t width, the thinner flange plate makes no di1Tcrcnce on the strength but thicker flange plate makes
s1gmficant d1ffcrcncc on the strength. 1l1is phenomenon is caused by the difference in tl1e failure mode of
the spl1t tee flange joint. In case of t11e tJunncr flange plate, the flange plate is considered to be bent
unifonnly in width direction and it failed, so t11at the tensile yield strength per unit width is almost same.
On the other hand, in case of thicker flange plate, the behavior is considered to depend on the failure of
the bolt, not the flange plate. 11lcrefore, if the estimation of yield strength is made based on the tensile
load per unit width, the yield sLTength becomes lower as the width of the flange plate becomes larger. As
shO\m in Fig. 4.14(a), whether the behavior is mainly aiTectcd by the flange plate or by the bolt can be
judged according to the yield strength vs. width relation. In other words, if the yield strength is constant in
spite of the diiTcrcnt width, the behavior of the joint depends on the flange plate; on the other hand, if the
yield strength is not same for the different width, the behavior of the joint depends on the bolt. In addition,
the more decrease in yield strength as the width becomes large is obtained, the more significantly tJ1e
bcha' ior of the joint is aiTccted by the bolt. Furthcnnore, if the flange plate is assumed to be in elastic or
ngid, the cun c of) icld strength vs. width relation can be given by hyperbolic type.
It is discussed about the applicability of the existing drafi specification[2] for tl1e split tee
flange joint "hose thickness of the flange plate is 22 (mm), comparing "ith the analytical results. TI1c
strength obtained by the specification is also shown in Fig. 4.14(a). From this figure, it is found that tJJC
CYaluation of the) icld strength by t11c existing specification is in a good agreement with analytical results
and it is considered to be applicable to the case of thick flange plate. However, the error of the yield
strength evaluation IS considered to be larger as the ,, idth of the flange plate becOmes larger. 11lcrcfore
this evaluation procedure used in the specification should be modified by taking into consideration the
width as a parameter for accurate evaluation. In addition, for the split tee flange joint whose thickness of
the flange plate is out of the range by the specification, the yield strength evaluation is proposed and
96
carried out by the following procedure: 11le simple model as sho"n in Fig. 4. I 5 is considered. 11Jis model
is the beam witl1 the same cross section as t11at of the flange plate fixed at the edge of the web plate and
the location of t11e bolt center. 1l1en, the yield load is defined by t11c reaction force when tJ1e maximwn
stress reaches to the yield stress as the rotationally fixed end of t11is beam is displaced. TI1e result of the
evaluation obtained by this procedure is compared in Fig. 4.14(a). It is found from this figure tJ1at the
yield strcngtl1 is estimated to be a little lower than that of the ana1)1ical results. Accordingly, 1t 1s
considered that tl1e proposed yield strength evaluation method gives a safe value and it may be enough
applicable.
2)Load-bolt force relation
Tensile load vs. bolt force curves obtained by the analysis arc shown in Fig. 4. 16. TI1e horizontal axis is
bolt force, and the vertical axis is one half of the applied tensile load to tJJC model because of the
comparison with the bolt force. In addition, the tensile load vs. bolt force curve in case of the rigid flange
plate, is also shown in this figure. If the flange plate is rigid, the bolt force is kept constant until the
applied tensile load reaches to the bolt pre-stress force; and then the bolt force increases equally as the
applied tensile load increases.
It is understood from this figure that t11e bolt force increases before the applied tensile load
reaches to the bolt pre-sLTcss force for all tJ1e cases. Here, the increase of the bolt force in case of the
tl1inner flange plate is larger t11an tl1at in case of the thicker flange plate. Although it is considered to be
caused by t11e prying force, tlus increase of the bolt force particularly for the case of thinner flange plate is
considered to depend on the defom1ation of the flange plate not like the conclusion obtained in the past
studies(refcr to section 1.2). 111at is, in case of the thicker flange plate, it mainly depends on the prying
force, on the other hand, in case of tl1inner flange plate, it depends on the tension and bending of the bolt
caused by only the local defonnation of the flange plate. This will be furtl1er discussed in tJ1e following
section on the contact surface.
As for tl1e width of the flange plate, the change of t.he bolt force for each case witJ1 di!Terent
widt11 of the flange plate is almost the same at tl1e initial loading stage, however, t11e behavior of the bolt
force is di1Tcrent afier the difference of tl1e load-separation curve is obsened. It is found that tJ1e larger
the width of the flange plate is, the smaller the bolt force is at the same le\'el of applied tensile load. Since
the dcfonnation of the flange plate becomes large as the "idt11 of the flange plate is much narrower, t11e
increase of the bolt force is considered to occur.
3) 3-dimensional deformation behavior on the contact surface and the distribution of the nodal
force.
Dcfonnations of the contact surface for all the cases when the applied tensile load reaches to the yielding
load arc shown in Fig. 4.17. In addition, deflection modes along the centerline of the flange plate in the
longitudinal direction are also shown in Fig. 4.18. In Fig. 4. 18, the deflection is magnified by 20 times of
the computed deflection in order to show the state of the deformation clearly.
97
 From these figures, in case of the thinner flange plate, it can be seen that the flange plate is bent
at the bolt position and that there is almost no gap between two flange plates at the area from the bolt
position to the outer edge of the flange plate. Moreover, it is understood that the deflection along the
center line of the flange plate is smaller than that at the side edge of the flange plate. On the other hand, in
case of the thicker flange plate, no significant bending dcfonnation of the flange plate is observed, that is,
the flange plate is rotated and translated and its rotation center is located on the outer edge of the flange
plate. In addition, the significant gap at the bolt position is observed and it is considered that tlte large
plastic defonnation of the bolt took place. It is concluded that in case of tlte thicker flange plate, the bolt
thread is elongated severely, on the other hand, in case of the thinner flange plate, the flange plate is
defonned.
Distributions of the nodal forces on the contact surface when tlte applied tensile load reaches to
the yielding load arc shown in Fig. 4. 19. In case of the thinner flange plate, the nodal forces arc
distributed between the bolt position and the outer edge of the flange plate; on the otlter hand, in case of
the thicJ...er flange plate, nodal forces arc distributed only around the outer edge of the flange plate.
TI!crcforc, it is understood from the distributions of the nodal forces that in case of tlte thicker flange plate,
the mechanical behavior of the joint is affected by the prying force, on the other hand, in case of the
tltinncr flange plate, tlte mechanical behavior of the joint has no relation with t11e prying force. As for tlte
widtl1 of the flange plate, in case of the narrower flange plate, the nodal forces tend to be distributed all
o,er the flange plate in width direction. On the other hand, in case of the wider flange plate, the nodal
forces do not tend to be distributed in width direction; namely, the nodal force at the side edge of t11e
flange plate is not large. This is caused by the difference of the deflection along t11e width direction due to
t11e less stiffness of the wider flange plate. Moreover, from these results, the effective width of the flange
plate where t11c applied tensile load would be carried should exist and this eiTcctive width of the flange
plate is considered to relate to the bolt pre-stress force.
4) The stiffness-load relationship
Stiffness vs. load curves for all the cases obtained from the analysis arc sh0\\11 in Fig. 4.20. The
horizontal axis shO\\S nonnalitcd applied tensile load divided by 2 times of the yielding load of tl1e bolt.
ll1c vertical axis shows the stifTncss of the joint per unit width based on the load-separation curve which
has already sho,,n in Fig. 4. 13. The stiiTncss in this figure is tltc tangential slope of the load-separation
curve, and it is calculated by di' iding the increment of the load by the corresponding increment of the
separation.
It is found from this figure that the thicker the thickness of tlte flange plate is, the higher the
stiffness of the flange plate is obtained. In addition, it is considered that the initial stifTncss until the
nom1alized applied tensile load is about 0.2, is significantly afTcctcd by the thickness of the flange plate.
Morco\cr, it is found that in case of thicker flange plate, initial stiffness is so high, but its decrease is
significantly large. As for the" idth of the flange plate, in case of the thicker flange plate, it is found that
98
the shape of the curve is almost same for different width of tl1e flange plate; on the other hand, in case of
the thinner flange plate, tlle stiffness becomes different as the applied tensile load becomes larger, and tlle
smaller the width of the flange plate is, the higher the stiiTncss is kept. lltcrefore, it is considered that
there is almost no difTerence of the initial stiffness among the anal)1ical cases with various widtll of the
flange plate if tlle stiffness is estimated by tlte unit width of the flange plate. In addition, for both tlte case
of tlticker flange plate and the case of thiru1er flange plate, it is understood from this figure that the
normalized applied tensile load is smaller t11an unity when tlte sti!Iness becomes almost equal to zero.
ll1erefore, it is considered tltat tl1e load carrying capacity of tlte bolt is not utilized etrectively. Especially,
even in case of the thicker flange plate, the nonnalized applied tensile load is about 0.8 when the stiffness
becomes zero. ll1e additional increase of t11c bolt force occurs in addition to that of t11e bolt force caused
by tl1e applied tensile load. Accordingly, in order to prevent this additional increase of the bolt force and
to make it zero, the considerable tlticker flange plate may be required. Moreover, it is also found that the
stifiness at tl1e yielding load specified by the draft specification of t11e hjgh strengtlt bolted tensile joints
for bridge structurcs[2] is almost zero, so that tlte strength until the stiffness becomes to zero is expected
in this specification.
In order to assess tl1e effect of tlte initial bolt pre-stress force, tl1e stiffness of tlte high strcngili
bolt, the stiffness of the flange plate and total sti!Iness of the joint for all the cases are listed in Table 4.4.
Here, tllese values are divided by the width of tlte flange plate; therefore, tllese are tl1e stifiness per unit
widtll of tl1e flange plate. TI1e stiffness of the bolt is also detennined by using the model that assume tlle
bolt to be tlte bar. It is defined by the following equation.
(4. 1)
in which, K9 is the stiffness of the bolt, A 9 , tlle effective cross sectional area of tlle bolt, 9 , Young's
modulus of the bolt, /8 , one half of t11e clamped length of the bolt, liiJ3, t11e diameter of the bolt.
Next, the stiffness of the flange plate is also detennined by using the model as shown in Fig. 4.21. lltis
model is the plate which has three fixed edges and one free edge subjected to concentrated load at t11e
center of the free edge. ll1is concentrated load corresponds to the load applied to the bolt. The definition
of tlle stifTness of tlte flange plate is as follows:
(4.2)
in which, Kp is the stifTness of the flange plate, p, Young's modulus, h, the thickness ofthc flange plate,
v, Poisson's ratio, /p, the length between the edge of the web plate and t11e center of the bolt, wp, the width
of tlte flange plate.
In addition, the total stiffness of the joint is dctcnnincd by using the model in which the flange plate model
99
and the bolt model is assumed to be connected in series. l11e equations of the total stiffness is shown as are detem1ined based on tl1e yield strength of the joint, namely, 1/4 of yield strength and 1/2 of yield

follows. strength as the basic pattern. However, in case of thinner flange plate, yield strength is very low, so that
3/4 of the yield strength and 1.0 times of the yield strcngtl1 are dctcnmncd to be applied to tJ1e test

(4.3) specimens.
TilC specimens are carefully manufactured by paying attention to the welding section where the

in which, KT is the total stiffness, Ke, the stiffness of the bolt, Kp, the stiffness of the flange plate. flange plate and tee web plate are connected because t11e less fatigue strcngt11 due to t11c stress

From t11c results, it is found that the stiffness of the joint obtained by the analysis is higher than the total concentration is expected. As for the welding of the flange plate to the tee web plate, part1al group

stiffness for alithe cases. The ratio of the stiffness obtained from the analysis to t11e total stiffness is about welding is used, and after then, the notch smoothing of the welding surface is made by hand grinding in

3 in case of thinner flange plate or about 2 in case of thicker flange plate when the load level is about 0.2. order to reduce unexpected extremely local stress concentration.

l11ercfore, the increase of the stiffness is also expected for the case that has thinner flange plate. Testing machine by SHIMADZU is used in the experiment. l11c detail of the testing machine
can be referred to section 2.2.1 or 4.2.1(a). The test setup as used for the monotonic loading test is also
utilized. An example of time history of the applied tensile load is shown in Fig. 4.23, where a frequency is

4.3 Fatigue Behavior of the Split Tee Flange Joints set to 2.0 (Hz) by considering the loading capacity of the testing machine. TI1e procedure of the applying
the load is swmnarized as follows: at first, the tensile load is applied up to one half of the load range; after

4.3.1 Outline of fatigue test then, the sinusoidal tensile load is applied by the function generator. Tensile load and the displacement of

High strength bolted tensile joints have promising superiority such as easiness of t11e erection, good the actuator are only measured from the beginning to the end. At the specific nw11ber of cycles of loading

mechanical beha,~or due to high pre-stress force given to the bolt. Especially, according to the fact that as shown in Fig. 4.23 for example each I 0,000 cycles, the loading is made very slowly with frequency of

the variation of the bolt force can be kept as small as possible due to the bolt pre-stress force, high 0.001 (Hz). During tl1is loading cycle, the bolt force and the defonnation oftl1c flange plate in addition to

strength bolted tensile joints seems to have a good characteristics for fatigue behavior as well. However, tensile load and displacement of tl1e actuator arc measured continuously for assessing the damage

in order to apply tensile joints for bridge structures, it is important to investigate fatigue behavior of this accumulated in the specimens.

type of joints, very little study on fatigue strength or fatigue behavior of tensile joints has been carried out In this loading test, average bolt force and the defonnation of the bolt and the flange plate are

except for only a rcfcrcncc[4) in civil engineering. Considering such a present situation, the necessity of focused. l11e bolt force is measured by four strain gages glued on tl1e surface of the bolt shank. Since t11c

further study based on both analytical and experimental approaches is one of motivations of current study. bolt shank remains in the elastic state, the bolt force is calculated based on the average strain of four

Therefore, in order to understand the cyclic behavior of split tee flange joints, the cyclic loading test up to strain gages. The location of these four strain gages is shown in Fig. 4.24. It is detcnnined in such a way

fatigue failure is carried out and the fatigue behavior of split tee flange joints is discussed based on the that the bending dcfonnation of the bolt can be in\'estigated. Moreo\'er, t11e strain gages to evaluate stTess

experimental results. concentration are also glued on the flange plate in order to investigate tl1e local bending defonnation of

In this fatigue test, the effect of the thickness of the flange plate on the mechanical behavior of the flange plate. 11lC location of these strain gages arc dctcnnined by considering the possible distribution

tl1e joint under cyclic loading is focused because it is understood from the discussion in the previous of strain on the flange plate surface ncar the bolt head as well as tl1e web plate and shown in Fig. 4.24.

section that the thickness of the flange plate is very important parameter to control the mechanical Total nwnber of the measuring points including the strains of the bolt and the flange plate is 48.

behavior of the joint. Dimensions of all the specimens arc shom1 in Fig. 4.22. Geometrical configurations l11e complete measurement of strains is carried out for t11c pre-set nwnber of loading cycle such

of each specimen arc the same except for the thickness of the flange plate. In addition, the high strength as each I 0,000 cycles. l11is measurement is carried out through high speed data logger with sampling

bolts used in these specimens arc M 12(FI OT). In order to compare the behavior Lmdcr cyclic loading to speed of I 000 points per a second.

that under monotonic loading, dimensions of the specimens are dctcnnincd to be same with those of the When the specimen is assembled by tightening the high strength bolts, t11e bolt pre-stress force
specimen used in the previous monotonic tcns1le loading test. is gi\'en up to 6.26 (toni) according to the specification of JIS[3] by checking the strain reading of the bolt

Stress ranges applied to each specimen arc tabulated in Table 4.5 and the yield strength shank

obtained from the monotonic loading test is tabulated in Table 4.6 for reference. Here, the stress ranges

100 101
 4.3.2 Results of fatigue test
(a) Fatigue failure mode
The number of loading cycles at fatigue failure arc tabulated m Table 4.7, \\here the fatigue failure is
defined by c1thcr breaking of bolts or large dcfonnation of the flange plate due to cracking at toe of tee
web plate less than 2,000,000 cycles. The modes of fatigue failure is classified as follows: In case of
thinner flange platc(SPT I B, SPT I C), the specimens arc failed by the crack at the toe of welding section
of the flange plate to the web plate. On the other hand, in case of thicker Oangc platc(SPT2A), the
specimen failed at the boll thread near the nut. The location of these failure is shown in Fig. 4.25. It is
understood that the failure mode under cyclic loading is similar to that of the dcfonnation mode under
monotonic loading. Therefore, the fatigue strength should be dctcnnined by either the fatigue strength of
the \\clding section or that of the high strength bolt. Especially, the fatigue strength of the welding section
highly depends on the thid. ncss of the flange plate as well. For example, in case of thicker flange plate,
the fatigue failure at the welding section docs not occur because of the low bending stress on the flange
plate surface. On the other hand, in case of the thinner flange plate, the fatigue failure occurs at tl1e
welding scct1on due to the high bending stress as well as high stress concentration at the same location.
For the cases that specimens did not fail, namely, fatigue life is more than 2,000,000 cycles,
bolts arc chccJ..cd carefully. There exists no remarkable change of bolts such as bending or torsion both at
the bolt shank and the bolt thread for both thicker flange plate case (SPT2C) and thinner flange plate
casc(SPTlA). In these cases, it is considered that the applied load is not large to cause fatigue failure.
Particularly, the additional stress increase given to the bolt due to applied tensile load to the joints is not
high; therefore, it is thought that the bolt did not fail. In addition, bending stress of the flange plate is also
not large enough to cause the fatigue failure at the toe of welding section of the flange plate to the web
plate.
For the case that stress range of the bolt is predicted to be the highest, namely SPT2A, the load
range of the boll is up to 20% of the applied load range to the joint. Therefore, it is understood that the
applied tensile load range should be limited to very low level by taking into account the high bolt pre-
stress force given initially, namely 80% of yield strength.
(b}Change of the bolt force under one cycle
TI1c d1fTcrcnce in time history of both the boll force and applied tensile load during a particular loading
cycle arc compared 1n Fig. 4 26. The horizontal axis shows elapsed time for a loading cycle, and tile
vcrt1cal axis sho,,s the bolt forcc(lcft) and the applied tensile load(right). The bolt force is calculated
based on a'craging the readings of 4 strain gages on the boll shank.
Regardless of the thickness of the flange plate and applied tensile load range, time history of the
bolt force is summarit cd as follows: Since the applied tensile load is given in the form of the sinusoidal
wa\e, the load is applied to the specimen up to 1/2 of the load range at first. After then, the cyclic loading
is given and the bolt force is varied with the same frequency of applied tensile load. It is found from these
102
ligures that the higher tl1e tensile load is or the thinner the flange plate is, tile larger tile increase of the
bolt force is. In addition, it can be understood tl1at the decrease of bolt force while tile applied load 1s Jess
tl1an mean value of applied tensile load is rclath ely small and this relates to tl1e release of the
compressive force between two flange plates gi,cn by the bolt pre-stress force. TI1is minimum value of
the bolt force corresponded to the bolt pre-stress force. Comparing time history of the bolt force at
different nwnbcr of loading cycles, it is found that the time history of the bolt force at tl1e initial cycle is
quite difTerent fom1 those at following loading cycles. That is, at tl1e initial cycle, tl1c bolt force after one
cycle is completed is lower than that before the loading cycle. On tl1c oilier hand, at the following loading
cycles, tl1e bolt force before at the beginning of loading cycle is almost the same witll that at the end. It is
concluded that loss of the bolt force is considered to occur at the early stage of cyclic loading.
Nun1ber of loading cycles vs. the maximum and minimwn bolt force relation for all the cases
arc sh0\\11 in Fig. 4.27. TI1e horizontal axis shows the number of loading cycle, the vertical axis shows t11e
bolt forcc(left) and the amplitude of the bolt force(right). The amplitude of the bolt force means the
increase of the bolt force during a loading cycle. As for tl1c case that the fatigue failure did not occur, tl1at
is, SPT 1A and SPT2C which has different tllickncss of tl1c flange plate, the an1plitude of the bolt force is
similar each other in spite of tl1e difference in tl1e thickness of tl1e flange plate. Here, it is also found that
some amount of tl1e bolt force loss occurs at the initial loading cycle; after then the bolt force decreases
very little cycle by cycle. This tension loss of the bolt force is 3.0 (%) of tile bolt pre-stress force given at
the beginning and this obsenration is good agreement with the result obtained in tile short tem1 tension
loss test[5]. The tension loss at tl1c initial loading cycle is considered to be caused by not the applied
tensile load but the plastic defonnation of the weak layer on the surface of tile flange plate such as paint
and rust. In addition, it is found that the amplitude of the bolt force is very small as compared to that of
tl1e applied tensile load. For exantple, in case of the thinner flange plate, the ratio of tl1e amplitude of the
bolt force to that of tl1e applied tensile load is 5-7(%), and in case of thicker flange plate, t11c ratio is
about 3 (%). TI1is results is also in good agreement with those obtained from the monotonic loading test,
that is, the bolt force does not increase until the applied tensile load reaches to the bolt pre-stress force .
Especially, in case of thicker flange plate, the amplitude of the bolt force remained to be very small, so
tl1at the tltickcr flange plate should be desirable from the view point of the fatigue resistance of the bolt.
On the other hand, in case of tl1e thinner flange plate, the reason why this ratio is a little higher than t11at
of the tlticker flange plate is considered that the bending force is given to the bolt head and then the bolt
force tends to increase. As for tl1c cases tl1at specimens failed by cracking, tl1at is SPTI C and SPT2A, the
tension loss of the bolt force is also compared against the case that specimens did not failed at cracking.
TI1c ratio of tl1e amplitude of the bolt force to that of the applied tensile load is about 18-20 (%) for both
tl1ickcr flange plate and thinner flange plate. This ratio is very higher than that of the unfailcd cases.
Moreover, it is found that the amplitude of the bolt force becomes very large suddenly at tl1e
loading cycle just before the failure for both thinner and thicker flange plate cases, and that the thinner the
103
 nwnber of tl1c measuring point whose location can be referred to Fig. 4 24. ll1cy are nwnbercd from one
thickness of the flange plate 1s, the higher the increase of the bolt force is. ll1e reason of this phenomena
side of the flange plate(left) to the other side(right) as sho"n in Fig. 4 24. Since the strain gages used in
is thought as follows. In case of the thicker flange plate, fatigue fatlurc occurs at the bolt, and the bolt
the fatigue lest is for measuring stress concentration, the local dcfonnation of flange plate can be
thread is considered to yield. ll1crcforc, it is considered that the minimwn bolt force decreases due to
examined.
plastic elongation of the bolt, and that the applied tcnstlc load is directly transferred through bolts because
In case of the iliickcr flange plate, it is fowtd that tl1e strain distribution on the flange plate is
of less compressive force between flange plates. On the other hand, in case of the thinner flange plate, it
almost the same as that at the different loading cycle and its magnitude of ilie strains is very small. For
also can be seen that the maximum bolt force increases and the minimwn bolt force decreases. ll1e
example, the maximum strain of the flange plate for SPT2A, \\hich failed at the bolt, IS about 500
increase of the bolt force is caused by the crack at the welding section of the flange plate to the web plate.
(micro); whereas, the yield strain of the flange plate is about 1400 (micro) according to the material test.
Namely, the large bending dcfonnation occurs by the crack at the welding section, so that the bolt is
Accordingly, the flange plate is considered to be still clastic. Moreover, from the results of strain
considered to be pulled and bent severely. The decrease of the bolt force is considered to be caused by the
distribution, the flange plate is thought to be defonned as if a concentrated load is applied at the bolt head
yield of the flange plate due to the bending dcfom1ation as well as the yielding of the bolt. For both
location. Furthcnnorc, in case of the thinner flange plate which did not fail at t11e bolt or the toe of
thicker flange plate and thinner flange plate, although failure modes arc difTcrcnt each other, the increase
welding, that is SPTIA, the maximwn strain on the flange plate is about 800 (micro). On t11e other hand,
of the bolt force takes place. Therefore, the prediction of the fatigue life of the split tee flange joint can be
in case of other thinner flange plate, namely, SPT I B and SPTI C which failed at the welding section of t11e
made based on the increase of the amplitude of the bolt force.
flange plate to the web plate, the maximwn strain is about 1200 (micro). In these cases, the flange plate is
(c) The deformation of the Bolt
considered to yield. ll1e maximWll strain of the thinner flange plate is higher tl1an tltal of the thicker
Time-history of the strains of the bolts are shown in Fig. 4.28. llle horizontal axis shows elapsed time
flange plate; therefore, the large defom1ation of the flange plate takes place. In addition, if it did not fail
during a loading cycle and the vertical axis shows the strain. lllc location of each strain gage can be
by fatigue cracking such as the case of SPTI A, the strain distribution was not changed even as the
referred to Fig. 4.24. In this figure, the strain is set to zero when the applied tensile load reaches to the
nwnber of loading cycle increases and the flange plate remained in elastic state up to 2,000,000 cycles.
mean value. It is understood from these figures that the difference in strain-time curve for thicker flange
On the otl1er hand, if it did fail, such as the case of SPTI B and SPTI C, it is foWld that the strain
plate and thinner flange plate is found. In case of thinner flange plate, the strain of the bolt at the side
distribution is changed as the number of loading cycle increases. In addition, the pem1ancnt strain was
facing to the tee web platc(R-82) becomes significantly large tensile strain as the applied tensile load
observed.
becomes large; on the other hand, the strain at the side to the other edge of the flange plate(R-84)
llle time history of the stress on the flange plate at ilie location close to the tee web plate are
becomes relatively small compression(the decrease of strain). lllcrcforc, the bolt is considered to be bent
shown in Fig. 4.30. ll1e location to discuss the stress is I 0 (nun) far from tllC edge of the tee web plate.
significantly. Furthcnnorc, it is also understood that the bolt is not unifonnly elongated by tightening the
Here, the horizontal axis shows the nwnber of loading cycle, and the vertical axis shows ilie normalized
nuts and its in-Wlifonnncss is very sensitive in case of tl1inncr flange plate because of Jess stiffness of the
stress by nominal stress of the tee web plate, namely, applied load divided by cross sectional area of t11e
bolt. Whereas, in case of the thicker flange plate, all the strain of the bolt at four locations remain in
tee web plate. It can be seen from this figure that the nonnalized stress of the thicker flange plate and t11at
tension. ll1crcforc, it is considered that the bolt is unifonnly pulled, not bent. However, as the strain at the
?f tl1e thinner flange plate is 4.0 to 9.0 and 1.0 to 2.0 respectively, and that the nom1alized stress in case
side to the tee web plate is a little larger than that to the outer edge of the flange plate, so that the bolt is
slightly bent. of thinner flange plate is significantly higher than that in case of the thicker flange plate. In addition, as
shown in this figure, in case of SPTl C, it is observed that ilie nom1alized stress turns to decrease
According to the pre' ious study, 11 is thought that the bolt force may increase by the prying
suddenly from 60,000 cycles to 80,000 cycles; therefore, ll1is is due to stress redistribution of the flange
force; however, in accordance "ith tl1e thickness of the flange plate, the cause to have bolt force increase
plate, so that the cracking at the welding section of the flange plate to the web plate is considered to
can be concluded as follows: When the tensile load is applied to the joint, the bolt is bent and tensioned
occur.
by the flange plate. Therefore, the bolt force is considered to increase by its bending and tensile
dcfom1ation. Regardless of the thickness of the flange plate, the distribution is not w1ifom1 over the flange
plate, so tlwt the eccentric loading to the flange plate is more significant to the defonnation of the flange
(d) Deformation of the flange plate and maximum stress on the flange plate
plate. Furthcnnorc, it is concluded that it is difficult to assume the pure tension in the bolts. So that, hjgh
ll1c strain distribution of the flange plate at a particular loading cycle obtained from the strain gages for
strength bolts should be assessed against combined loading such as tension, shear force and bending.
all the cases arc shom1 in Fig. 4.29. The vertical axis shows the strain. The hori.t.ontal axis shows the

104 105
(e) Fatigue strength
S-N diagram obtained from the fatigue test is sh0\\11 in Fig. 4.31 . The horizontal axis shows the number
of loading cycle to the failure, and the vertical axis shows the applied load range ratio which is a load
range nonnalizcd by the yield load of the joint obtained from the monotonic loading test. It is calculated
using the following equation.
P = p (4.4)
n YL
in which, P,, P, YL is the normalized applied load range, the applied tensile load range and yielding load
obtained from the monotonic loading test, respectively. Yielding load is defined by the load level at the
intersection point between the tangential line at the initial clastic region of load-separation cune under
monotonic loading and the tangential line in the plastic region( referred to Fig. 4.14(b)).
It is found fonn this figure that the fatigue strength in case of thinner flange plate is higher than
that in case of thicker flange plate. Tilercforc, if the fatigue strength is assessed according to the yield
strength based under monotonic loading, the split tee flange joint with the thinner flange plate is
considered to be superior fonn the view point of allowable stress design although the yielding load used in
nom1alization is difTcrcnt each other.
Considering these results, it is concluded that the thickness of the flange plate should be very
important parameter to C\ aluate the fatigue behavior as well. Even though the thicker flange plate is used,
the allowable applied load range for the fatigue failure is about 6.5 (ton{) , which is much smaller than
total bolt-pre stress force. Therefore, in order to usc the high strength bolt efficiently, the additional
increase of the bolt force should be prevented from the viewpoint of the fatigue strength, so that
considerable thicker flange plate or longer connection type should be required. In addition, if the applied
load to the joint is small, the split tee flange joint which has thinner flange plate is desirable from the view
point of the allowable stress design because it is very cfTcctive for ultimate state, such as initial yielding
of tl1c joint. On the other hand, if tl1e applied load is larger, the joint which has thicker flange plate is
desirable. Finally, the fatigue strengtl1 and failure mode of the split tee flange joint can be easily
controlled by varying the thickness of the flange plate,. Furthcm10re, the design of the split tee flange joint
under cyclic loading could be more rational if this procedure to control failure mode and this
characteristics should be utilized aggrcssi\'cly.
4.3 .3 Simple Estimation of Fatigue Strength
Instead of the nonnalited load range, S-N diagram is dra\\11 for the absolute applied load range in Fig.
4.32. Fatigue strengths of the bolt and the flange plate arc estimated by the following procedure and these
arc compared in this figure. Here, only the fatigue strength under 2,000,000 cycles is considered: The
fatigue strength of the bolt is evaluated based on the guideline for fatigue design of steel structurcs(6], in
106
which, the fatigue strengtl1 is asswncd to be level "K4" for tl1e high strengtl1 bolts. As a result, it is
detennined to be 6.6 (kgf/nun2 ) for the bolt. Then, it is converted to the strength in force, 557 (kgf) by
multiplying the efTcctivc cross sectional area of the bolt prescribed in JJS[3]. The bolt force is further
converted to the applied tensile load by using applied tensile load \ 'S. bolt force cune obtained by tl1e
experiment under monotonic loading. As a result, t11e fatigue strength of tl1c joint in case of bolt failure is
evaluated. In case of the tllinncr flange plate, this strength is 4.1 (ton{), and in case of thicker flange plate,
this is 6.2 (ton{).
On the otllCr hand, as for the fatigue failure at the welding section of flange plate to tee web
plate, fatigue life estimation is made by referring the strcngtl1 of the joint as shown in Fig. 4.33 in tllC
guideline for fatigue design for steel structurcs(6], in which fatigue strength of the joint is classified to be
level "C" ( 12.8 kgflnun2 ). This strength is for the joint which has welding smoothened for quality control.
Then, the following model for the split tee flange joint as sh0\\11 in Fig. 4.34 is assumed in order to obtain
the working stress of tl1e flange plate. TI1is model is a beam whose both ends are fixed, and unifonnly
distributed load which corresponds to the applied tensile load at the web plate is given to the center of the
fixed beam. TI1e applied tensile load is obtained for the condition that the maximwn stress in this model
corresponds to tllc fatigue strengtl1 defined in the guideline for fatigue design. As a result, the fatigue
strengtll of the thinner flange plate and tl1at of the thicker flange plate is estimated to be 1.30 (toni) and
(6.25 ton{) respectively. Finally, the fatigue strength of tlle split tee flange joint is obtained by the smaller
value of allowable applied tensile loads defined by two possible fatigue strength. It is found from this
result, tl1at the fatigue strength of the thicker flange plate is higher than that of the thinner flange plate as
obtained in the fatigue test.
In addition, it is understood that in case of thinner flange plate, the fatigue strength is estimated
to be smaller than tl1at of the experiment, which is conservative. Moreover, fa1lure mode is estimated to
be by a failure of tllC flange plate and this is same as the experimental results. Accordingly, it is thought
tllat tlle estimation of the failure mode by proposed procedure is applicable in the case of thinner flange
plate. On the other hand, in case of the thicker flange plate, t11e split tee flange joint failed at the bolt and
fatigue strcngtl1 of the bolt is well-predicted. However, the poor estimation of the fatigue failure mode IS
obtained. Although the fatigue strength of the bolt is estimated by a little lower than that of the flange
plate which may result in failure of the bolt, the fatigue strength of bolt and that of the flange plate arc
almost same which cannot conclude that either failure mode should be considered. Therefore, in order to
estimate the fatigue failure mode accurately, the proposed model is still further modified.
107
4.4 Conclusions
ln this chapter, the mechanical behavior of the split tee flange joints, which is the most simple high
strength bolted tensile joints, under monotonic and cyclic loading is investigated experimentally and
analytically. As for the behavior under monotonic loading, the contact/separation behavior and the joint
stiffness is assessed in detail by using 3-dimcnsional analysis. On the other hand, as for the behavior
under cyclic loading, the cyclic loading test is carried out considering the local dcfonnation behavior,
such as the defom1ation of the bolts and the flange plate. Based on these results, the fatigue strength is
assessed and its estimation procedure is also proposed. TI1e following conclusions and future research
needs arc obtained as follows:
I) The mechanical bcha\'ior of split tee flange joints under monotonic tensile load depends on the
mechanical behavior of both of the flange plate and the high strength bolt, in particular it is
fatigue design for steel structures and good agreement with the monotonic experimental results
are sh0\\11.
In the future, in order to make this results reliable, further cyclic loading test is carried out for the joints
with various geometrical configurations. Furthennore, considering the practical usc for steel bridge
structures, the joints with multiple bolts should be considered. In special, the eiTectiYe area of one bolt and
the interaction among the bolt group on the mechanical behavior should be investigated.
References
I) Architectural Institute of Japan : Rcconuucndation for the Design Fabrication of High Strength Bolted
Joints, Maruzcn, Mar. 1993(in Japanese).
2) Japanese Society of Steel Construction : Recommendation for the design of tensile joints for bridge

significantly afleeted by the thickness of the flange plate. In case of the thinner flange plate, the structures (draft), Feb. 1993(in Japanese).
bcha\'ior depends on that of the flange plate and the high defom1ation capability is observed. 3) Japanese Industrial Standard Committee : Sets of High Strength Hexagon Bolt, Hexagon Nut and
On the other hand, in case of thicker flange plate, the behavior depends on that of the high Plain Washers for Friction Grip Joints(B1186), 1979(in Japanese).
strength bolt and high load carrying capacity is obtained, but the failure is brittle. 4) Y.Miki, K.Horikawa : Fatigue Behavior on Split Tee Flange Joints, Proc. of the 46th A.rulUal

2) The stiffness of the joint depends on the geometrical configurations of the flange plate: the Conference of JSCE, I, JSCE, Sep. 1991, pp. 606-607(in Japanese).
thicker and the narrower the flange plate is, the higher the stiiTncss of the joint is obtained. In 5) A.Ahmad, E.Watanabe, K.Sugiura, K.Hatanaka : Tension Loss in High Strength Bolts, Journal of
addition, the eiTect of the bolt pre-stress force on the stiiTness is made clear, in fact, the stiffuess Structural Engineering, JSCE, Vol. 41A, Mar. 1995, pp. 87-94.
increases by the bolt pre-stress force. 6) Japanese Society of Steel Construction : Recommendation for Fatigue Design, Apr. 1993(in
3) In addition to the prying force ciTcct understood in the previous studies, the increase of the bolt Japanese).
force is considered to be caused by pulling up the bolt by the local defonnation of the flange
plate.
4) The mechanical behavior under cyclic loading depends on the thickness of the flange plate as
well. In case of thinner flange plate, the joint failed at the welding section' on the other hand, in
case of thicker flange plate, the joint is failed at the bolt thread. The amplitude of the bolt force
under cyclic loading is small compared with that of the applied tensile load and the significant
eiTect of the bolt pre-stress force is observed for any thickness of the flange plate.
5) TI1c significant increase of the bolt force is observed before the failure of the joint for both bolt
failure and flange plate failure.
6) In case of the evaluation of the fatigue strength of the split tee flange joint based on the yielding
load of the joint under monotonic loading, the fatigue strength of the joint with thinner flange
plate is higher than that of the joint with thicker flange plate, therefore, the joint with thinner
flange plate is eiTcctiYc from the view point of the design based on yield strength. Moreover,
the simple estimation method of the fatigue strength is proposed using the guideline of the

108 109
 Tabl e 4. I L.1St 0 fS>peclmens Table 4 5 L.1St 0 fLo ad.mg Pattem
Specimen Thid.ness of Width of High Strength Bolt Bolt Specimen Minimunt Maximum Amplitude Stress Ratio Pmax/YL Loading
the Flange Plate the Flange Plate Pre-stress Force Load Load of (Pmin/Pmax) Period
ST-1 5 63 Ml2 (F10T) 6.26 (tonf) (Pmin) (Pmax) Load
ST-2 10 63 Ml2 (FJOT) 6.26 (tonf) SPTlA 100 2,500 2,400 0.040 0.49 2.0Hz
ST-3 22 63 Ml2 (FIOT) 6.26 (tonf) SPTlB 100 5,000 4,900 0.020 0.98 2.0Hz
(unit : nm1) SPTIC 100 3,750 3,650 0.027 0.74 2.0Hz
SPT2A 100 7,500 7,400 0.013 0.54 2.0Hz
SPT2B 100 11,000 10,900 0.0091 0.78 l.OHz
Tabl e 4. 2 L.1St 0 fA na1yt1ca
I . I Cases SPT2C 100 3,750 3,650 0.027 0.27 2.0Hz
ST-Al ST-A2 ST-A3 ST-A4 ST-A5 ST-A6 ST-A7 ST-A8 (unit : kgf)
YL : Yielding Load of the Specimen under Static Loading
Tilickness of 10 15 22 10 22 10 15 22
the Flange Plate
Width of 63 63 63 78.75 78.75 94.5 94.5 94.5
Table 4 6 Results of Monotonic Tensile Loading Test
the Flange Plate
Tilickness of Yielding Load
(unit : nmt)
Ute Flange Plate
I 0 nun 5.1 tonf

I .
Table 4 3Maten. aI Propcrt1es us ed.In tIte AnalySIS 22 Dilll 14.0 tonf
Young's Modulus Yielding Stress
Flange Plate 21,000 26.9
Bolt Shank a 1gue Tes t
T abl e 4.7 Resu ts oflLte Ft"
of 2 1,000 90.0
High Strength Bolt Specimen Nwnber of Cycle Failure Mode Pmax/YL
Bolt Titread to Failure
of 15,700 67.1
SPTlA - not failed 0.49
High Strength Bolt
SPTlB 17,848 FL 0.98
SPTlC 82,000 FL 0.74
SPT2A 139,480 BT 0.54

Tabl e 4 4 L.IS t 0 fSt"Ii

1 ness per a uLc
mt ngltI at EIS SPT2B 17,113 BT 0.78
ae t cellon
Stiffness ST-AI ST-A2 ST-A3 ST-A4 SPT2C - not failed 0.27
ST-A5 ST-A6 ST-A7 ST-A8
YL : Yielding Load under Monotonic Loading
Bolt 1.88 1.26 0.856 1.51 0.685 1.26 0.837 0.571 Pmax : Maximwn Load applied to the Specimen
Plate 0.0336 0.113 0.358 0.0269 0.286 0.0224 0.0756 0.238 FL : Failure at Ute Toe of Welding between Flange Plate and Web Plate
Total 0.0330 0. 10-J 0.252 0.0264 BT : Failure at the Bolt Thread
0.202 0.0220 0.0693 0.168
(unit : tonf/nun/mm)

110 111
 Hinge

Loading Frame
Specimen

Flange Plate
c---t--r- Hinge

Load Cell

Fig. 4.1 Overview of the Split Tee Flange Joint

Fig. 4.3 Testing Setup

Output/Input Line

63 mm

206mm
Tee Web Plate High Strength Bolt

I
ST-1 : 5mm
16 (unit: mm) 12
Flange Plate ST-2 : IOmm
ST-3 : 22mm

Fig. 4.4 Displacement Transducer for Measuring the Gap

Fig. 4.2 Geometrical Configurations of the Specimens

112 11 3
 ' Strain Gage for Measurment
.(of the Bolt Force

20 ,-----------~======~ 20 . - - - - - - - - - - -
18
-<>- ST-1
-o-- ST-2
18 ~Bolt Side A
Bolt Side B
16 -~::r- ST-3 16 Center
'E 14 'E 14
0 0
~ 12 ~ 12
nl nl
.3 10 .3 10
Q)
..!!
u; 8 u; 8
c c
~ 6 ~ 6
4 4
2 2J._
0~---,--~c-----,-----.-----l 0~,--.--~-.--.--.---.--.~
0 2 3 4 5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Separation(mm) Separation(mm)
High Strength Bolt
(a) Average Separation (b) ST-1
Fig. 4.5 Strain Gage buried into the Bolt

20 ~---------r========~ 20 ~-----------~
-<>- Bolt Side A
18 18
-o-- Bolt Side B
16 -~::r- Center 16

'E 14 'E 14
0 0
~ 12 ~ 12
nl nl
.3 10 .3 10
..!! Q)
"iii 8 ~ 8
c c
~ 6 ~ 6 -<>- Bolt Side A
4 4 -o-- Bolt Side B
-~::r- Center
2 2
0 ~----~r-.-,--.-.-.-.-~ 0~-.--~r-.-.--.-.--.-~~
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Bolt Side A Separation(mm) Separation(mm)

(c) ST-2 (d) ST-3

Measuring Point Fig. 4.7 Load-Separation Curves (Experiment)

Fig 4.6 Location of Measurement of Gap

114 115
 Tensile Load

20
18 - o - ST-1
- o - ST-2
16 --6- ST-3
cc 14
0
~ 12
"'0
ro
0
_J 10
Q)
u; 8
cQ)
I- 6 High Strength Bolt
4
2
Tensile Load
0
0 2 4 6 8 10 12 14 16 18 20
Bolt Force(tonf)
Fig. 4.9 Overview of the Analytical Model

Fig. 4.8 Load-Bolt Force Curves (Experiment)

220mm
52.0mm 29.0mm I 8.0mm
1/ v: v: ~
1/ 1/ 1% :% Side View
1/. ~ IO.Omm
~ ~
~
...,.. Contact Surface Contact Surface
ll .Omm 12.0mm
High Strength Bolt(M 12,F I01)

31.5mm Grand View

103.0mm

Fig. 4.10 Example of Finite Element Discretization of the Analytical Modei(ST-A I)

117
116
 y ~ ~ v ~
D
v v v t/. Side View 0.30
~ ~ ~ STF1
-o- STF2
E
t/. ~ ..........
0.25 ----c.- STF3
Contact Surface
~ Contact Surface E
E
-0-- STF4
-6---- STF5
High Strength Bolt(MI2,FIOn 1 0.20
A .9 ------ STF6
.........
"0 ---- STF?
<U 0.15 __..._ STF8
0
_J
Q)
u; 0.10
Grand View c
Q)
1-
0.05

0.00
in which 0 2 4 6 8 10 12
AB : Displacement in y -direction is fixed Separation(mm)
BC : Displacement in x -direction is fixed
B : Displacement in both x-direction andy-direction are fixed
Fig. 4.13 Load-Separation Curves (Analysis)
DE : Displacement in y -direction is fixed

Fig. 4.11 Boundary Conditions of Analytical Model

High Strength Bolt

II>I
15.
~ lflinge Plate
Pre-stress Force
(a) Step 1

Applied Tensile Load m High Strength Bolt

:;;y Boundary considering
'CiL----~~~--___,~I Flange Plate Contact or Opening

(b) Step 2, Step 3

Fig. 4.12 Loading Procedure Analysis

118 119
 0.30 - . - - - - - - - - - -
S- -t-,
h
t::. Design treng

-
E
0.25
-at.- TF = 22mm
= 15mm
- t t - TF
__.._ TF = 10mm
~_..l._--1--l"--J ~----, * t

E (Analysis) L-~---r-~~--~,-~ ~
1 0.20
~
0
'"0

~ 0.15
....J
0>
c 0.10
:g
<1>
)::

0
(a) Split Tee Flange Joint (b) Estimated Model

0.05 =
TF 10mm
m TF = 15mm Fig. 4.15 Estimated Model for Yield Strength
(Estimated Strength)
0. 00 J_------r----r--.------!===r==r=====;=:::::::::=====i
0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6
Width

(a) Yield Strength

'Yield Lo~d (Y)

Separation

(b) Definition of the Yield Strength

Fig. 4. 16 Load-Bot Force Curves (Analysis)

Fig. 4.14 List of Yield Strength of All Cases

120 121
 1.0 1.0 1.0 1.0
E EE
-
E
c
Q)
0.8

0.6
-c
Q)
E
0.8

0.6
E
~
Q)
0.8

0.6
E
.,
cC1>
0.8
0.6
E E E
Q)
t.>
1:1)
~
1:1)
120 ~
g ~
1:1)
"li ~0 ~ "li
100 1$' "li "li 120 ~
6
rJ>
ad
60
.
{::>;t-
6
rJ>
80 +~
60 {::>,
rJ>
6 6
rJ>
100 .
80 +
40 !:::-0 40 ~ 60 {::>,
20
0 v
# 0
20
v
# 40 ~
20 8'
LOcation - Y I LOcation - Y I 0 ..J
LOcation _Y 1
,rnrnJ ,rnrnJ ,rnrnJ
(a) STF l (b) STF2 (e) STF5 (f) STF6

E
.s
1.0
-
E
1.0
- 1.0
eE
1.0

cC1>
E
0.8
0.6
-c
E
Q)
E
0.8

0.6
E
--
E 0.8
cQ)
0.6
--
c
C1>
0.8

0.6
E E
~ C1>
~
()
1:1) 1:1) 8
g. "li
rJ>
120 ~
100 1$'
1:1)
"li 120 ~
1:1)
"li
6 6 80 +~
rJ>
6 100 . rJ>

60 {::>, 80 + 6
60 ,
40 ~0 40 .o~
20 0fli
60 40 20
0 ..J 20 #
LOcation _Y 1 LOcation 0 v0
,rnrnJ -Y (rnrnJ Location Lecafion
- Y (rnrnJ - Y (rnrnJ
(c) STF3 (d) STF4 (g) STF7 (h) STF8

High Strength Bot

High Strength Bolt

X
X

y y

Fig. 4. 17 Defonnation of the Contact Surface (continued)

Fig. 4. 17 Dcfonnation of the Contact Surface

122 123
 120 120 120 120

100 100 100 100

......... ......... ......... .........
80
E 80
~ E 80 E 80
-E
c:
Q)
E
60
-~
c:
60
-E
c:
Q)
E
60
-E
cQ):
E
60
8<0 8<0 Q)
(.)
ctl
Q)
(.)
ctl
Ci. 40 Ci. 40 Ci. 40 0.. 40
1/) 1/)
.~.::. 1/)

......
1/)
0 0 0
... ...=~:
0
----
--- 20
. .: :......: ..:....:.....................
:~==-=- ...
:._~~E:.... . . . ..
...:.~.. .. . . :. .....
.. ... . .................. ... ...
20 20 ........ 20 .-:::...__,
-
...~
. ......:..-=;,:.......
......... -
.. ....._._.. ..
... .... .
....... ,......
..... ..--:;.
.....;.-....::......
.__ . . . . .... ._.... .. .. . .
.::::.;:::::::.
0 ""::';..... .
~.
0 0 -- . 0 .. ..::=~:.:.: ....
0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20 40 60 80 100 120
Location(mm) Location(mm) Location(mm) Location(mm)

(a) STF1 (b) STF2 (e) STF5 (f) STF6

120 120 120 120

100 100 100 100

.........
E 80 'E 80 E' 80 -E 80
-E

E
c:
Q)
60
.s
-c:
Q)
E
60
- E
c:
Q)
E
60
-E
c:
Q)
E
60
Q)
(.)
ctl
Q)
(.)
ctl <0
8 8
<0
Ci. 40 Ci. 40 a. 40 a. 40
.. .. . .
:-~.
. ..~.
1/) 1/) 1/) 1/)

...............:......-:. ..:::::.;:................
:-::-=::.~.
:. :. ..:~:.. .
0 0 i:5 0
.:. .::::.~.....
. . ... .. :-~==::-: .....
. ...... ..__. .........
.-....-
..: ........:......:--...."::.:......
20 .::::::~:. 20 20 .:~.::::;:
::::-::=-::-: .... . 20 ..:.~.
. . .:
0
~

_
................
... ...... . ..
.. . ....
0
......":::,.':.',
. :::::::.-:........ ..
. . .....,. .. . 0 . -._
,~:...
.......
- .
0
:.-=:_.:.....
.......:--=
- . ... ...
t

0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20 40 60 80 100 120

Location(mm) Location(mm) Location(mm) Location(mm)

(c) STF3 (d) STF4 (g) STF7 (h) STF8

Fig. 4.18 Defonnation at the Center of the Model in Longitudinal Direction (continued) Fig. 4.18 Deformation at the Center of the Model in Longitudinal Direction

124 125
 200 200 600 200
c c <;::- c
~ 150
~
0
~ 150
Q)
._
(.)

0
~
Q)
._
(.)
0
500
400 -
~ 150
~
._
0
U- U- U- U-
(0 (0 (6 te
-g 166 ~~ "0
0
166~~ "0 120 ~ "0
0 120 ~
z 80 +-.;::,
z 80 z
0
100 ~ z 100 ~
60 , 60 ,+ 80 + 80
60 ,
+
60 ~,
40 ~0(;;- 40 ~0~ 40 ~ 40 ~(;;-
20 8' 20 8' 20 fJ'll 20 #
0 ...J 0 ...J 0 ...J 0 ...J
LOcation LOcation Location
Y (rnrn) - Y (rnrnJ Y (rnrnJ

(a) STFl (b) STF2 (c) STF5 (b) STF6

600 200 600 600

cOl 500 c cOl 500
~

0
8._ 400 -
~ 150
8._
c
~
8._ 400
~
Q)
._
(.)
400
U- 300 0
U- 0 0
U-
U-
(0 (0 (6
(6
166~~ 166~~
"0
166~~
"0 "0 120 ~ "0
z0 z
0 0
z 100 ~ z
0
80 + 80 + 80 + 80 +
60 ~, 60 , 60 , 60 ~,
40 ~ 40 ~0(;;- 40 0(;;- 40 ~
20 8' 20 {JI 20 ou~ 20 fJ'll
0 ...J 0 ...J 0 0 'v 0 ...J
Location Location LOcation
Y (rnrn) Y (rnrn) Y (rnrnJ

(c) STF3 (d) STF4 (g) STF7 (h) STF8

High Strength Bolt High Strength Bolt

X X

y y

Fig. 4. 19 Distribution of Nodal Force on the Contact/Separation Surface (continued) Fig. 4.19 Distribution of Nodal Force on the Contact/Separation Surface

126 127
 1.5 -o-- STF1
-o- STF2
-E
E
---t:.- STF3
-0-- STF4
--&- STF5
E
E 1.0 _._ STF6
;;:::: ---11- STF7
c: ___._ STFB Tee Web Pl ate High Strength Bolt
.....
0
.........
C/)

\
C/)
Q)
: 0.5
.,
en

Flange Plate
0.0 0.2 0.4 0.6 0.8 1.0
Load(Tensile Force I Bolt Yielding Force)

Fig. 4.20 Stifiness-Load Curves (Analysis)

Fig. 4.22 Dimensions of the Specimens for the Fatigue Test

c>
"'C
ro
0
--'
Q) -
Initial Loadin Level ~--
Modeling c;;
I
I
I
c:
Q)
I- Normal Loading Period ~
v
14----+--t--- ~
!
I
- Extra Loading Period
(a) Split Tee Flange Joint (Measuring Strains)
(b) Clumped Beam at both Ends

Time

Fig. 4.21 Model for Estimation of the Stiffness of the Flange Plate
Fig. 4.23 Time History of Applied Tensile Load

128
129
 High Strength Bolt High Strength Boli
R-B2 R-BI L-B2 L-Bl
7. ~ Bol1 7. - Bol1

7.2 ~ Boft2
7.2 - Bo2
- Tens~e Load 3
7.0 - Tensile Load 3
c- 7.0 c-
c- c c
g 6.8 0 'E
0 6.8 0
<:::-
~ 6.6
& 6 .
2~
~ i 6.6
& 6 .
2~
"'0
...J

Tee Web Plate s 6.2

6.0
1~
~
s 6.2
6.0
1~
c
41
1-
5.8 0 0
5.8
- Strain Gage for 5.6 5.8
Stress Concentration
0 200 .00 600 800 1000 0 200 400 600 800 1000
Strain Gage for Time (sec) Time (sec)
Axial Bolt Force (a) SPTIA (I cycle) (b) SPT lA (l ,OOO,OOOcycles)

Fig. 4.24 Location of Strain Gages glued on the Specimen

7.
~Bol1
7. - Bo1
6

7.2 - Bo.2 72 s
- Tensile Load 3
7.0 c- 7.0
~<:::-
c- c
g 6.8 0 'E
0 6.8 0
Bolt Head <:::-
8 6 .6
2;- ~ 6.6 3"0
~ e "'
0
"" 6.4
& ...J
& 6.4 ...J

Bolt Shnak
s 6.2
6.0
1~
Q)

cQ)
1-
s 6.2
6.0
22
.,;
1~
c

5.8 0
5.8 0
5.6 5.6
a) Bolt Failure 0 200 400 600 800 1000 0 200 400 600 800 1000
Time (sec) Time (sec)

(c) SPTA (2,000,000cycles) (d) SPTIB (lcycle)

Tee Web Plate

7.4
r--------,======~--~ 4
7.4
.---------.----------rr 4

7.2 7.2
3 3
7.0 7.0
C'
Welding g 6.8 '2 6.8
<:::- .9
~ 6 .6 i 6.6
High Strength Bolt 0 6.4
& 6.4
~- -~ s I.L.
2::0 6.2
6.2 0
!D 6.0
6.0
5.8 5.8
5.6
5.6
~--.----.---,---~-.---- 5.4 -'-----.----.----.-----.==.--
0 200 400 600 800 1000 0 200 400 600 800 1000
Flange Plate Time (sec) Time (sec)

(e) SPTIC (!cycles) (f) SPT l C (80,000cycles)

(b) Flange Plate Failure

Fig. 4.25 Location of the Fatigue Failure

Fig. 4.26 Time History of Bolt Force and Applied Tensile Load under One Cycle (continued)

130
131
 8 8
7.4 7.4
7 7
7.2 7.2
6 7.0 6
7.0 c- c-
ce 6.8 a
5 c:. ce 6.8 5 a
c:.
0
c:.
8
... 6.6 4-g
0
80
6.6 4-g
0 6 .5 0.190 8 1. 1
..J 0 6.4 ..J
af 6.4 3j! u. 3j!
6 .4 - Min
ill ~ 6.2 'ill
~ 62 2 cGl 2 cGl 0.185 -Max 1.0
cB 1-
cB 6.0 1- 6.3 cc - - Bolt Force Range cc
6.0 1 7
5.8
1 5.8 ~ 6.2 0.180 g cc 0.9 g
0
5.6
0 5.6 g 6.1
0 .175 ;
Q)
01
0
::::. c
Q)
C)
5.4 0.8 111
~
~ 6.0 0:
Q)
~ 6
0 200 400 600 800 1000 0 200 400 600 800 1000 0 0
0:
0 .170 ~
Time (sec) Time (sec) 't
0
5.9
0 ...
u.
0
0.7
Q)
0
~
0
co 5.8 u. m u.
(h) SPT2A (60,000cycles) 0.165 ~ 0.6 ~
(g) SPT2A (!cycle) - Min
5.7 -Max
0.160
s 5 0
m
5.6 - - Bolt Force Range 0.5

5 .5 0.155 4 0.4
4 4 0 500000 1000000 1500000 2000000 0 20000 40000 60000 80000
7.4 7.4 - aon1
7.2
~ BoH2 Cyde Cycle
7.2 - Tensileload
- Tensile Load 3 3
7.0 c- 7.0 c-
c- c c- c (a) SPTIA (b) SPTlC
g 6.8 .2 g 6.8 0
c:. 2:;' 2;-
8 66 ~
~ 6.6 "'0
!:!
.f 6.4 ..J
af 6.4
..J

s 6.2
6.0
1~
Ill

c
Ill
1-
s 6.2
6.0
1~
Gl

cGl
1-

5.8
0
5.8 ~ 0
9 1.50 6.5 0.130
5.6 5.6
1.48 6.4
0 200 400 600 800 1000 0 200 400 600 800 1000 0.125
8 1.46 ~ 6.3 c
Time (sec) Time (sec)
0.120 c
cc 1.44 g cc 6.2 g
(i) SPT2C (!cycle) (j) SPT2C (l,OOO,OOOcycles) g7 Q)

Q)
1.42 c C)
g 6.1 0.115 Q)
CJ)
c
0
111 Q)
111
1.40 0: ~ 6.0 0.110 0:
0
't
0
6 1.38
Q)
~
0
a:0 5.9
0

0.105
Q)
~

7.4
4 m
- Min
1.36
0
a: m 5.8
-........_.. e Ieee._.

0.100 ~
0
u.
5 1.34 m Min
7.2 5.7 m
- TensHe Load - Max -Max
3 0.095
7.0 c- ~ Bolt Force Range 1.32 5.6 - - Bolt Force Range
c- c
!5 6.8 g 4 1.30 5.5 0.090
8 6.6
2-o
13
0 10000 20000 30000 40000 50000 60000 0 500000 1000000 1500000 2000000
.2 6.4
..J
Cycle Cycle
s.
cB 6.2 1~
c
Ill
6.0 1- (c) SPT2A (d) SPT2C
5.8 ~ 0
5.6
0 200 400 600 800 1000
Time (sec)

(k) SPT2C (2,000,000cycles) Fig. 4.27 Change of the Maximum and Minimum Bolt Force

Fig. 4.26 Time History of Bolt Force and Applied Tensile Load under One Cycle

132
133
 1200 1000 100
70
~ RB-1 - LB-1 _.._ LB-1
1000 600 60 - - RB-1
- - RB-2 - - LB-2 - LB-2
~ RB-3 - LB-3 - - RB-2 80
600 600 _.__ LB-3
- - RB-4 - - LB-4
50 - RB-3
-- LB-4
600 40 - - RB-4 60
400
3 400
3 330 3
.5 c 200 c
c
~
(/)
200 ~
(/) ~ 20 ~
40
0 en en
0 10 20
-200
-200 0
...4()0 0
...4()0
-10
-600 -600 -20
-20
0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000
0 200 400 600 800 1000
Time{sec) Time (sec) Time (sec) Time (sec)

(a) SPTlB-R (I cycle) (b) SPTlB-L (lcycle) (g) SPT2C-R (I cycle) (h) SPT2C-L (I cycle)

400 400 120

80
- RB-1 - LB-1
-RB-1 - LB-1
300 -- RB-2 300 - - LB-2 70 100
- LB-3 - - RB-2 -- LB-2
- RB-3 - RB-3
_._ RB-4 - - LB-4 60 - LB-3
200 200 - - RB-4 80 -- LB-4
50
3 3 340 3 60
c 100 . 100 c
c
~ ca
._ 'jij
en
0
(i)
0
~ 30
en
20
=
en 40

20
-100 -100 10
0
0
-200 -200
-10 -20
0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 0 200 400 600 800 1000
Time (sec) lime (sec) Time (sec) Time (sec)
(c) SPTIC-R (lcycle) (d) SPTI C-L (I cycle) (i) SPT2C-R (l,OOO,OOOcycles) G) SPT2C-L ( l ,OOO,OOOcycles)

1000 1000 100 120

-- LB-1
800 - LB-1 - LB-1
-- LB-2 800 -- RB-1
-- LB-2 100
- LB-3 80 - RB-2 LB-2
_._ LB-3
600 -- LB-4 - RB-3 ~ LB-3
600 -- LB-4 80
60 -- RB-4 - - LB-4
3 400 3 400
3c 3 60
c c c
~ 200 ~ 'iii 40 ;;;
en
0
(i) 200
=
en
20
=
en 40
0 20
-200 -200 0 0
-400 -400
-20 -20
0 200 400600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000
Time (sec) Time (sec) Time (sec) Time (sec)
(e) SPTlC-R (80,000cycles) (1) SPT IC-L (80,000cycles) (k) SPT2C-R (2,000,000cycles) (I) SPT2C-L (2,000,000cycles)

Fig. 4.28 Time History of the Strain at the Bolt Shank (continued) Fig. 4.28 Time History of the Strain at the Bolt Shank

131 135
 1000 1000

1500 800 800

1500
600 600

~ ~
1000 1000 400 400
:; 200 :; 200
500 500 c c
:; :; ii)e 0 n; 0
c c 0 -200 ~ -200
; 0 "iij
!::>
(J)
-500
!::>
(J)
-500
-400
-600
-800
~-- Maximum Strain
-<>- Minimum Strain
I -400
-600
-800
~ -- Maximum Strainl
-<>- Minimum Strain

-1000 -1000 - - Maximum Strain -1000 -1000

-<>- Minimum Strain
0 10 20 30 40 50 0 10 20 30 40 50
-1500 -1500
Channel Number Channel Number
0 10 20 30 40 50 0 10 20 30 40 50
Channel Number Channel Number (g) SPT2A (I cycle) (h) SPT2A (60,000cycles)
(a) SPTIA ( I cycle) (b) SPTIA (l,OOO,OOOcycles)

1000 1000
800 800
1500 1500
600 600
1000 1000 400 400

:; 200 :; 200
500 500 _; _; 0
0
:; ~ "'
!::>
(J) -200
"'
!::>
(J) -200
c
ii)ec 0 "iii
._
0

-500
ii)
-500
-400
-600 ~-- Maximum Strain
-<>- Minimum Strain
I -400
-600 - - Maximum Strain
-800 -<>- Minimum Strain
-800
-1 000 - - Maximum Strain -1000 -1000
-1000
-<>- Minimum Strain 0 10 20 30 40 50
0 10 20 30 40 50
-1500 -1500 Channel Number
Channel Number
0 10 20 30 40 50 0 10 20 30 40 50
Channel Number Channel Number (i) SPT2C (I cycle) (j) SPT2C (I ,OOO,OOOcycles)
(c) SPTIA (2,000,000) (d) SPTIB ( l cycle)

1000
800
1500 1500
600
- - Maximum Strain
1000 1000 -<>- Minimum Strain 400
:; 200
500 500 c 0
:; :; ~
(J) -200
c
~
(J)
0
eii)c 0
-400
.50() -600 - - Maximum Strain
-500
-<>- Minimum Strain
-800
-1 000 -1000 -1000
0 10 20 30 40 50
-1500 -1500 Channel Number
0 10 20 30 40 50 0 10 20 30 40 50
Channel Number Channel Number (k) SPT2C (2,000,000cycles)
(e) SPTIC (lcyclc) (f) SPTI C (80,000cyclcs)

Fig. 4.29 Strain Distribution on the Flange Plate

Fig. 4.29 Strain Distribution on the Flange Plate (continued)

137
136
 ........ 1.1
10 "0
ra
.3 1.0
fth=1 O(mm) fth :
9 fth=22(mm)
0) Thickness of
8
.E 0.9
"0 the Flnage Plate
Ill
Ill
Q.l 7 ~ 0.8
~
C/)
"C 6 ~ 0.7
Q.l _J
.~
'iii
E
5 ;gra 0.6 Not Failed
....
0 4 a::<1,) 0.5 ..__
z 0)
3 - SPT1A ; 0.4

--
- SPT1B a::
2 Not Failed
_..,_ SPT1C -g0 0.3
_J

Oe+O 2e+4 4e+4 6e+4 8e+4 1 e+5

Cycle
103 105 106
Cycle

(a) SPTI Type

Fig. 4.31 S-N Diagram (1)

3 -r---- - - - - -- - - ---,
- SPT2A 8
_._ SPT2C
F.S.:
7 F.S. for Bolt (fth=22mm)
Fatigue Strength
6 F.S. for Plate (fth-22mm)
c
-g
..........
<1,)

g>4
5
. --
F.S. for Bolt (fth=1<!._~m_2_ ___ _N_o1Eailed
----------
ra
'-
"0

-
ra 3 fth=10mm Not Failed
0
_J
fth=22mm
2
o l-~~~==~==~==~ F S for Plate (fth=1 Omm)
~ ~ ~---- ------- ----------
Oe+O 2e+4 4e+4 6e+4 8e+4 1 e+5 1
Cycle 0
1Q3 105
(b) SPT2 Type Cycle

Fig. 4.32 S-N Diagram (2)

Fig. 4.30 Time History of the Stress on the Flange Plate

138 139
 Plate Chapter 5

Simple Analysis on Split Tee Flange Joints using

2-dimensional Finite Element Method

5.1 Introduction

Fig. 4.33 The Estimated Model for the Fatigue Strength of the Flange Plate
Generally speaking, in order to investigate the behavior of structures and assess its reliability against
external loads, 3-dimensional analysis may be required . Particularly, for complex structural clement such
as the bolted joints, loading transferring mechanism and stress distribution should be examined by stress
analysis. However, such an analysis requires much effort in pre-/post data preparation as well as
computation. Even though computation teclmology such as finite element method has been developed
significantly, 3-djmensional stress analysis may face with many difficulties such as computation time and
memory. In addition, it is almost impossible for engineers to carry out 3-dimensional analysis to design
each structural clement of structures. It is better to provide the rational design procedure with parameter
limitations based on a data-base or simple analysis method to check the actual 3-dimensional behavior. In
particular for the design of joints, the structures have many joints and they are not always same; namely
every joint may have its 0\\11 design condition. Therefore, such a simplification can save a lot of time for
tedious joint design. In thls chapter, simple analysis procedure by 2-dimensional fmite clement method for
split tee flange joints is discussed.

~} In order to carry out 2-dimensional analysis for 3-dimensional structures, the assumption such

~~~--~--~~~ ~
Q_ as plane stress or plane strain should be made. As for the split tee flange joints which consists of flange
plates and high strength bolts, the above assumption can not be made directly. Accordingly, any
Modeling
modification or proper structural model should be made in order to make it possible to reproduce
mecharucal behavior and stress state by quasi 2-dimensional analysis. In this study, the concept of
I effective width for flange plate and bolts is proposed. Based on the data base of complete infonnation on
mechanical behavior of the split tee flange joints obtained by 3-dimcnsional finite element analysis,

(a) Split Tee Flange Joint effective width coefficients in 2-dimensional finite clement analysis with asswnption of plane strain are
(b) Clumped Beam at both Ends
discussed. Furthcnnore, the state of local stress at tl1e flange plate and bolts are assessed for calibrated
effective width coefficients.

Fig.4.34 Model of Working Stress Verification on the Flange Plate

141
140
5.2 Quasi-2 dimensional Analysis on the Split Tee Flange Joints
5.2.1 General procedure for calibration of effective width coefficients
At first, 3-dimensional analysis is carried out in order to make a database of complete
information on mechanical behavior of split tee flange joints varying the thickness and the width of the
flange plate. Anal)1ical model is shown in Fig. 5.1 and this model is the same as the model considered in
the previous Chapter 4. The material properties, boundary conditions and the procedure of the analysis are
all the same and can be referred to 4.2.2. In addition, the finite element discretization and its procedure
also can be referred. All of the analytical cases are tabulated in Fig. 5.2. As shown in this figure, both the
thickness of the flange plate and the width of the flange plate are varied. As for the selection of these
an'atytical cases, attention is paid to the applicable range in the current design procedure for building
structures[ I) and the design procedure for bridge structures(draft)[2], in wruch structural dimensions such
as the thickness of the flange plate is limited to prevent prying force. In this figure, such limits of the
thickness and the width of the flange plate are also shown.
Secondly, 2-dimensional analysis is carried out with a certain cffect1ve coefficient of the flange
plate and that of the bolt. TI1e ciTective width coefficients are defined as follows:
w1 = k 1 w10
(5. 1)
wb = kb. wbo
in which, ll[ is the cfTective width of the flange plate, kf, the eiTective width coefficient, WJO, the width of
the flange plate, l1'b, the effective width of the bolt, kb, the eiTective width coefficient of the bolt, wbO, the
width of the bolt. The effective width coefficients used in the 2-dimensional analysis are determined based
on the load-separation relationship for all the cases in order to reproduce the 3-dimensional mechanical
behavior of the joint. Til is 2-dimensional analysis is repeated-by varying coefficients until good ~greement
of results of 2-dirucnsional analysis with t110se of 3-dimensional analysis is obtained.
5.2.2 Analytical model
In this analysis, the analytical model of the split tee flange joints is considered to be plane strain problem.
As for the clement, constant strain triangle element is used considering simplicity of fonnulation. TI1e
analytical model and its finite clement discretization are shown in Fig. 5.3. Boundary conditions, such as
the condition at the contact surface between two flange plates, the condition at the contact surface
between the bolt and the flange plate, material properties arc the same as those used in 3-dimcnsional
analysis. Titcse boundary conditions and material properties arc also sho\\n in Fig. 5.4 and Table 5.1
rcspcctiYcly. TI1crcforc, the difTcrencc between 2-dimcnsional analysis and 3-dimcnsional analysis is
basically only a dimension. This 2-dimcnsional model is based on a plane cut of the 3-dimcnsional model
at the center of bolt line. And finite clement discrctit.ation is the same as that of 3-dimcnsional analysis on
142
the out surface. In addition, in this 2-dimensional model, the nodal points of the bolt section, and those of
the fl ange plate section are conunon in order to reproduce the 3-dimensional structure by unifom1 cross
section in width direction. In this analysis, maximum thickness of t11e flange plate is 22 (nun) and
maximum width of the flange plate is 94.5 (mm), so that the number of the elements and degrees of
freedom for the largest analytical case is 576 and 656 respectively. Comparing with 3-dimcnsional
analytical cases, the number of the clements and degrees of freedom is 1/32, 1/21 of 3-dimcnsional
analytical case. Therefore, it is understood tllat 2-dimensional model is very efficient as compared to 3-
dimensional model from the view point of tlle dimension of the array prepared for the analysis on t11e
micro computer.
5.3 Results and Discussions
5.3.1 Effective width on mechanical behavior
The results of the load-separation curves for several effective widtll coefficient of the flange plate arc
shown in Fig. 5.5. The horizontal axis shows t11e separation between two flange plates at the tee web plate
and the vertical axis shows the tensile load applied to t11e split tee flange joint. In case of the thinner
flange plate, it is found from these figures that the behavior of contact/separation is considerably affected
by the effective width coefficient of the flange plate and it is understood t11at the larger the coefficient is,
the rugher tlle strength is. On the other hand, in case of tlle thicker flange plate, tlle behavior of
contact/separation is not affected by tlle effective "~dt11 coefficient of tlle flange plate. Thls difference is
considered to be caused by the difference of the failure mode. Namely, in case of the thinner flange plate,
the behavior significantly depends on the flange plate, so that t11e coefficient of tlle flange plate is very
important. On the other hand, in case of tlle thicker flange plate, the behavior significantly depends on the
bolt not the flange plate, so that the effective width coefficient of the flange plate is not important for t11e
mechanical behavior.
Next, load-separation curves for several effective width coefficients of the bolt obtained from
2-dimensional analysis are shO\m in Fig. 5.6. The horizontal axis shows the separation between two
flange plates at the tee web plate and the vertical axis shows the applied tensile load to the joint in the
same way of Fig. 5.5. In case of the thinner flange plate, it is found from theses figures that the shape of
the curve is almost the same in spite of varying the effective width coefficient of the bolt. Particularly, this
tendency is significant for the joint which has narrow width flange plate. On the other hand, in case of the
t11icker flange plate, it is understood that the shape of the curve is significantly arected by the effective
width coefficient of the bolt, that is, the larger this coefficient is, tJ1e higher the strength is. Such a
difference of the cfTect of effective width coefficient on load-separation curves for the different thickness
oftlle flange plate is considered to be caused by the difTcrent failure mode as mentioned above. That is, in
143
case of the thinner flange plate, the mechanical behavior of the joint depends on the behavior of the flange
plate, as a result, it is not affected by the effective width coefficient of the bolt, on the other hand, in case
of the thicker flange plate, the behavior of the joint depends on the that of the bolt; accordingly, it is
affected by the effective width coefficient of the bolt.
5.3.2 Calibration of effective width coefficients for the flange plate and the bolt
Based on all the results obtained from 2-dimensional analysis varying the effective coefficient of both the
flange plate and the bolt, these coefficients is calibrated by try and error method. When load-separation
curve obtained from 2-dimensional analysis is good agreement with that obtained from 3-dinlensional
analysis, these coefficients are detcmtined to be effective coefficients. As shown in Fig. 5.5 and Fig. 5.6,
if 'the effective coefficient of the flange plate is 0.90 and the effective coefficient of the bolt is 0.65, it is
found that the load-separation curve is well predicted by 2-dimensional analysis.
5.3.3 Deformation and stress verification by 2-dimensional analysis using effective coefficients
Defonnations of all the cases using calibrated effective width coefficients, namely 0.90 for the flange
plate and 0.65 for the bolt, arc shown in Fig. 5.7. ln these figures, in order to show the defonnation
clearly, the displacement is magnified by 20 times of the actual displacement, in which the results of 3-
dimcnsional analysis arc also shown for comparison. In case of the thinner flange plate, it is found from
these figures that the bolt is not defonned significantly. As for the flange plate, significant defonnation
occurs at the bolt and it isn t defonned at the outer edge of flange plate and remained original shape. On
the otJ1cr hand, in case of thicker flange plate, it can be seen from these figures that the flange plate is not
defonned significantly. In addition, it is observed that flange plate is rotated at the edge of the flange plate
as a rigid body and that tJ1e significant defom1ation of the bolt occurs at the bolt thread. Such results of
the mechanical behavior of the joints obtained from this 2-dimensional analysis is good agreeinent with
experimental results and 3-dimcnsional exact analysis. Therefore, it is concluded that this quasi 2-
dimensional analysis using effective coefficients of both the flange plate and the bolt can reproduce the
defonnation mode correctly. However, it is observed that the displacement obtained from 2-dimcnsional
analysis is a little different fom1 that obtained from 3-dimensional analysis. Especially, the difference is
observed at the section between the tee web and the bolt. In order to carry out more accurate analysis, this
analysis method should be modified.
Sccondal), the maximum stress at the flange plate obtained from both 2-dimensional analysis
using tJte calibrated effective \\idth coefficients( 0.90 for flange plate and 0.65 for bolt) is tabulated in
Table 5.2. In this table, the results obtained from 3-dimcnsional analysis arc listed for comparison. In iliis
table, the location of the maximwn stress are also shown. From this Table in case of the thinner flange
plate, it is understood that the maximum stress and its location are well predicted by 2-dimcnsional
analySIS with cfTccti\c "idth coefficients. On the other hand, in case of the thicker flange plate, it is found
144
that the maximun1 stress obtained by 2-dimensional analysis is quite different from that obtained by 3-
dimensional analysis. Particularly, it is understood that the larger the width of the flange plate is, the
larger the difference between 2-dimcnsional analysis and 3-dimensional analysis arc made. However, the
location where tJ1e maximwn stress occurs is in good agreement with the results of 3-dimensional analysis.
Accordingly, it is concluded that it is difficult to utilize tJ1is analysis directly for stress verification. This
difference between 2-dimcnsional and 3-dimensional analysis is considered to be caused by the fact tl1at
the behavior of the joint significantly depends on tJ1at of the bolt in case of tJ1e thicker flange plate.
Finally, bending stress vs. load curYes for all tJ1e cases obtained from bo!Jl 2-dimensional
analysis with calibrated effective \\~dth coefficients( 0.90 for flange plate and 0.65 for tJ1e bolt)and 3-
dimensional analysis are shown in Fig. 5.8. The horizontal axis shows the tensile load applied to the split
tee flange joint and the vertical axis shows the bending stress at the fixed edge of the bolt. The bending
stress is calculated based on the bending moment applied to the edge of the bolt wltich is obtained from
ilie nodal force at the end of the bolt and elastic section modulus of the bolt. It is found from these figures
that !Jle bending stress in case of the thinner flange plate is about 2.0 times higher than that in case of tJ1c
thicker flange plate. Moreover it is observed that there is significant difference between the result of 2-
dimensional analysis and that of 3-dimensional analysis for any thickness of tJ1e flange plate, and it is
found that the result obtained from 2-dimensional analysis is higher !Jlan that obtained from 3-
dimensional analysis. Therefore, based on these results, it should be concluded that there exist many
difficulties to estimate the stress occurred at the bolt, as if the behavior of joint depends on that of the
bolt.
Considering the comparisons mentioned above, it is understood that the prediction of stress
state by 2-dimensional analysis using the effective width coefficients only is very difficult. In order to
reproduce stress state by 2-dimensional analysis, further study should be requtred.
5.4 Conclusions
In this chapter, quasi-2-dimensional analysis on split tee flange joints for the load-separation behavior is
proposed. In this analysis, the effective width coefficients of the bolt and the llange plate is utilized in
order to reproduce 3-dimcnsional behavior. lltis analysis is very useful for parametric study of the split
tee flange joint or tJte design of tltis type of joints because of its simplicity compared with 3-dimensional
analysis. TI1e following conclusions and future research needs are obtained as follows:
I) 3-dimcnsional behavior of the split tee flange joint, such as load-dcfonnation cunc,
dcfonnation characteristics, can be reproduced by quasi-2-dimensional finite element analysis
using the effective width of the bolt and the flange plate. In this study, the coefficient of the
145
 flange plate and that of the bolt is determined to be 0.90 and 0.65 respectively by try and error. Ta ble 5 I Ma ten.a I Propert1cs us ed.m 2-d"JmCnSJOna I An aJ)'SI
I s
Young' s Modulus Yielding Stress
2) The stress obtained by this 2-dimcnsional analysis is very conservative. Furthem1ore, accurate
Flange Plate 21,000 26.9
stress verification is difficult by this 2-dimensional finite clement analysis using effective width
Bolt Shank
coefficients. Particularly, in case of thicker flange plate, the difference between 2-dimensional
of 21 ,000 90.0
analysis and 3-dimcnsional analysis tends to be large. High Strength Bolt
Bolt Thread
of 15,700 67.1
In the future, this 2-dimcnsional analysis should be applied for the cases with various geometrical Hiclt Stren~th Bolt
configurations, such as various thickness of the flange plate, width of the flange plate, the number of the
bolts, and its applicability should be made clear for the application of this analysis for actual joints.
Furthermore, in order to estimate the stress state at the flange plate more accurately, this 2-dimensional
analysis should be modified. I .
Table 5.2 Comparison of Maximum Stress between 2-dimensional an d 3-dtmensJOna AnalySIS
Analytical 2-dimensional Analysis 3-dimensional Analysis
Case Maximwn Stress Location Maximwn Stress Location
References
STF2 27.4 A 27.7 A
I) Architectural Institute of Japan : Recommendation for the Design Fabrication of High Strength Bolted
STF4 27.2 A 28.7 A
Joints, Maruzen, Mar. 1993(in Japanese).
STF13 21.3 A 27.2 A
2) Japanese Society of Steel Construction : Rcconunendation for the design of tensile joints for bridge
STFI5 14.6 A 27.2 A
structures (draft), Feb. 1993(in Japanese).
[NOTE]
A: Adjacent of the High Strength Bolt at the Loading Side
STF2,STF4 : Thinner Flange Plate
STF13, STF15 :Thicker Flange Plate

146
147
 22 em
52. em
/ /:
~ ~
/. I/:
29.tm l s.em
Side View
/:

~ ~
/.

-
10. 0ntn
100

90
sy~
STF4 151 F'8
Slj11

STF15
11. em
Contact Surface
12. <mri ~ Contact Surface I
I
High Strength Bolt(M12,FJOT) I
80

- E
E
.........
70
STF3 II
I

I
STF10 STF14
STF2 I S~ F'7 STF13
31 . 5nm
-~
J::
"0 60 STF9
S~FS
Grand View
STF1
I
I

STF12
50 I
I
I ~- Ref.10
.---- Ref. 15 .
I
40
l
103.em l
30 -
Fig. 5. l 3-dimcnsuional Analytical Model
8 10 12 14 16 18 20 22 24
Thickness( mm)

L.1St 0 fAl l An ai)1Ica

I . I Cases ( urut
. : mm)

Case STFl STF2 STF3 STF4 STF5 STF6 STF7 STF8

TI1ickness of 10 10 10 10 12 12 13 13
the Falnge Plate
Width of 56.7 63 78.75 94.5 63 94.5 63 94.5
tl1e Flange Plate
Case STF9 STFIO STFl l STF12 STF13 STF14 STF15
Thickness of 15 15 15 22 22 22 22
Ole Flange Plate
Widtll of 63 78.75 94.5 56.7 63 78.75 94.5
the F alnge Plate

Fig. 5.2 List of Anal)1ical Cases

148 149
 6 10

5
8
c cc:
4 g
~
22.0mrn "0 "0
6
(11 (11
Loading Area 8.0mm 0 3 0
52.0mrn ..J ..J
~ --30 ~ 30
u; u; 4
c:
Q) 2 - 20 (W8 =0.65, W1 =0.9) c:
Q)
20 (W8 =0 65, W1 =0.90)
I-
_...._ 20 (W8 =0.65, W1 = 0.85) I- __.__ 20 (W8 =0.65, W1 =0.95)
I O.Omm -<:~- 20 (W8 =0.65, W1 = 0.95) 2 --.,- 20 (W8 =0.65, W1 = 0.85)

0 0
ll.Omrn 0 2 3 4 5 6 7 8 0 2 4 6 8 10
Separation (mm) Separation (mm)
High Strength Bolt (a) STF l (b) STF2
(M 12 FlOT)
10 10

Fig. 5.3 2-dimcnsional Anlytical Model

8 8
cc: cc:
0
~ g
"0
6 "0
6
(11 (11
0 0
..J ..J
~ ~ - -30
u; 4 u; 4
c:
Q) --30 c:
Q) 20 0N8 =0.65, W1 =0.90)
I- 20 (W8 =0.65, W1=0.90) I-
__.__ 20 0f'J8 =0.65, w, =0.95)
2 2 --.,- 20 (W8 =0.65, W1 =0.85)

0 0
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Unseparated Condition Separation (mm) Separation (mm)
Loading Area
(c) STF3 (d) STF4

16 16

14 14

Contact Surface Contact Surface -g

_..... 12
c:
10
cc:
g
12

10
,. --=-"~ -
Axis of S)1nmetry "0
(11
"0
(11
0 8 0 8
..J ..J
Fig. 5.4 Boundary Conditions of 2-dimcnsional Analytical Model ~ ~

I: ::_;g
IJ)
6 u; 6
c: c:
Q)
I-
Q)
I-
(W, = 065. w, = 090)]
4 - 20 (W8 = 0.65, W1=0.90) 4
__.__ 20 (W8 =0.65, W1 =0.95)
2 2
--.,- 20 (W8 =0.65, W1 =0.85)
0 0
0 2 3 4 5 6 7 8 9 10 0 2 4 6 8 10
Separation (mm) Separation (mm)
(c) STF5 (Q STF6

Fig. 5.5 Load-Separation Curves changing the Coefficient of the Flange Plate (continued)

151
150
 16 16 20 ~----------------------, 20
18 18
14 14
16 16
12 c 12 ~ 14
c
c c -E 14
g 0
g 10
"C
10 ;- 12 ~
"C
12
"C
ro ro ro
ro
0
_J 8 0
_J 8 .3 10 .3 10
~ ~
~ 30 ~ v; 8 v; 8 30
VI
c 6 20 0Ne = 0.65, W1 = 0.90) VI
c 6 c -30 c 20 (We = 0.65, W1 = 0.90)
Cl Cl - -30 ~ 6 - 20 0Ne =0.65, W =0.90) Cl
1- 6
1- _.__ 20 (We= 0 65, W1 = 0.85) 1- 1 _.__ 20 (We= 0.65, W1 = 0.95)
4 4 - 20 (We = 0.65, W 1 = 0.90) 4
____..,__ 20 (We =0.65, W1 =0.95) 4 -v- 20 (We = 0.65, W1 = 0.85)
2 2 2 2

0
o._- ----,.---..---.---..- -_, 0
0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8
0 1 2 3 4 5 0.0 0.5 1.0 1.5 2.0 2.5 3.0
10
Separation (mm) Separation (mm)
Separation (mm) Separation (mm)
(g)STF7 (h)STF8 (m) STF14 (n) STF15

20 T---------------------~ 16
18
14
16 Fig. 5.5 Load-Separation Curves changing the Coefficient of the Flange Plate
~ 14
~
0
12
-g
........ 12
c
10
"C "C
ro ro
0 10 0
_J _J 8
~ ~
'iii 8 v; 30
c 6

~g
Cl
c
6 Cl - 20 (We= 0.65, W 1 = 0.90)
w, 0.90) I
1- 1-
(W 0 65, 4
4
2 2

0 0
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Separation (mm) Separation (mm)
(i) STF9 G) STFI I
18 20 .---------~--~

16
14
C' cc 15
c 12
g g
"C
ro 10 .,
"0

0
_J
..3 10 30
~ 8 -l 30 .S! - 20 (W8 = 0.65, W 1 = 0.90)
'iii
'iii 20 (We= 0.65, W1 = 0.90) cG)
c 6 -&- 20 (W8 = 0.65, W 1 = 0.95)
Cl
1- - - 20 (We= 0.65, W, = 0.85) 1- 5 -v- 20 (W8 = 0.65, W 1 = 0.85)
4 ~ 20 (We= 0.65, W 1 0.95) =
2
0~--.-----.---.--,---,-----j
0
0 2 3 4 5 6 7
0 2 4 6 8 10 Separation (mm)
Separation (mm) (1) STF13
(k)STF12

Fig. 5.5 Load-Separation Curn:s changing Ole Coefficient of Ole Flange Plate (continued)

152 153
 8.----------------------- 10 . -- -- - - - - - - - - - - -- -- - ,

8
c-6 C"
c: c:
g .s
"0 ::; 6 _60 ~--------~~~==~ - 60 ~-----;::=--======::;J
cu cu
~ I~
E 30 Analysis 30Analysis
.3 4 0
...J ~ 40 o 20 Analysis 20Analysis
~ ~ 4 c: ~
c:
40
v; v; 30 C1l C1l
c: 30
~
c:
C1l
1- 2 20 (We= 0.65, W, = 0.90) C1l
1-
20 (We= 0.65, W1 = 0.90)
(.)
20 0 0
~G 0
0
o<?9 ~
u
20
_._ 20 (We= 0.80, W1 = 0.90) _._ 20 (We= 0.50, W1 = 0.90) co "'?ctg90 cu
2 Ci. Q Q
0..
_,_. 20 (We= 0.50, W1 = 0.90) _,_ 20 (We= 0.80, W1 = 0.90) 6 0 6 0
o v----..-- ---,,-----r---.-----~
0 20 40 60 80 100 120 0 20 40 60 80 100 120
0 2 4 6 8 10 0 2 4 6 8 10 12
Separation (mm) Separation (mm) (a)STF2 (b)STF13

(a) STF2 (b) STF4

Fig. 5.7 Deformation ofthe Bolt and the Flange Plate(Yield Loading Stage)

20 .-----------------------. 20 .------------~

50 ~------r===========~
C" 15 C" 15 - STF 2(2d-fem)
c: c: - STF13(2d-fem)
0
~ g --<>- STF 2(3d-fem)
"0 "0 - o - STF13(2d-fem)
cu cu
.3 10 .3 10
~ C1l
v; 30 v;- 30 -
c:
C1l 20 (We = 0.65, W1 = 0.90) c:
C1l 20 (We =0.65, W1 =0.90)
1- 5 _...,._ 20 (W8 = 0.80, W1 = 0.95) 5
1- - A- 20 (W8 =0.50, W1 =0.90)
_,_. 20 (W8 =0.50, W1 =0.90) _,_ 20 (W8 =0.80, W1 =0.90)

0~-----r--.--.--.-~
0 2 3 4 5 6 7 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Separation (mm) Separation (mm) O fed~l::!jl:~-..--,--.---.---,--l
0 2 4 6 8 10 12 14 16 18
Load(tonf)
(c) STF13 (d) STF15
Fig. 5.8 Bending Stress(Bolt)-Load Curves

Fig. 5.6 Load-Separation Curves changing the Coefficient of the Bolt

154 155
 effective for preventing prying action of the flange plate at tension side of the joints. TI1crcfore, they
Chapter 6
concluded that the practical use of the high strengtl1 bolted tube flange joints with thicker flange plate
without any rib plates can be feasible for members under bending.
Application of the High Strength Bolted Tensile Flange Joint
In this study, the strength and the design of the tube flange joints with thicker flange plate is
-High Strength Bolted Tube Flange Joint-
further studied based on not only the previous experimental works by Watanabe, E. et at. but also
experimental and anal}1ical studies for high strength bolted tensile joints described in previous chapters.
In order to investigate the mechanical behavior of Ll1e tube flange joint subjected to general loading such
6. 1 Introduction
as combined loading of bending and tension, a series of experiments is carried out. In addition, based on
the experimental observations such as the co-relation of global behavior of the tube flange joints to the
In previous chapters, fundamental structural elements of high strength bolted tensile flange joints such as
local behavior of bolt groups discussed in tile previous chapters, the rational design procedure for the tube
high strength bolts, BAF model and split tee flange joints are investigated. ln Litis chapter, the high
flange joints subjected to the combined loading is proposed and its applicability is assessed.
st;engtl1 bolted tube flange joint is focused as an application of the high strength bolted tensile joint. This
type of joints is mainly used in the offshore structures, the steel erosion control dam, transmission tower
and other temporal facilities. Typical application for the steel erosion control dam and its joint detail are
6.2 The Mechanical Behavior of the Tube Flange Joints
sho\\11 in Photo 6.1. Tilis type of joints is classified into 2 types. One type is the joint with rib plates, the
other type is the joint without rib plates. Typical examples of these two types of tube joints are
In order to investigate L11e mechanical behavior on the tube flange joints \\~th thicker flange plate
schematically sho\\11 in Fig. 6.1. As shown in Fig. 6.1, the structural details of Ll1e latter is very simple
subjected to combined loading, two series of experiments are carried out. The outline of the experiments
because of no rib plates welded, so that its fabrication can be done easily by the automated machine such
is described in the following.
as cutting and welding.
TI1e pipe structures arc often applied for the offshore structures and the pipe lines for water/gas
6.2. 1 Outline of experiments
supply; and particularly utilized for the structures in water environment because of no dependence of
(a)Pure bending test
loading directions on bending strength. In addition to this structural superiority, the cross section shape of
In order to investigate the mechanical behavior subjected to pure bending, 4 points bending test on Ll1e
pipe can give better impression to observers than that with sharp comers and it is desired from the view
specimen connected with heavy non-destructive box beam by high strength bolts is carried out using Ll1e
point of aesthetic design. However, in order to construct the pipe structures, the problems such as how to
servo-controlled hydraulic actuators ''ith capacity of 400 (kN). This test setup is shown in Fig. 6.2. Tile
connect members have arisen. Generally speaking, joints ar~ made of either bolt connection o~ welding.
loading is controlled manually to set a certain vertical displacement at the loading point. TI1e unifonn
Since welding is superior in water problems, in Ll1e past, Ll1e complex welding joints have been studied
bending applied to the specimen given by this test setup is shown in Fig. 6.3. Basically, t11e loading is
cxtensi\ely and implemented for the large offshore structures by taking into consideration for the fatigue
continued until the toad reaches to t11e peak. In the experiment, the displacement of tile actuator, the
failure. At present time, as a result, the connecting methods of pipe members such as welding, gusset plate
applied load, the average cunature of t11e specimen and local defonnation of tile joint elements such as
joints and direct splicing by high strength bolts have been established[ l ][2]. However, there is a little
the fl ange plate and the tube, are measured by a micro computer through GP-IB. Each measurement item
study available on high strength bolted tube flange joints, and in particular no study on the tensile joints
is sunuuarized as follows:
subjected to bending can be found because axial tension/compression is only taken into consideration as a
( I )The applied load and the displacement
result of design assumption Ll1at the structures are truss structures. Recently, Watanabe, E. and co-
n 1cse arc measured by load cell and displacement transducer built in the actuator. The bending
workers carried out the experiment on high strength bolted tube flange joints both with rib plates and
moment M can be given by
without rib plates subjected to bcnding[3 ]. It is found that the joint rigidity is as high as that of the tube
and not necessary to be asswned pin structures It is also found that both the joints stiiTcned by the rib I
M= PL (6.1)
plates and the joints not stiffened by the rib plates but "ith thick fl ange plate can perfonn very well under 2
bending and their mechanical behaviors arc very similar. It is shown that both structural details arc

157
156
 in which P and L are applied load and the distance between the loading point and the supporting
point respectively.
(2)The average curvature of the specimen
The average curvature of the specimen is measured by the displacement transducers installed on
the specimen at the compressive/tensile sides. It is calculated by the following equation.
(6.2)
in which, , u~, 112, d and L are the average curvature, the displacement at the compressive side,
the displacement at the tensile side, the distance between two displacement transducers and
length of the specimen respectively. Schematic view of the setup of those displacement
transducers is shomt in Fig. 6.4.
(3)Local defonnation
llte local defonnation of the joint elements, such as the defonnation of the flange plate, the
defonnation of tlte tube and the buckling shape of tlte tube arc measured by the strain gages.
The local defomtation of the tube is measured by uni-axial strain gages glued on tlte inner/outer
surface of the tube, and the local defonnation of the fl ange plate is measured by rectangular
rosette strain gages glued on the surface of the flange plate. llte locations of the measuring
points are shown in Fig. 6.5.
(b)Combined loading test
In order to investigate the mechanical behavior subjected to combined loading of tension and bending, the
test setup shown in Fig. 6.6 is utilized. As shoNn in Fig. 6.6, two servo-controlled hydraulic actuators
\\itll the capacity of 400 (kN) are used in this test. These t\vo actuators are connected to the tipper and
lower beams by universal joints. llte loading control system is shown in Photo 6.2. The test is controlled
by tlle load of the actuators and the operation of loading is made through the micro-computer. The loading
procedure is sununarized as follows: at STEP I, the load of each actuator set to zero after finishing the
installation of the specimen for starting the loading(bccause the compressive load about I (kN) is already
applied to tltc specimen in order to install the specimen easily); at STEP 2, the tensile load is applied to
the specimen by increasing the loads of two actuators equally, namely, bending moment will not be
applied to the specimen.; at STEP 3, the bending moment is applied by increasing the load of one actuator
and decreasing the load of another actuator by the same rate. lltis procedure of loading is schematically
shown in Fig. 6.7.
In tltc experiment, the applied load, the average curvature of the specimen, the bolt force, the
defonnation of the bolt shank, the dcfonnation of the flange plate and defonnation of the tube arc
measured by on-line measurement using the micro computer through GP-18. Each measurement is
158
swnmarized as follows:
(I)The applied load
The applied load is measured by tlte load cell built in the actuator.
(2)The average curvature
The average curvature of the specimen is calculated by tlte displacements at the compressive
and tensile side. The formula to obtain the average curvature is already described in section
6.2.1 (Eq. 6.2).
(3)The bolt force and the strain distribution at the bolt shank
The bolt force and the strain distribution are measured by the strain gages glued on the bolt
shank as shown in Fig. 6.8. The average bolt force is obtained by averaging readings of 4 strain
gages, and bending defonuation of tlle bolt can be obtained by subtracting a reading of the
strain gage from other reading.
(4)1lte local defonnation of the flange plate and the tube
The local defomtation is measured by the strain gages glued on the flange plate and tlle tube.
The local defonnation of the flange plate is measured by the stress concentration strain gages
glued on the flange plate adjacent the tube members. The local defonnation of the tube is also
measured by the uni-axial strain gages glued on the inner/outer surfaces of the tube. lltcsc
locations are shom1 in Fig. 6.9
The number of total measuring points is I 00, that is, 40 points for measuring bolt force, 20 points for
deformation of the flange plate and 40 points for defonnation of the tube.
(c)Test specimens
In these experiments, 12 specimens are made, 6 specimens are used for tlte pure bending test and the rest
arc used for tlte combined loading test. The dimensions of 12 specimens are sununarized in Table 6.1 . lllc
geometrical configurations of the specimens are shown in Fig. 6.1 0. As for the specimen name, t1te first
character " BL" or " CL" indicates tltc experiment type, " BL'' indicates the specimen for tlte pure bending
test and "CL" indicates the specimen for the combined loading test. The second character ''S" or "L"
indicates the diameter of the tube, '"S'' indicates the specimen with smaller diantcter tube and " L"
indicates tlte specimen with larger diameter of the tube, 216.3 (nun) and 267.4 (nun) respectively. lltc
nuntber following the character ''TH'' denotes the t1tickness of the flange plate in millimeters. The nuntbcr
following character " P" denotes the axial force ratio to the total bolt pre-stress force.
llte materials used to fabricate the specimens arc STK400 steel for the tube and SS400 steel
for the fl ange plate. High strength bolts used in the specimens are M20(FlOT), whose nominal diameter is
20 (nuu). Tite bolt pre-stress force prescribed in JIS, namely, 178.4 (kN), is given to the bolt[4]. In case
of pure bending test, the bolt pre-stress force is given by the torque wrench, on the other hand, in case of
159
 combined loading test, the bolt pre-stress force is given by checking the reading of 4 strain gages glued on
the bolt shank.
Since the thickness of the flange plate is concluded from the results of previous study[3J to be a
very important parameter which affect on the mechanical behavior, t11e tl1ickncss of the flange plate is
varied as the primary parameter. In addition, tl1c diameter of the tube is varied as the secondary parameter
in order to investigate the scale effect for application to the actual structures. The thickness of the tubes is
determined to keep the radius to thickness ratio constant, so t11at bending strength of the tubes can be kept
constant. Besides, in order to investigate the effect of t11c axial force, ti1e axial force is varied as t11e third
parameter. The axial force given to the specimens is dctem1ined based on t11e total bolt pre-stress force
given to 16 bolts, tl1at is, 2854 (kN). By considering the actual axial force applied to tlle real
structures(stcel erosion control dam), tlle ratio is set to I0% and 17 %.
Before carrying out tl1cse two experiments, the material tests by JIS were carried out for
STK400 steel(tubc) and SS400 steel (flange plate). The results of the material test are sununarized in
Table 6.2. It is understood that the material constants are considered to be standard values of steel
generally used and prescribed in JIS[5][6).
6.2 .2 Experimental results and discussions
(a)Pure bending test
1)Load-deformation characteristics
Specimens were loaded until tl1e ultimate strength was obtained except for BL-S-TH6. The bending
moment-curvature curves of all the specimens arc shown in Fig. 6.10. The horizontal axis shows the
curvature nonnalized by yield value of the tube and the vertical axis shows the bending moment
nom1alized by yield value of the tube as well. These values are corresponded to the initial yielding of the
tube based on tl1e material test results. TI1e initial yielding s-trength and corresponding the cun~ature, tJ1e
position of initial yielding, ultimate strcngtl1 and tlle corresponding curvature for each specimen are
summarized in Table 6.3. As for the detemtination of initial yielding, it is based on the local strain of the
tube and the flange plate. That is, when any strain gage indicated to reach to yielding strain, tl1e load is
determined to be the initial yielding load.
The initial yielding strength under the bending of both S type and L type specimens increases as
the flange plate becomes thicker( except for BL-L-TH I 9 and BL-L-TH25). All the specimens initially
yield at the tube except for BL-S-TH6, whose initial yielding occurred at the flange plate. TI1e use of
thinner flange plate rclati\c to the wall thickness of the tube causes the premature failure at the flange
plate. On tl1e other hand, in case that the flange plate becomes much thicker, it becomes hardly defonned
and the failure may taJ...c place in the tube such as the buckling of the tube. Furthcnnorc, the ultimate
strength and initial stifTncss of the joint increase as the thickness of the flange plate becomes larger.
However, the ultimate strength becomes almost the same C\'Cn though the thickness of the flange plate
160
increases because of the failure at tile tube. As a matter of fact, the load-defom1ation cunes of BL-L-
TH19 and BL-L-TH25 can be seen to be almost identical; moreover, that of S type can be also identical
by using thicker flange plate. It is considered in case of thicker flange plates ti1at the strcngtl1 depends
only on the buckling strength of tl1c tube and tltat the load-defonnation curve can be given by that of the
tube.
2)Deformation characteristics and failure mode
The local defonnation characteristics of each specimen are sununarized in Table 6.4, and the local
deformation of the specimens are also shown in Photo 6.3. These photos arc taken from ti1c specimen split
longitudinally.
From tl1ese photos, it is found that the deformation of t11e flange plate of BL-S-TH6 is very
severe; on the other hand such a severe deformation of tlle flange plate is not observed in BL-S-TH 16.
The load transferring mechanism of the tube flange joint can be sununarized based on these local
deformation as follows: When the joint is subjected to bending, the compressive stress of t11e tube at the
joint is transferred to the other side through the flange plate only. On the other hand, the tensile stress at
the joint is transferred to the otl1er side through both tl1e flange plate and tl1e high strength bolts. TI1crcfore,
the failure mode of the tube flange joint can be classified into the following three modes: 1)local buckling
of tlle tube, 2) local deformation the flange plate and 3) breaking of the bolt. Parameters to affect the
above failure modes are considered to be radius-to-tlllckness ratio of the tube, tlle tllickness of the flange
plate, bolt pre-stress force of tl1e bolt and the tensile strength of tl1e bolt and so on.
3)Deformation of the tube
The strain distributions of the tube for BL-S-TH6 and BL-S-TH 16 based on d1e strain gages glued on
inner/outer surface of the tube are shown as a typical example in Fig. 6.12. The strain in t11ese figures is
nom1alized by its yield strain obtained from the material test as already shown in Table 6.2. TI1c mean
strain is obtained by averaging tl1e strains measured on botil inner/outer surface of the tube. The bending
strain is also calculated by multiplying one half by tile difference between inner and outer strain , which
can represent the local out-of plane dcfonnation of the tltin wall of tile tube. If the bending strain is small,
such a defom1ation by local buckling docs not occur, on tltc other hand, if it is large under the
compression, a local buckling may occur because of losing stability w1der compression.
At the early stage of loading when the local buckling is not observed yet, the strain distributions
of BL-S-TH 16 both at the section adjacent to the flange plate and far from it are the same and the
Bernoulli-Euler's hypothesis is found to be applicable. In addition, the strain at both compressive and
tensile sides is distributed unifonnly along the axial direction of the tube. On t11c other hand, that of BL-S-
TH6 is distorted significantly and docs not correspond to that of the distribution given by Bernoulli-
Euler's hypothesis. Tit is is caused by the local defonnation of the flange plate. At ti1e same time, the axial
strain at the compression side of the tube becomes non-unifonn, \\hich implies the local out-of -plane
defonnation of the tube may occur. Since the thick flange plate can prevent the cross section of the tube
161
from the distortion as well as the out-of-plane defomtation, it is considered that the location where out-of-
plane deformation occurs moves away from the flange plate as the thickness of the flange plate increases.
Therefore, it is concluded that in case of the joints with a thinner flange plate such as BL-S-TH6 and BL-
L-TH 12 type, such an distortion of the strain distribution is considered to be significant; and that for the
joint with thicker flange plate, the distortion of the strain distribution of the tube is not so significant by
further loading even beyond the initial yielding.
4)Deformation of the flange plate
The strain distributions of the flange plate in both tlte tangential and radial directions of BL-S-TH6 and
BL-S-TH 16 are shown as a typical example in Fig. 6.13 as well as Fig. 6.12. llte dashed line in tltis
figure is drawn by connecting the location of triangular rosettes.
It is understood from this figure that the flange plate is compressed and tensioned in the radial
direction at the compressive and tensile sides, respectively. This is caused by the preventing tlte cross
sectional change of the tube, namely, the expansion of the tube at the compressive side and the contraction
of the tube at the tensile side In addition, it is found that tltc strain in the circwnferential direction is not as
high as that of the radial direction. In general, these strains of the flange plate are not so large at the early
stage of loading, except for the thinner flange plate in which the level of the strain of the flange plate
becomes more than I 0 times of yield strain for BL-S-TH6 when the joint reaches to its ultimate state.
Howe\ cr, for BL-L-TH J 9 and BL-L-TH25, it is just as the same as yield strain, namely the failure
occurred in the tube; not the flange plate.
The principal strain and its direction of BL-S-TH6 and BL-S-TH16 are shown as a typical
example in Fig. 6.14. It is understood that tlte principal direction coincides with the radial direction at
both upper and lower section of a tube. Although tlte principal direction is not orthogonal to the tube at the
middle section, tlte difference is not so large tltat tlte asswnption used in tlte split tee type of the joint can
be applied directly for a tube flange joint subjected to bending.
(b) Combined loading test
1)Load-deformation characteristics
The bending moment vs. average curvature curves of all the specimen obtained from tlte experiment are
shown in Fig. 6. 15 and the yielding moment and the corresponding cunature dctcnnined based on Joad-
defomtation cunes arc listed in Table 6.5. The horizontal axis shows the nonnaliL.ed cunature and the
vertical axis shows the nonnalized bending moment. lltcsc are nonnalizcd by the yield values of the tube
subjected to pure bending based on the material test results. In this figure, yielding moment considering
the axial force. MP>., MP> 10 and MJ-17 arc also shown for reference. MP>, Mpy1o and Mpy17 are the yield
moment under tension whose magnitude is 0 % , I 0 % and 17 % of the total bolt pre-stress force
respectively. The yielding moment and corresponding cunaturc of the tube are shown in this table for
reference. llte ~ iclding moment of the specimens is defined by the point of intersection of both the
162
regression lines in the clastic area and plastic area of the moment-cunature cune.
It is fowtd from this figure that the higher the tensile axial force is applied and the thinner tlte
flange plate is, the lower the stiffness and the yield moment are obtained. In particular, it is found that
CL-S-TH I 0-P 17, which has tlte thinnest flange plate and the highest axial tensile load applied, is
significantly defomted at the early stage of loading. On the other hand, in case of thicker flange plate, the
.
load-deformation characteristics ofCL-S-TH22-POO is similar to that ofCL-S-TH22-PIO tltcrefore it is .
concluded that tlte decrease of yield moment can be small enough to neglect axial force if it is I0 % of tltc
total bolt pre-stress force.
The yield moment of CL-S-TH I 0-PIO is about 46% of tltc yield moment of CL-S-THlO-POO
and that of CL-S-TH 10-P 17 is only 17%, so tltat the effect of axial force on bending strengtlt is
significant if the flange plate is tltinner. On the other hand, the yield moment ofCL-S-TH22-PIO is almost
same as tlte yield moment of CL-S-TH22-POO, that of CL-S-TH22-Pl7 is about 78%. It is concluded that
the joint with tJunner flange plate is more affected by the axial force tl1an tl1e joint witl1 tlticker flange
plate. In addition, in case of thicker flange plate, it is found that the ratio of yield moment of tlte specimen
to that of the tube considering the tensile axial force is high, so tltat it is understood that the yield moment
of the specimen with tltickcr flange plate can be expected as high as that of the tube. But, in case of
thinner flange plate, this ratio is not so high, it is understood that the yield moment of the specimen is
much lower than that of the tube. Therefore, the high strengtlt bolted tube flange joint with tJucker flange
plate is considered to be comparative to the tube without joint section from the view point of dcfomtahon
characteristics and the strength.
2) Initial yield state
The location of first yielding and corresponding strain of several sections such as the tube, the flange plate
are listed in Table 6.6. In this table, masking cell denotes tlte section where the first yield occurs.
It is found from this table that tlte specimen with thinner flange plate firstly yielded at tlte flange
plate; on tlte other hand, tlte specimen with thicker flange plate firstly yielded at tlte tube. For cxantple,
the first yield section of CL-S-TH I 0-AOO is the flange plate at the tensile side, CL-S-TH22-POO is tlte
tube at the compressive side and CL-S-TH22-P I0 and CL-S-TH22-P 17 is tltc tube at tlte tensile side.
Particularly, tlte specimen with thinner flange plate yielded at the loading stage when the tensile load IS
being applied, but tlte bending moment is not applied; therefore, the flange plate is considered to be axi-
symmetrically defonned wtdcr tension. Moreover, it is found that first yielding of the specimen with
tltinner flange plate occurs earlier than that witlt thicker flange plate. lltcrcfore, it is concluded that the
mechanical behavior of the joint with thicker flange plate depend on that of the tube; on tlte other hand,
the mechanical behavior of the joint with thinner flange plate depends on tltat of the flange plate.
3) Deformation of the flange plate
Strain distribution of the flange plate for all the specimens obtained from the reading of the strain gages
on tlte flange plate arc shown in Fig. 6.16. llte hori.wntal axis shows the strain nonnalizcd by the yield
163
 strain of the flange plate obtamed from the matenal test. 11te ,ertical axis sho,,s the location where the
strain is measured. The center of the tube is set to 1.:ero, positive value indicates compressive side and the
negative value also indicates tensile side. In addition, the strains at the compressive and tensile side of the
tube obtained from the uni-axial strain gages glued on the tube are shown for reference.
It is found from these figures that in case of the thinner flange plate, the flange plate rs
significantly defom1ed but the tube is not defonned. ln particular, the strain at the adjacent the tube is
much higher than the yield strain of the flange plate. On the other hand, it is found that in case of thicker
flange plate, the tube is significantly defornted, and tl1e flange plate is not defonned. In general, it is
observed that tlte tensile strain at the tensile side further increases and little compressive strain at the
compressive side occurs as the applied moment becomes larger. It is considered to be caused by the tube
defonnation such as that the tube contracts at the tension side and expands at tl1e compressive side under
bending. As a result, it is considered that the flange plate is compressed in radical direction at the
compressiYe side and it is tensioned at the tensile side as sho"n in Fig. 6.17. l11is phenomenon is also
observed in the pure bending test and discussed in the previous section. In addition, it is understood that
the strain is relati\e small in case of thicker flange plate.
4)Average axial strain distribution of the tube and the bolt force distribution
Average a\.ial strain distribution of tlte tube and the bolt force distribution of all the specimens are sho\m
in Fig. 6.18. In these figures, the vertical axis shows the location where the strain gages are glued on the
surface of tlte tube. The center of the tube is also set to zero and the positiYe value indicates compressive
side and negative value indicates tensile side. Tite upper horizontal axis shows tlte increase of the bolt
force and the lower horizontal axis shows the average axial strain of the tube.
At the loading stage when the tensile load is being applied, tl1e strain distribution of the tube is
unifonn except for CL-S-TH I 0-P 17; therefore, it is consi~ered that the tensile load is applied unifom1ly
even though two acttrators were controlled independently. But in case of CL-S-TH I O-Pl7, lhe strain
distribution of the tube is not unifonn due to the large defonnation of the flange plate. At this loading
stage, in case of thicker flange plate, the bolt force is not changed; whereas, in case of thinner flange plate,
the bolt force decreased significantly. Particularly, in case of CL-TH I 0-Pl 0, the decrease of the bolt
force is about I0% of bolt pre-stress force, and in case of CL-TH 10-P 17, tl1e decrease of the bolt force is
about 20% of the bolt pre-stress force. Tite fact tltat the bolt force is not changed in case of thicker Oange
plate is good agreement with the previous study on split tee flange joints, namely, tlte applied load is
carried by the release of the compressi\'e force between two circular plates given by the bolt pre-stress
force. On the other hand, the phenomena that the bolt force decreased in case of thirmer Oange plate can
be explained by the decrease of the tl1iclncss of the flange plate due to by yielding.
At the loading stage where the bending moment is applied to the specimen, in case of thicker
flange plate, the strain distribution of the tube is linear and it is considered that Bernoulli Euler's
hypothesis is applicable. In addition, it is found that change of the bolt force is almost zero. It is caused by
164
the exist of the bolt pre-stress force as already mentioned abo\e. As for the neutral axts, it is found that
tlte position of the neutral axis subjected to pure bending is the center of the tube and the position of tlus
axis moves to the compressive side as the applied tensile force increased. Particularly, in case of CL-
TH22-P 17, the position of neutral axis is considered to be out of the cross section of the tube. However,
at the loading stage when MIMy is about 1.0, the strain distribution of the tube becomes non-linear, it is
considered to be caused by the out-of-plane deforn1ation of the tube. The strain distribution of the tube is
distorted significantly after the moment reaches to My because of yielding of tl1e tube.
ln case of thinner flange plate, the strain distribution of the bolt is not linear even at tltc earlier
loading stage. Particularly, in case of CL-TH I0-P J7, the strain distribution of tl1e tube is not unifonu and
bending defonnation of tlte bolt took place. Therefore, Bernoulli Eulers hypothesis is not considered to
be applicable in case of thinner flange plates. It is considered to be caused by the significantly
deformation of the flange plate. As for the neutral axis, even in case of CL-S-TH I0-POO under pure
bending, the position of tl1e neutral axis is not located at the center of the tube. It is caused by the
significantly defonnation of the flange plate. In addition, it is also found that the bolt force is kept
constant while the bending moment is being applied, therefore, it is understood tl1at the decrease of the
bolt force is considered to occur only when the axial tensile load is applied to the specimen.
5) Load-bolt force relation
Load vs. average strain of tl1e bolt curves of all tlte specimens arc shown in Fig. 6.19. In tltesc figures, tlle
horizontal axis shows the average strain of the bolt given by averaging readings of 4 strain gages glued on
the bolt shank as shown in Fig. 6.8, and the vertical axis shows the load normalized by the yielding
moment of the tube. This average strain can be converted tlte bolt force by multiplying cross sectional
area of the bolt shank and elastic modulus.
In case of thinner flange plate, it is found that the bolt force is not changed \vhile tlte axial force
is being applied up to about 100 (kN), after tltcn tlte bolt force decreases as the axial force is applied. 1n
addition, in case of CL-S-TH I 0-P 17, it is observed tl1at each bolt force is different even though the axial
force is applied as compared with the case of CL-S-TH I 0-P I 0. Such a decrease of tl1e bolt force is caused
by the decrease of the thickness of the flange plate due to the yield of the flange plate. Furthern10rc, at tl1e
loading stage when the bending moment is being applied, the bolt force decreases significantly at the
tension side as the bending moment becomes large, at the compressive side, the bolt force almost does not
decrease. Titerefore it is considered that tlte compressive load is transferred through only the tube.
On the other hand, in case of thicker Oange plate, even at the loading stage when the tensile
load is being applied, the bolt force didn t change. It is due to the thick flange plate, and this result is good
agreement with that of the split tee flange joints. 1n addition, at the loading stage when the bending
moment is being applied, at the compressive side, the bolt force decreases a little; on the other hand, at the
tension side, the bolt force increases. It is understood tl1at the change of the bolt force in case of thicker
flange plate is smaller than that in case of the thinner flange plate. This result is also good agreement with
165
 that of the split tee flange joints.
6) Deformation of the bolt
Load vs. bending strain of the bolt curves of all the spccnnens are shown in Fig. 6.20. In these figures, the
hori.rontal axis shows the bending strain at the inner/outer side as shown in Fig. 6.8, and the vertical axis
shows the applied load. If this strain is positive, the bolt is bent in the radial direction to the outer side of
the tube, and if this strain is almost zero, the bolt is not bent.
In case of tJ1icker flange plate, at the loading stage when the tensile load is being applied, the
bolt is not bent. Therefore, tJ1e tJ1ickness of the flange plate is considered to be tJ1ick enough not to cause
the local dcfonnation of the flange plate and it is understood that the bolts arc unifonnly elongated. In
addition, at the loading stage when the bending moment began to be applied, it is found d1at the bolt at the
tension side is significantly bent. Moreover, in case of CL-S-TH22-Pl7, it is observed that ilie strain at
the compressive side decreases. It is caused by the change of the thickness of the flange plate. Therefore,
even in case of thicker flange plate, the bending defonnation of the bolt at the tension side is observed.
On the other hand, in case of ilie thinner flange plate, at the loading stage when ilie tensile load
is being applied, it is found that tJ1e bending strain decreases. TI1is phenomenon is considered to be caused
by not the dcfom1ation of the bolt but tJ1e defom1ation of U1e flange plate. In addition, at the loading stage
"hen tJ1c bending moment is applied, the bending strain at the tension side increases, so that the bolt is
considered to be bent to outer side similar to the case witJ1 thicker flange plate. Therefore, in case of
thinner flange plate, the change of the bolt shank strain is affected significantly by the defom1ation of the
flange plate.
6.3 Simple Design Procedure for the Tube Flange Joints
6.3.1 Current design procedure
In order to design tJ1c high strength bolted tube flange joints, the verification of the working stress to yield
stress is necessary. Since the shape of the cross section of the tube is circular and at the joints the cross
section changes along the member direction, the rational verification of the working stress is very difficult
without any assumptions. Therefore, simple modeling of the joint should be made based on the mechanical
bchaYior.
TI1e current design procedure for the high strength bolted tube flange joints is analogous to that
of the reinforced concrete members, where the cross section of the flange plate above the neutral axis
(compressive side) is asswned to behave as concrete and the high strength bolts below the neutral axis
(tensile side) as steel reinforcement as sh0\\11 in Fig. 6.21. The working stress of the joint has been
traditionally computed based on these simple asswuptions. However, there exist many ambiguities such
that the flange plate in the compressive side can be effective and that the resistance at the tensile side can
166
be provided only by the bolts just like the reinforcing bars m tJ1e RC section without constdering the effect
of the bolt pre-stress force. Watanebe, E. ct al. modified this assun1ption by assuming that tJ1e cross
section of the tube in the compressive side can be effective and that the discrete bolt cross section can be
replaced by that of continuous circular ring for simplification; then, the working stress of the flange plate
is computed based on the model as shown in Fig. 6.22[7]. This model is cantilever circular ring plate
fixed along the inner edge and subjected to concentrated load corresponding to tJ1e bolt force at tJ1e bolt
location. The bending stress at the fixed inner edge is computed by solving the differential equation of the
circular plate with Fourier series[8]. It was found that the proposed assun1ptions can provide tJ1e rational
and conservative evaluation of U1e working stress. It is also suggested U1at the effective cross sectional
area both at the compressive side and the tensile side must be assessed in order to evaluate the accurate
working stress. Particularly, the increase of the effective cross sectional area at the tensile side due to the
bolt pre-stress force is suggested to be studied.
6.3.2 Simple design procedure
In this section, based on the effective cross sectional area evaluated in the previous chapters, further
modification of current design procedure was made in order to establish tJ1e rational design procedure.
TI1e flow of current simple design procedure is in the following.
STEP 1: Determination of the neutral axis position and calculation of tJ1e working stress at
the joint plane;
STEP 2: Calculation of the working stress at the tube; and,
STEP 3: Calculation of the working stress at the flange plate.
At STEP 1, the assumption on eiTective cross sectional area at the joint plane carrying the external
applied load, is the most important and it is assessed that tJ1e cross sectional area at both compressive and
tensile side as shown in Fig. 6.23(a) are effective. At tJ1c compressive side, the effective cross sectional
area, Acrr may be larger than tJ1e cross sectional area of the tube, Ap. Therefore, the ratio of Acn/Ap is
detennined by using axisynm1etric finite element analysis for the model as shown in Fig. 6.23(b). Based
on analytical results for the case that the thickness of the flange plate is I0 (nm1) and 22 (mrn), Acn/Ap is
obtained to be 2.24 and 3.55 respectively. According to this ratio, the fictitious tube thickness as shown in
Fig. 6.23(a) is considered as the effective cross sectional area at the compressive side. On the other hand,
the effective cross sectional area at the tensile side is dctcnnined by using BAF model as discussed in
Chapter 3, where tJ1e proposed fonnula (Eq. 3.7) for the evaluation of tJ1e stiffness of BAF model is used
and the stiiTncss is converted into the cross sectional area. Tile effective cross sectional area at tl1e tensile
side is detcnnined to be 112 of the effective cross sectional area of BAF model considering the difference
of loading condition between tJ1c tensile tube joint and BAF model. As discussed in Chapter 3, the initial
167
high stiffness decreases and tends to be constant until the sudden decrease occur due to the yielding of
bolts as extemal applied load increases. Although it is relatively small compared to t11e initial high
stiffness of BAF model, this constant value is considered for conservative evaluation. Therefore, the
effective cross sectional area at the tensile side is assun1ed to be 2.45 (cm 2) for t11e flange plate thickness
of 10 (mm) and 2.57 (cm 2) for 22 (rnm). Based on the effective cross sectional area at both compressive
and tensile sides, the stress at the joint plane is calculated by assuming Bemoulli-Euler's hypothesis in the
same way as the current design procedure. TI1is hypothesis is considered to be applicable according to t11e
results obtained from t11e experiments (referred to section 6.2) if the applied external load is below the
yield strength. At STEP 2, the simple model as shown in Fig. 6.23(c) is asswned. This model is a strip
with unit width cut in dcpt11 direction. Using this model, t11c working stress at the tube is computed from
the working stress at t11e joint plane considering equilibrium of forces from the joint and the tube.
Compressive force P1 and tensile force P2 at the tube arc obtained by following equations.
(6.3)
P..2_ L+ac (6.4)
- 2
L obtained from the current design procedure. In particular, it is found t11at the location of the initial yielding
in which, C~, C2, P~, P2, L and a is compressive force at the joint plane, tensile force at the joint plane,
compressive force at the tube, tensile force at t11e tube, distance between the center of the tube wall and
distance between the center of t11e bolt and the center of the tube wall respectively.
At STEP 3, in order to compute the working stress at the flange plate, a strip is considered as t11e clamped
beam shown in Fig. 6.23(d). The length of this beam corresponds to the distance to the position of the
high strength bolt at the tensile side from the neutral axis. ~= boundary condition at bot11 ends of ~e strip
is also considered to be applicable from the results of split tee flange joints in Chapter 4. Maxiniwn stress
C7ma.x is obtained by t11e following equation.
(6.5)
in which, amA.x, t, a and b, I, Pis maximun1 stress, thickness of the flange plate, distance between the fixed
end and loading point, and applied load respectively.
6.3.3 Results and discussions
TI1c working stress is computed using proposed design procedure for the test specimens subjected to
tension and bending as described in the previous section 6.2. l11c working stress obtained from the
proposed simple design procedure are shown in Table 6. 7. Applied external load at the first yielding
obtained from the combined loading test is given as design load. In this table, the working stress obtained
168
from the current design procedure and experimental results are also shown for comparison. In addition, in
case of CL-S-THIO-PlO and CL-S-THIO-PI7, the first yielding occurred at the flange plate when the
axial force is applied, so that the strip with the boundary conditions as shown in Fig. 6.24 is used to
compute the working stress of the flange plate in case of only an axial force applying. Two cases such as
the simple supported beam and the clamped beam at bot11 ends arc considered.
As for the loading stage in which the axial force is applied, it is found t11at by t11c simple
supported beam t11e working stresses arc in good agreement wit11 the experimental results in case of
thinner flange plate; on the other hand, in case of thicker flange plate, by the clamped beam analytical
results agree well with the experimental results. Therefore, it is considered that the boundary condition at
the ends of the strip depends on the thickness of the flange plate, and should be considered as hinge for
thinner flange plate and t11e fixed end for thicker flange plate. In general, the evaluation of the working
stress of the flange plate can be conservative by the simple supported beam for any thickness of the flange
plate. Furthermore, the more rational evaluation can be made by adding elastic constraint at the ends of
the strip according to the thickness of the flange plate.
As for the working stress of the tube and the flange plate under bending, it is found from Table
6. 7 that the working stress obtained from the proposed design procedure is considerably accurate tJ1an that
is well predicted by using the proposed design procedure except for CL-S-TH-22-POO. Since t11e working
stress can be evaluated wit11 enough accuracy by the proposed design procedure, it is considered that tJ1c
proposed design procedure can be applied to design the tube flange joints. However, in order to evaluate
the working stress for t11e wide range of structural dimensions, further study on the effect of the bolt group
and the boundary conditions in modeling may be required.
6.4 Conclusions and Future Needs
In iliis chapter, the behavior of high strength bolted steel tube flange joints subjected to tension and
bending is studied experimentally as an application of high strength bolted tensile joints. Not only the
global mechanical beha,ior such as load Lransfcrring mechanism and failure mechanism but also local
mechanical behavior of the flange plate, the bolt and t11e tube are discussed based on the experimental
observations. In addition, t11e simple design procedure for t11e high strength bolted tube flange joints by
means of the results of previous chapters arc proposed and its applicability is assessed. The following
conclusions are obtained:
I) In case of thinner flange plate, the mechanical behavior of the joint depends on tl1at of the
flange plate and the bolts are forced to be bent significantly due to the local dcfonnation of the
169
 Table 6. 1 Geometrical Configurations of the Specimens
(unit mm)
flange plate; on the other hand, in case of thicker flange plate, the mechanical behavior can be Specimen Diameter Wall Thickness Diameter Length Radius to
identical to that of the tube. of Tube Thickness of Flange of Flange of Test Thickness
of Tube Plate Plate Section Ratio
2) When the thicker flange plate is used, the interaction between the flange plate and the tube is
BL-L-TH12 12 212
prevented, and Bernoulli-Euler's hypothesis is considered to be applicable at the joint section.
BL-L-TH19 267.4 6.0 19 450 219 44.6
In addition, the initial stiffness, the yield strength and the ultimate strength can be made larger. BL-L-TH25 25 225
However, the further increase of the ultimate strength by much thicker flange plate cannot be BL-S-TH6 6 206
BL-S-TH10 IO 210
expected by the buckling of the tube. BL-S-TH16 216.3 4.5 16 400 216 48.1
3) In case of thinner flange plate, the decrease of the bolt force is observed. The bolt force is found CL-S-TH10 10 210
to be very sensitive to the change of the flange plate thickness, even if it is induced by in-plane CL-S-TH22 22 222
[Note] Tube: STK400 Flange Plate: SS400 H1gl1 Strength Bolt : M20(FlOT)
defonnation of the flange plate.
Bolt Pre-stress Force : 178(kN)
4) By proposed simple design procedure considering the effective cross sectional area, the working
stress can be evaluated more accurately than the current design procedure. In particular, the
initial yielding and working stress of tJ1e tube arc in good agreement with experimental results. Table 6.2 Results of Material Tests
(a) BL-test
In the future, t11e fatigue strength of tl1e tube flange joints should be studied. Moreover, tlte proposed
Ltype s
type
SS400 0.323 0.323
design procedure should be further improved in order to be applied for tl1e various ultimate states of the
STK400 0.372 0.377
joints. 2..
(urut : kNinun )
(b) CL-test
TH10-POO THIO-PlO TH10-Pl 7 TH22-POO TH22-Pl0 TH22-P17
References SS400 0.2576 0.2576 0.2592 0.253 0.253 0.2489
I) Japan Road Association : Specifications of Highway Bridges(JSHB), Maruzen, 1991(in Japanese). STK400 0.4I9 0.4 19 0.417 0.419 0.419 0.417
2..
2) Architectural Institute of Japan : Reconunendation for the Design and Fabrication of Tubular [Note] (urut : kNinun )
Tube : STK400 Flange Plate : SS400
Structures in Steel, Maruzen, Jan. 1990(in Japanese).
3) T.Yamaguch, E.Watanabe, K.Sugiura, S.Kasai : Experin1ental Study on High Strength Bolted Tube
Ta bl e 6.3 Pure Bendin1g Test Resu ts
Flange Joints, Proc. of Annual Conference of Civil Engineers, JSCE Kansai Chapter, Jun. 1991(in
Specimen Thickness Radius Initial Yielding Strength Ultimate Strength
Japanese). of -to-
4) Japanese Industrial Standard Conunittee : Sets of High Strength Hexagon Bolt, Hexagon Nut and Flange Plate Thickness (M I My) (I (Jy) (M I My) ( I y)
(mm) Ratio
Plain Washers for Friction Grip Joints(B1186), 1979 (in Japanese).
BL-L-TH12 12 44.6 0.310 0.871 1.12 18.0
5) Japanese Industrial Standard Conunittee : Rolled Steel for General Structure(G31 0 I), 1987 (in
BL-L-TH 19 19 44 .6 0.365 0.477 1.40 8.39
Japanese). BL-L-TH25 25 44 .6 0.321 0.372 1.39 8.21
6) Japanese Industrial Standard Conunittee : Carbon Steel Tubes for General Structural BL-S-TH6 6 48 . 1 0.340 3.5 1 0.776 24. 1
BL-S-TH10 10 48 . 1 0.390 0.856 1.13 12.5
Purposes(G3444), 1994 (in Japanese).
BL-S-TH16 16 48 . 1 0.400 0.579 1.29 6 . 16
7) E.Watanabe, K.Sugiura, T. Yamaguchi, S.Kasai : Design Method of High Strength Bolted Tube
Flange Joints, Journal of structural Engineering, JSCE, Vol. 38A, Mar. 1992, pp. l-12(in Japanese).
8) S.Timoshenko, S. Woino,.,sJ..:y-krieger : Theory of Plates and Shells, Second edition, McGraw-Hill,
I968.

170
171
 Table 6.6 List of Strain at the Initial Yicldi

Table 6 4 Defonnation Characteristics Location

BL-S-TH6 BL-S-THIO BL-S-TH16 BL-L-THI2 BL-L-TH19 BL-L-TH25 CL-S-THIO-POO Tube Side! -0.6275 0.4078
Flange Plate TF TF TF( ) TF TF() TF( )
Side2 -0.6233 0.4078
Tube CT CT CT CT CT CT
epy= 1816 ()J)
[Note]
TF : Local deformation of the flange plate at the tension side is observed.
CL-S-TH 10-P I 0 Tube
CT: Local defonnation of the tube at the compression side is observed.
: Bernoulli-Euler's hypothesis may be applied at the joint section.
&py= 1816 ()J) Flange Plate

CL-S-TH 10-P 17 Tube

Ta ble 6.5 Com b'med Loa d'tng Test Resu ts

CL-TH I 0-POO CL-TH10-PIO CL-THIO-Pl7 CL-TH22-POO CL-TH22-PIO CL-TH22-Pl7

4 4 4 4 4
Mv(kNmm) 4.33410 1.99910 0.723510 6.964 10 7.1261 0 5.379 10 4
CL-S-TH12-POO Tube
v (1/m.m) 2.589lo5 L6811o 5 1.746to5 1.364105 1.6 181o5 1.749 105
4 4 Side2 -1.0105 0.8078 2
Mnv (kNoun) 6.508 104 6.5o81o 6.477 10 6.508 104 6.508 10 4 6.477 104
&py= 1816 ()J) Flange Plate Side I 0.3545 0.8078 F6
r.v ( 1/oun) 1.679105 1.679105 1.6951o5 1.6791 o5 1.679105 1.695 105 r-==~r-~~~~~~~~
Side2 0.3765 0.8078 F6
MJ M"" 0.6659 (0.3072) (0.1117) 1.070 (1.0950) (0.8305)
CL-S-TH22-P10 Tube
(M/~.)/(M~/~~") 0.43 19 0.3068 0.1084 1.317 1.136 0.8048
M) Mnvto - 0.3976 - - 1.4174
M/ Mnvt7 - - 0. 1827 - - 1.3582

CL-S-TH22-Pl7 Tube
LNotej
Mpyto = 0.7725 (kNnun), Mpy 17 = 0.6 11 4 (kNnun)
epy= 1834 ()J) Flange Plate Side ! 0.7142 0.6006 F6
r-~=--r~~~~~~~~~~
Side2 0.7080 0.6006 F6
&py : Yield Strain of the Tube
[NOTE] Strain Gage Location
Ery : Yilcd Strain of the Flange Plate

rn
Compressi\'e Side

Stress Concentration Gage

(Compressive Side)

Tensile Side Stress Concentration Gage

(Tensile Sic.lc)

0 : High Strength Bolt - A'\ial Strain Gage

172 17:3
 Table 6.7 Results of Proposed Simple Design Procedure
(a)Working Stress of the Flan e Plate(Axial Force appl ,ing)
Specimen Experimental Results Working Stress Working Stress
(Flange Plate) (Clwnpcd Beam) (Clumped Beam)
CL-S-THIO-PIO yield 0.0512(0.198) 0.303( 1.17)
CL-S-THI O-P17 yield 0.138(0.534) 0.818(3.17)
CL-S-TH22-P10 not yield 0.0486(0.187) 0.288(1.11)
CL-S-TH22-Pl0 not yield 0.0819(0.316) 0.488( 1.88)
[NOTE] (a) With Rib Plates
2
unit : kN/mm
Applied axial load for the case of thinner flange plate is initial yielding load and in case of thicker
flange plate is prescribed axial force.
The value in( ) is the stress nom1alized by the yield stress of each material.

(b)Working Stress (Bending Moment applying)(Proposed Design Procedure)

Specimen Experiment Stress at the Tube Stress at the Tube Stress at the Prediction
(Yielding (Compressive (Tensile Side) Flange Plate (Yielding
Position) Side) Position)
CL-S-TH 10-POO FL 0.0649(0.155) 0.130(0.312) 0.175(0.678) FL
CL-S-TH22-POO CT 0.239(0.572) 0.397(0.95) 0.120(0.461) TT
CL-S-TH22-P10 TT 0.337(0.806) 0.458(1.09) 0.107(0.414) TT
CL-S-TH22-P17 TT 0.319(0. 762) 0.394(0.943) 0.0922(0.356) TT

(e)Working Stress (Bending Moment applying)(Current Design Procedure) (b) Without Rib Plates
Specimen Experiment Stress at the Tube Stress at the Tube Stress at the Prediction
(Yielding (Compressive (Tensile Side) Flange Plate (Yielding
Position) Side) Position)
CL-S-TH I 0-POO FL 0.0787(0.188) 0.0459(0.11 0) 1.14(4.4:n FL Fig. 6.1 Typical Types of High Strength Bolted Tube Flange Joints
CL-S-TH22-POO CT 0.250(0.599) 0.146(0.350) 0. 752(2.90) FL
CL-S-TH22-P 10 TT 0.311(0. 743) 0.212(0.507) 0.883(3.41) F~
CL-S-TH22-Pl7 TT 0.279(0.337) 0.217(0.520) 0.801(3.09) FL
[NOTE]
unit : kN/nun2
FL, CT and TT denotes the flange plate, the tube at the compressive side and the tube at the tensile
side respectively.
The value in( ) is the stress nonualized by the yield stress of each material

171\ 175
 CD Vertical Actuator
Loading Beam
@ Loading Point (Roller) P/2
@ Roller Support
@ Test Specimen
Specimen
CD

0:
l!J L
Displacement Transducer
Fig. 6.4 Setup of the Displacement Transducer (BL Test)

Fig. 6.2 Test Setup for the Pure Bending Test

P/2 P/2

Specimen~:
'
I
I
i< ~ !!'<:...__--->;
..: ,. .!
: 700 : : 700 : Uniaxial Strain Gage De: Diameter of the Flange Plate
Rectangular Rosette Dt : Diameter of the Tube

Fig. 6.5 Measuring Points of Local Strain (BL Test)

Fig. 6.3 Applied Moment Distribution Diagram

176
177
 O.SP1

(a) Step 1 (Applying Axial Tensile Force)

O.SP1+P2

<D Loading Frame

Non-destructive Section M=2P2L
@Universal Joint
CV Vertical Actuator I
@ Vertical Actuator II (b) Step 2 (Applying Bending Moment)
@ Test Specimen
Fig. 6.6 Test Setup for the Combined Loading Test

Step 1 ~ Step 2
Actuator II

Step 2
Actuator I

Step 1
Time Axial Force

Fig. 6. 7 Loading Procedure (CL Test)

178 179
 364

8 ------
'V

Flange Plate

Bolt
0 0
~ V'\
C"'l 'V
Tube
(unit: mrn)
0
'V
Fig. 6.8 Strain Gages Glued on the Bolt Shank (CL Test)

(a) L Type

I
364

-r
8----
'V
--
0
0
<:;1"

Flange Plate

30
0 0
N 0
....., 'V

- Uniaxial Strain Gage

(unit : mm)
Stress Concentration Strain Gage

Fig. 6.9 Measuring Points of Local Strain (CL Test) (b) S Type

Fig. 6.10 Dimensions of the Specimens

180
181
 My : bending moment of the tube corresponding to
the 1st yielding
y : curvature of the tube corresponding to
........ the 1st yielding

f~ BL-L-TH19
initial yielding state
....
c~
CD-
e
0
z:
0
c<n ..._.
-ci 1.0E'y
"0
c
(!)
en X : radial direction
...,
d 0 : tangential line direction

0
ci
0 5 10 15 20 25 ultimate state
Curvature (1</>y)

Fig. 6.11 Bending Moment-Curvature Curves(BL Test)

Fig. 6.13 Strains at the Radial Direction and at the Tangential Line Direction
flange
convex
~ ~
at the Flange Plate(BL Test)
compression compression I. 0 E't 15.0fy

tube

tension

tension
convex
initial yielding state ultimate state
initial yielding state
(a) BL-S-TH6 * X : one bali the difference
t:::. 0 O: average

ultimate state

initial yielding state ultimate state

(b) BL-S-THlO
Fig. 6.14 Principal Strains and their Directions at the Flange Plate(BL Test)
Fig. 6.12 Strain Distribution at the Axial Dircction(BL Test)

182 183
 Axis for Location
1.4 - , - - - - - - - - - - - - - - - - - - - - ,

1.2 Flange Plate

105.9 - - -- - - '\
\
1.0 I \
I
I '~ Stress Concentration
i5: 0.8 0 '
I Strain Gage

-
~

~ 0.6 -105.9
'
I
I

-160
0.4
' '\
Tensile Side
Uniaxial
0.2
Strain Gage
Uniaxial Strain Gage Tube
0.0 Stress Concentration Strain Gage
0 1 2 3 4 5 6 7 8 9 10 11 12

1 py

---- CL-S-TH22-POO
~ CL-S-TH1 0-POO

- t l- CL-S-TH22-P1 0
Fig. 6.16 Strain Distribution of the Flange Plate (CL Test) (continued)

-o-- CL-S-TH1 O-P1 0

- 4 - CL-S-TH22-P17

-C::r- CL-S-TH1 O-P17

Fig. 6.15 Bending Moment-Curvature Curves (CL Test)

184 185
 1~ .-------~----------------,
1~ ~----------------------. 1~ .---,------,,--------------,
1~ .-------------------------~
120 l~ 120
120 120
I ~
110 110
110 110
e .o. M I M =0.25
e .o. M I M1 = 0.25 e.. e
.. 100 1 .. 100 .o. M I M = 0.25
c v M l M1 = 0.50 c v M I M1 = 0.50 100 1 .. 100 & Ml M =0.25
1
0 c v M I M =0.50 c
0
o M I M1 =1.0 +> o M I M1 =1.0 1 v M I M1 = 0.50
g -100
+>

_,
!3
_,0
-100
0
+>
!3 -100
o M I M1 =1.0 0
'D
!3 -100 o M I M1 = 1.0
-110
_,0 _,0
-110
'' -110 -110
'

-3 -2 -1 0 1 2 3 4 5 6 7 8
-120

-3 -2 -1
Jl
-130 +--.--i--.---,f-.---,--.---r--,--.--1
0 1 2 3 4 5 6 7 8
-120 -120

-130
f!
+--.----.---....:._.,--,--.---,.--.--.,..--.-~
2 3 4 5 6 7 8 -3 -2 -1 0 1 2 3 4 5 6 7 8
cl cy cl ey
c I &y c I liy
(a) CL-S-THIO-POO (b) CL-S-TH22-POO
(g) CL-S-TH 10-POO (h) CL-S-TH22-POO

130 ~-----------------------, 1~.---------~--------------,

1~ .---~---------------------, 1~ ~------------------------~
120 1\ 120
120 120

110 0.5P0 110

110
~

E MI M1 =0.00 e..

0.5P0
M I M1 = 0.00 e..


0.5 P0
M I M1 = 0.00 e
110
e 0.5P0
.. 100 MI M1 =0.25 100
c
.o.
c .o. M I M = 0.25 100 .o. M I M = 0.25 .. 100 M I M1 0.00 =
0 v MI M1 =0.50 ..,
.0 1 c
0
1
M I M1 = 0.50
c
..., .o. M I M =0.25
~ -100 rl v M I M1 = 0.50 v 0
o =1.0 -100 +> 1

_,g
MI M1 ~ -100 ~ =0.50
_,0 o M I M1 = 1.0
_,
o M I M1 = 1.0 -100 v M I M1

-110 -110
0
o M I M1 =1.1
_,0 o M I M1 =1.0
-110 -110

I
-120
0 0

=~
-120 -120
0
-130 -1--.--,-_:;::..::...;.-,--,--;--.--.--,---l ?
-130 +--.---+---.:-+--.---.--.--.--.--.---!
-3 -2 -1 0 1 2 3 4 5 6 7 8 -3 -2 -1 0 1 2 3 4 5 6 7 8
-3 -2 -1 0 1 2 3 4 5 6 7 8 -2 -1 0 2 5 6 7 8
cl ey c l ey
c I &y
(c) CL-S-THIO-PIO (d) CL-S-TH22-Pl0
(i) CL-S-THIO-PIO (j) CL-S-TH22-PIO

130 ~------------------------, 1~ --~----.-----------~---,

130 ~------------------------~ 1~ ~------------------------~
120 120
120 120

110 110
110
e
.. 100


0.5P0
M I M1 =0.00
e.. 100


0.5P0
M I M1 = 0.00 e.. M I M1 = 0.50 e..
110
0 ~ 0.5P0
100 '
c c M I M1 = 0.00 100 Ml M1 =0.00
.o. Ml M = 0.25
...,0 1 ...,0 "' Ml M1 = 0.25 c
0 .o. Ml M = 0.25 c .o. M I M = 0.25
~ -100 v M I M1 : 0.50 ~ -100 v M I M1 = 0.50 +> 1 0
'Q 1
_,0 o MIM 1 =1.0 _,0 -.;,: 0 o M I M1 = 1.0
~ -100 v Ml M1 = 0.50 rl -100 v M I M1 = 0.50
-110
_,0 o MI M =1.0 _,0 o M I M = 1.0
-110 1 1
-110 -110
-120

-3 -2 -1
~
0
.... v
o
-130 +--.---r---.:-..---,--,.---.----,:-..---,---1
1 2 3 4
0

5 6 7 8
-120

-2 -1 0
r
-130 -J--;---,::=-,,-,--,----r-,r-;--,--l
2 3 4 5 6 7 8
-120

-130 -1--.---.---.,--.,..--,---,---.--,-.--,---j
-3 -2 -1 0 1 2 3 4 5 6 7 8 -3 -2 -1 0 1 2 3 4 5 6 7 8
cl ey
c l ey cl &y
(e) CL-S-THIO-Pl7 (f) CL-S-TH22-P 17
(k) CL-S-TH10-Pl7 (I) CL-S-TH22-Pl7

Fig. 6.16 Strain Distribution of the Flange Plate (Side!) (CL Test) (continued)
Fig. 6.16 Strain Distribution of the Flange Plate (Side2) (CL-tcst)

186
187
 Incremental Bolt Force (kN)
Incrementa l Bolt Force (kN)
50 -40 -30 20 10 0 10 20 30 40 50
160 ..5() -40 30 20 10 0 10 20 30 40 50
\> 160

Compressive Side
120

80
~
120

80
,.
I
...,
c
40

0
/} I
c
40

Flange Plate 0

!J
T 3 0

.9 -40 g
~
-40
-80

120 <
,... -80
~
120
,~ , ~ = 1816 (p)
-160 ~= 1816 (p)
160
'-' 1000 500 0 500 1000 -1000 0 500 1000
Average Axial Strain (p)
Average Axial Strain (p)
(a) CL-S-TH10-PIO (b) CL-S-TH22-PIO

Incremental Bolt Force (kN)

,-' Tube 50 -40 30 20 10 0 10 20 30 40 50
Incremental Bolt Force (kN)
' -' 160 ,.....~............................~t-..a...............................~ 50 -40 30 20 10 0 10 20 30 40 50
160
120 I?
120
,~, ,~,
80
80
'-' ' -' e
sc 40 E' 40 '
0
:Q

lJ0
0

-40
sc
.2
(;
0 '
~ 8
~
-40
-80
Tensile Side -80
-120 I
120
~= 1834 (p)
Fig. 6.17 Stress induced by Change of Cross Section of the Tube 160 +--~,.....{;>,-.-~~r-.-~-.-.~~--i ~ =1834 (p)
160
1000 500 0 500 1000 1000 500 0 500 1000
Average Axial Strain (p)
Average Axial Strain (p)
(c) CL-S-TH10-Pl7 (d) CL-S-TH22-Pl7

Axis for Location

Compressive Side

160
105.9

Average Axial Strain (Side!)

0
.A Average Axial Strain (Side2)
0 Incremental Bolt Force
-105.9
-160

Tensile Side

Fig. 6.18 Axial Strain Distribution of the Tube and Strain Distribution of the Bolts (Po) (CL Test)
(continued)

188
189
 Incremental Bolt Force (leN) Incremental Bolt Force (kN)
Incremental Bolt Force (kN) Incremental Bolt Force (kN)
50 .4() 30 -20 10 0 10 20 30 40 50 50 .4() -30 20 10 0 10 20 30 40 50
-50 -4() .3Q 20 -10 0 10 20 30 4() 50 -50 -40 30 20 10 0 10 20 30 4() 50
160 160 160+-~~~~~~~~~--~ 160 +-........................c.........L......<}w-................~..............~
120 120 120 120
eo eo eo eo
e 40 e 40
I
.s .s 4()
,& o+----~--------1

] .4()

-80

120 120 120

~ "1816 {II) "= 1816 {II)
-160 +----.-...,----,..--j:)---,..-,..--,--1 160 +----.--,-----.--{>--.,.--,---.,.--i
160 +---r--r---r-d-+---r-....---r--1 160 +---r-..--,--~-,--,,.-.,.--j
1200 -900 -600 300 0 300 600 900 1200 1200 900 -600 300 0 300 600 900 1200
1 600-1200 -800 -400 0 400 800 1200 1600 1600-1 200 -800 -400 0 400 800 1200 1600
Average Axial Strain (p) Average Axial Strain (p)
Average Axial Strain (p) Average Axial Strain (p)
(a) CL-S-THIO-POO (b) CL-S-TH22-POO
(a) CL-S-THlO-POO (b) CL-S-TH22-POO
Incremental Bolt Force (kN) Incremental Bolt Force
Incremental Bolt Force (kN) Incremental Bolt Force (kN)
-50 .4() 30 20 10 0 10 20 30 4() 50 50 -40 30 20 10 0 10 20 30 4() 50
160~ww~~~~~~~~~~
-50 -40 -30 20 -10 0 10 20 30 4() 50 .so -40 -30 20 10 0 10 20 30 50
160 160
4()
160~~~~--~~~~~~~

120 120
120 120
~
eo 80 eo \ eo
I
c
40 ~ I 40 e
sc:
4()
e
s
40
,..
0 0 ..,
.~ 0+------~-~~----
:8 0
-~ -t------ "*-4t-- - -
g
..J
-40 . ] -40 g .4()
0

..J ] -40
-80 -80
;.. -80
i.
-80
120 < -120 120 <
160
~=1816{11) ~ = 1816 {II)
160
"= 1816 {II)
120
" . 1816 {II)
160 +---r--r---r--1-<5---r--.--....--1
1200 900 -600 300 0 300 600 900 1200 -1200 -900 -600 -300 0 300 600 900 1200
1 600-1200 -800 -400 0 400 800 1200 1600 1600-1200 -800 -400 0 400 800 1200 1600
Average Axial Strain (p) Average Axial Strain (p)
Average Axial Strain (p) Average Axial Sl1a1n {II)
(c) CL-S-THIO-PIO (d) CL-S-TH22-Pl0 (c) CL-S-THIO-PlO (d) CL-S-TH22-Pl0
Incremental Bolt Force (kN) Incremental Bolt Force (kN) Incremental Bolt Force (kN) Incremental Bolt Force (kN)
50 -40 30 20 10 0 10 20 30 40 50 -50 -40 30 20 -10 0 10 20 30 40 50 50 -40 -30 20 10 0 10 20 30 40 50
160 50 -40 -30 20 10 0 10 20 30 40 50
160+-~~~~~+-~~~--~
160 -t-..J..........L.......&.~~+- ........~................L...-...j
120 120 120 120
eo eo eo
e
se
40
.s 40

~ 0 +--~---+---+~~---!
g
..J
-40

-80

120 -120 120

Cy. 1834 {II) ~ = 1834 (p) "y = 1834 {II)
-160 +----r----<>r--,--+----.--,--,---1 -160 +---.-.,.---.r--~-,..--.--..---i . 160 +---r--<>r--r- + --.--,--,.--!
1200-900 -600 300 0 300 600 900 1200 1200900 -600 -300 0 300 600 900 1200 . 1600.1200 -800 -400 0 400 800 1200 1600 1600-1200-800 -400 0 400 800 1200 1600
Average Axial Strain (p) Average Axial Strain (p) Average Axial Strain (p) Average Axial Strain (p)
(e) CL-S-TH10-Pl7 (f) CL-S-TH22-Pl7 (e) CL-S-THIO-Pl7 (f) CL-S-TH22-Pl7

Average Axial Strain (Side!)

Avcrage Axial Strain (Side I)
.A Average Axial Strain (Side2) .A A veragc Axial Strain (Side2)
0 Incremental Bolt Force 0 Incremental Bolt Force
Fig. 6.18 Axial Strain Distribution of the Tube and Strain Distribution of the Bolts (0.25My) (CL Test) Fig. 6.18 Axial Strain Distribution of the Tube and Strain Distribution of the Bolts (0.50My) (CL Test)
(continued) (continued)

190 191
 Incremental Bolt Force (kN) Incremental Bolt Force (kN)
-50 .4() -30 -20 -10 0 10 20 30 40 50 14 ~------------------------.
-50 -40 -30 -20 -10 0 10 20 30 40 50
160 Flange Plate 12
120
I
I~ c 1816 (}Jl 120 :~ = 1816 (JJ) - - Bolt09 - Bolt01
I I 1.0
80 I 80 I
I I
e. 40 I e
.
40 I
I
~'fi. 0.8
I
c c
0 0 ~ 0.6
,g .8
R .4()
g
...J
.4()
0.4
...J

.00 .00
0.2
-120
Tensile Side
-1 20
0.0 +--~--,---.-~-,.~--~~..,.............--~
-160 -160
Bolt 01 -t:r- Bolt 06 -900 -600 -300
-5000 -3000 -1000 1000 3000 5000 -5000 -3000 -1000 1000 3000 5000 0 300
Average Axial Strain (JJ)
-o- Bolt02 -If'- Bolt 07
Average Axial Strain (JJ) Average Strain (p)
Bolt 03 -..J\J- Bolt 08
(a) CL-S-THIO-POO (b) CL-S-TH22-POO -a- Bolt04 -+- Bolt09 Bending Moment-Bolt Strain Curves
- A - Bolt05 -<>-- Bolt 10
Incremental Bolt Force (kN) Incremental Bolt Force (kN) (a) CL-S-THIO-POO
-50 .4() 30 -20 -10 0 10 20 30 40 50 50 .4() -30 20 -10 0 10 20 30 40 50
160 160 500
~: : ~ = 1816 (JJ)
120 ( I 120 h= 1816 (p) 450
1.4
I ~ I I
400 1.2
80 I 80 I
I I 350
e
.
40 I e
.
40 I
I
1.0
c II c
-
z 300 ~'l;. 0.8
-
,g 0 0 0
'D ~ 250
g -40
I
I
I
I
g .4() a. 200 0.6
...J ...J ~
I I
.00 I I .00 150
I i I 0.4
-120 ~ I I 120 100
-160 I l 160
50 0.2
-5000 -3000 1000 1000 3000 5000 5000 .300() -1000 1000 3000 5000 0
0.0
Average Axial Strain (p) Average Axial Strain {)1) -900 -600 -300 0 300 -900 -600 -300 0 300
(c) CL-S-TH IO-PIO (d) CL-S-TH22-P l 0 Average Strain (p) Average Strain (p)
Incremental Bo Force (kN) Incremental Bolt Force (kN) (i) Axial Tensile Load-Bolt Strain Curves (ii) Bending Moment-Bolt Strain Curves
-50 .4() 30 20 10 0 10 20 30 40 50 50 .4() 30 -20 -10 0 10 20 30 40 50
160 160+-~~~~~~~r-~~~
(b) CL-S-THIO-PIO
120 120

80
500 r-:::;;;;;;;:;~::------, 1.4 . . . - - - - - - - - - - -- ----.
80
450
e
.
40 e
.
40 400 1.2
c 350
.2 0 1.0
-;
-300
~ .4()
z ~'fi. 0.8
~ 250
..8Q
a. 200 ~ 0.6 - Bolt01
120 -120

. 160 -t--~--..l.~-l-,..--1.-..,--~--!
5000 -3000 -1000 1000 3000 5000
Average Axial Strain (JJ)
-160 -t-~-,-.....,.1--,.~l-><~.1.,--,-~--1
-5000 3000 1000 1000 3000 5000
Average Axial Strain (JJ)
150
100
50
0.4

0.2
I
(e) CL-S-TH 10-Pl7 0 +-~~-r-~~-.---~~----~~ 0.0 ........_ _
-t--->..:>..o.....,~~r..-."" ~--~-l
(f) CL-S-TH22-P l 7
-900 -600 -300 0 300 -900 -600 -300 0 300
Average Strain (p) Average Strain (p)
Average Axial Strain (Side!)
.& Average Axial Strain (Side2) (i) Axial Tension Load-Bolt Strain Curves (ii) Bending Moment-Bolt Strain Curves
0 Incremental Bolt Force
(c) CL-S-THIO-Pl7
Fig. 6.18 Axial Strain Distribution of the Tube and Strain Distribution of the Bolts (Yielding Stage) Fig. 6.19 Load-Bolt Strain Curves (CL Test) (continued)
(CL Test)

192 193
 1.4 -r-------------------------::B:-o-::-
lt -:
0~9
1.4 .,...--------------------------,
Flange Plate 12 - Bolt01
)
1.0

~"ii 0.8
vv Flange Plate 1.2

1.0
- - Bolt09 - Bolt01

-
~ 0.6
~~ 0.8

-
~ 0.6
0.4
0.4
High Strength Bolt 0.2
Tensile Side 0.2
Tensile Side
0.0 +--~-----r-~------.-~~--?~-...---1
-900 -300 0 300 0.0 +-..............~......,-....................r--.~-.-.-~......~~
Bolt 0 1 -t:r- Bolt 06 -600
__._ BoltOI -t::r- Bolt 06 -1500 -1000 -500 0
-o- Bolt02 - - Bolt07 Average Strain (p) -o- Bolt02
500 1000 1500
~ Bolt08 - . - - Bolt07
----- Bolt 03 Bending Strain (/1)
Bolt 03 ---'V- Bolt 08
-o--- Bolt 04 Bolt 09 Bending Moment-Bolt Strain Curves
-o--- Bolt 04
Bolt 05 -<>- Bolt 10 ---+-- Bolt 09 Bending Moment-Bending Strain Curves
(d) CL-S-TH22-POO -a- Bolt05 -o- Bolt 10
(a) CL-S-THIO-POO

500 .-----------------------~ 1.4

500 -r--------------------------, 1.4
450 - 8olt01 450
1.2
400 400 1.2
350 1.0
350 1.0
-
z 300 ~~ 0.8 - 300
~
a..
250
200
-
~ 0.6
z
6 250
a.. 200
~"ti. 0.8

~ 0.6
150 0.4 150
0.4
100 100
0.2
50 50 0.2
0 +-~~-r~~--.-~~~~~-; 0.0 0 ~~~~~~~~~~~~~~
-300 300 0.0
-900 -600 -300 0 300 -900 -600 0 . -1500 -1000 -500 0 500 1000 1500 -1500 -1000 -500 0 500 1000 1500
Average Strain (p) Average Strain (/1)
Bendig Strain (/1) Bending Strain (/1)
(i) Axial Tensile Load-Bolt Strain Curves (ii) Bending Moment-Bolt Strain Curves
(i) Axial Tensile Load-Bending Strain Curves (ii) Bending Moment-Bending Strain Curves
(e) CL-S-TH22-Pl0
(b) CL-S-THIO-PIO

500 1.4
450 - Bolt09 5oo r-:::~:;;;;;;:;;;~:;:;::::::::::=:J 1.4
1.2 - Bolt01 450
400 1.2
400
350 1.0
350 1.0
- 300 ~"ii 0.8
z
~
a..
250
200
-
~ 0.6
-
z
6
a..
300
250
200
:l
~
0.8
0.6
150 0.4 150 0.4
100
0.2 100
50 0.2
50
0 0.0
0 ~~~~~~~~~~~~...........~ 0.0
-900 -600 -300 0 300 -900 -600 -300 0 300
-1500 -1000 -500 0 500 1000 1500 -1500 -1000 -500 0 500 1000 1500
Average Strain (/1) Average Strain (p)
Bending Strain (p) Bending Strain (p)
(i) Axial Tensile Load-Bolt Strain Curves (ii) Bending Moment-Bolt Strain Curves
(i) Axial Tensile Load-Bending Strain Curves (ii) Bending Moment-Bending Strain Curves
(f) CL-S-TH22-P 17
(c) CL-S-TH10-PI7
Fig. 6. I9 Load-Bolt Strain Curves (CL Test)
Fig. 6.20 Load-Bending Strain of the Bolt Curves (CL Test) (continued)

191
195
 1.4
- Bolt01 - Bolt09
Flange Plate 1.2 /
1.0 ... .I
Flange Plate Compression
-:i."t;. 0.8
-
-:i. 0.6

0.4
High Strength Bolt 0.2
Tensile Side
0.0 -h-....-.-.....-~~...-.-....-.-~..--.-,.-,-....-.-.....,-~...-.-;
Bo lt 0 1 - f r - Bolt06 -1500 -1000 -500 0 500 1000 1500
-o- Bolt 02 - "- Bolt 07 Bending Strain (p) Bolts Tension
Bolt 03 -Q- Bolt 08
-o- Bolt04 Bolt 09 Bending Moment-Bending Strain Curves
-----A- Bolt 05 -o-- Bolt 10
(d) CL-S-TH22-POO Fig. 6.21 Hypothesis used in tl1e Current Design

500 ~-------------------------, 1.4 -,-----------------------------..,

450
1.2
400 - Bolt01
350 1.0
- 300 -:i."t;. 0.8
z
~ 250
-:i. 0.6
Fixed Edge
Q. 200
150 0.4
100
0.2
50
o ~~-r~~~~~~~~~~~ 0.0 -j-..~...,..-~--.....-.-....-.-~~--...~--r~~
-1500 -1000 -500 0 500 1000 1500 -1500 -1000 -500 0 500 1000 1500
Bending Strain {.u) Bending Strain (p)
(i) Axial Tensile Load-Bending Strain Curves (ii) Bending Moment-Bending Strain Curves
(e) CL-S-TH22-POO

500 -r----------------------------, 1.4

450
400
1.2
1/.. l
7.'
Fig. 6.22 Model for Computation of Working Stress at tlle Flange Plate (Current Design)

350 1.0
- - - Bolt09
-300
z -:i.~ 0.8
~
Q.
250
200
r
-
-:i. 0.6
150 0.4
100
0.2
50
0 ~~-r~-.~~~--T-~~~~ 0.0
-1500 -1000 -500 0 500 1000 1500 -1500 -1000 -500 0 500 1000 1500
Bending Strain {.u) Bending Strain {.u)
(i) Axial Tensile Load-Bending Strain Curves (ii) Bending Moment-Bending Strain Curves
(f) CL-S-TH22-PI7
Fig. 6.20 Load-Bending Strain of the Bolt Curves (CL Test)

196 197
 Flange Plate Effective Area at Compressive Side

I-Iigh Strength Bolt

-- --I

M
_-_---if_-_--_-_-_--'~

Tube Effective Area at Tensile Side I

- ------ _i ~ ,'
I
L I
<
I
\
(a) Step 1 (Effective Cross Sectional Area) I
I
I
I
- _\_ I '
I
I
- - '-.....
a
Axis of Synunetry ~....---t -- ---- --
Circular Plate I

(Flange Plate),.__... '-------

(c) Step 2

r------------------4~ ft
: - - - - - -- - -- -rra-- - - - - + - t t
Bolt
1 - - - - - - - - - - - - - - .,
r---------------~
~------------------+ I

R '
Neutral Axis I

R : Diameter of Circular Prate

I '
t : Thickness of Circular Plate t I \
a I
I

(b) Step I (Evaluation Model of Effective Compressive Cross Sectional Area) p Tensile Force
I
\

Fig. 6.23 Asswnption of Proposed Design Procedure (Continued) Tube

(d) Step 3

Fig. 6.23 Assumption of Proposed Design Procedure

198 199
 r------------------------
I I
I I
I
I
I
I
I
I
I
tube I
I
I
...... ,. .
.-- ---
I
I

.- -- I
I
I
I
I
I
I
I
I
I
I
I

l(objcct
: photo
I I
I I
-------------------------~
(a)Loading Controller

(b) BL-S-TH16 (tr 16mm)

.; - I.'!

Photo 6.3 Local Deformation after Loading

(b)Display

Photo 6.2 Loading Contrler System

202 20:3
 Chapter 7
Conclusions
This study deals with the high strength bolted tensile joints which have been considered to be superior in
mechamcal behavior such as !ugh stiffness, high fatigue resistance and easiness of construction compared
w1th otl1cr connections. In thts study, in order to establish the rational design procedure of such joints, the
local and global behavior of the joints under monotonic and cyclic loading are investigated experimentally
and analytically paying attention to the mechanical behavior of structural elements such as high strength
bolts, the high strength bolt and its adjacent flange plate, split tee flange joints. As an application of the
tensile flange joints to pipe structures, tube flange joints are studied, where the assessment of global joints
behav10rs related to local behavior of the joint elements such as bolts, the flange plate and the tube is
made. TI1cn, the rational design procedure of high strength bolted tube flange joints subjected to combined
loading was proposed.
Chapter 2 focuses on the high strength bolt which is a basic structural element of the bolted
joints The mechanical behavtor of ilie high strength bolts under both monotonic and cyclic loading are
investigated experimentally and analytically. It is found that significant plastic defom1ation takes place in
bolt tluead and that load-displacement curve of bolts depends on relative length of the bolt thread to the
total bolt length. Therefore, the utilization of the bolt with long bolt wead should be prohibited from
viewpoint of strength, but should be recommended from viewpoint of ductility. In addition, the simple
modeling of high strength bolts is proposed based on the effective load-elongation relation of each bolt
section such as bolt shank and bolt tluead; then, its applicability is investigated. Proposed modeling is
verified to be very useful to analyze the global 3-dimensional behavior of the joint system by finite
element analysis because of saving the effort on data preparation and computation time. However. it is
also found that the stress check in the design for yielding by using this simple modeling may not be
accurate. As for the fatigue strength of the high strength bolts, it is found from the results of cyclic loading
test and finite element stress concentration analysis tl1at the fatigue strengtl1 is significantly affected by the
bolt pre-stress force. The local yielding is observed at the bolt tluead and incomplete bolt thread.
However, it is found that the fatigue strength obtained from the loading test is considerable higher than
that spectfied in guideline of faugue design for steel structures. so it is understood that the guideline is
very conservative.
Chapter 3 takes into consideration t11e high strength bolt and its adjacent flange plate
element(BAF model) for tl1e mvestigation of the contact/separation behavior of flange plates near tl1e bolt
Here, tl1e mechanical behavior of BAF model is investigated expenruentally and anal}1ically. In particular,
tl1e tensile stiffness of BAF model is assessed against the external applied load. It is w1derstood that tl1e
mechanical behavior of BAF model depends on the diameter-to-tluckncss ratio of circular flange plates
204
and bolt pre-stress force. fn case of small diameter-to-thtcl-ncss ratio. the tntllal stiffness is \Cry htgh and
the failure is caused by the bolt yteldmg. On the other hand, in case of large diameter-to-llltcl-ncss ratio,
the initiaJ stiffness is not high and the failure is caused by tl1e flange plate yielding and tiS energy
absorption capacity may be found to be higher. As for the effect of the bolt pre-stress force, it is found
that as the bolt pre-stress force given to the bolt increases, tl1e stiffness becomes large. Moreover, it is
observed tllat the mcrcase of the bolt force occurs at the carl) stage of loadmg tn case of thmner ctrcular
plate. This increase is considered to be caused by the local defonnation of the circular plate beneath the
bolt head, which is not so called prying action. In addition, tl1e simple evaluation fonnula for tl1e stiffness
to tile external applied load are proposed and coefficients in this fonnula are detennined by the multiple
regression anaJysis using non-linear least square metl1od. It is concluded tllat tl1e proposed cquatton for
tl1e evaluation of tile stiffness is applicable for the wide range of structural parameters of BAF model.
Chapter 4 studies the split tee flange joints which is the simplest and typical high strcngtJ1
bolted tensile joints. Its mechanical behavior under monotonic and cyclic loading is investigated both by
experimental and analytical approach. Especially, attention is paid to the contact/separation behavior, the
joint stiffness and local beha,,ior of the bolts and the flange plate. Under the monotonic loading, the
mechanical behavior is dominated by tl1at of eitl1er tl1e flange plate or the high strengtl1 bolt in accordance
with tllC thickness of the flange plate. ln case of tl1e thinner flange plate, tJ1c behavior of flange plate is
dominant and the high defonnabilty is obtained. On the other hand, in case of thicker flange plate, the
behavior of tile high strengtll bolt is dominant and tl1e high load carrying capacity is obtained, but the
failure becomes very brittle. From the "iew point of the energy absorption capacity, the use of thinner
flange plate is considered to be desirable. Moreover, it is also found that the joint stiffness becomes large
if the thicker and the narrower tl1e flange plate is used and that tl1e effect of the bolt pre-stress force on the
stiffness is observed to be significant. Furtllennore, tile significant increase of tl1e bolt force is observed
as the externally applied load increases and it is considered to be caused by tl1e local pulJ-up of the bolt by
the local defom1ation of tl1e flange plate, which is not like a prying force effect. Under the cyclic loading,
it is found that the mechanical behavior also depends on the thickness of the flange plate. ln case of
thitUler flange plate, the joint may fail by cracking at the welding section; on the other hand, in case of
thicker flange plate, tl1e joint may fail by breaking at the bolt tl1read. The amplitude of the bolt force under
cyclic loading is small compared with l11at of extemaJ applied load. It is also found tl1at tl1e fatigue
strengtll relative to the yield strength of tl1e joint is high in case of tlunner flange plate. TI1creforc, tl1e
ultimate state may be defined by the strength under the static loading. However, tl1e joint with thicker
flange plate may be required to be checked against fatigue in the design. Moreover, tl1e simple evaluation
method of tile fattgue strengtll is proposed m conjunction with the guideline of tile fatigue design for steel
structures and tl1e experimental results under monotonic loading, and its appltcabtlit) is proven.
Chapter 5 discusses the development of the quasi-2-dimensional analysis on split tee flange
joints considering the load-separation behavior. Proposed is the ciTecti\'e width coefficients of the bolt and
205
the flange plate ut11t.r.ed for 2-dimensional analysts in order to reproduce 3-dimensiOnal behavior. This
analysis is very useful for parametric study of the split tee flange joint or its design because of its
simplicity compared with 3-dimensional analysis. It is found that 3-dimensional behavior of the split tee
flange joint such as load-deformation curve and characteristics of local defonnatlon can be reproduced by
quasi-2-dimensional fimte element analysis using the effective width of the bolt and the flange plate.
However, the stress obtamed by this quasi-2-dimenstonal analysis is found to be very conservative, so
that accurate stress check may be difficult by the proposed 2-dimensional finite element analysis using
ciTcctivc width coefficients.
Chapter 6 makes an assessment of the tube flange joint as an application of high strength bolted
tensile flange joints. The mechanical behavior of high strength bolted tube flange joints is studied
experimentally in order to develop the practical design procedure. The fundamental characteristics such as
the load transferring mechanics in both compresstve and tensile sides, failure mechanism under the
combined loadmg arc dtscussed. It is found from the experimental results that the initial stiffness, the
yield strength and the ultimate strength can increase significantly by preventing the interaction of the
flange plate with the steel tube as the flange plates become thicker. However, the further increase of the
ultimate strength may be limited by the buckling of the tube. [n addition, the simple analytical model for
the stress check based on the load transferring mechanism both in the compressive and tensile sides are
proposed. It is understood that the initial yielding of the flange plate under bending is well-predicted by
this proposed procedure considering the effective cross sectional area. Therefore, the rational design of
tube flange jomts can be carried out by using the proposed procedure.
As mentioned above, the high strength bolted tensile flange joints can be utilized for primary
members of bridge structures. However, in order to make the proposed design procedure more rational
and reliable, further study to solve the following technical problems should be carried out.
I) The stgnificant increase of the bolt force is observed in case of thinner flange plate. Tius
mechantsm IS qwte different from the mechanism considered in the past study. Therefore, the
structural detatl to reduce this increase of the bolt force should be developed.
2) Generally spcakmg, the joints are made of numerous bolts. Ln order to design the economical
joints, the contribution of a bolt among the bolt group on the mechanical behavior should be
investigated.
3) It is concluded from the experimental results that the tensile joints with thinner flange plate have
high energy absorption capacity. Further study should be carried out in order to use the tensile
joints as energy absorption device from the view point of earthquake resistance.
4) l11e cycl1c loadmg tests for high strength bolts and split tee flange JOints were carried out for the
Limited number of test specimens. In order to develop the reliable S-N curve for the fatigue
strength evaluation, further cyclic loading tests should be required, particularly for the tube
flange joints.
206
Research Activities
l) T.Yamaguehi , E. Watanabe, K.Sug1ura, S.Kasai : Experimental Study on High Strength Bolted Tube
Flange Joints, Proc. of Annual Conference of Civil Engineers, JSCE Kansai Chapter, Jun. 199 1(in
Japanese).
2) T.Yamaguchi, E.Watanabe, K.Sugiura, S.Kasai : Design Method of High Strength Bolted Tube
Flange Joints, Proc. of the 46th Annual conference of JSCE, Sep. L99l (in Japanese).
3) E.Watanabe, K.Sugiura, T.Yamaguchi, S. Kasai : Design Method of High Strength Bolted Steel
Tube Flange Joints, Journal of Structural Engineering, JSCE, Vol. 38A, Mar. 1992, pp. 1-12(in
Japanese).
4) T.Yamaguehi, E. Watanabe, K.Sugiura, S.Kasai, T.Mitamura Experimental Study on H1gh Strength
Bolted Tube Flange Joints subjected to Bending, Proc. of the 47th Annual conference of JSCE, Sep.
1992(in Japanese).
5) E.Watanabe, K.Sugiura, T.Yamaguehi, S.Kasai : Behavior of High Strength Bolted Tube Flange
Joints subjected to Bending, Proc. of the 3rd Pacific Structural Steel Conference, Oct. 1992, pp.
391-398.
6 E.Watanabe, K.Sugiura, T. Yamaguchi , S.Kasai : Rigidity of High Strength Bolted Steel Flange
Joints, Journal of Construction Steel, Vol. l , JSSC, July 1993, pp. 3 1-38(m Japanese).
7) E.Watanabe, K.Sugiura, T.Utsunomiya, T.Yamaguchi, S.Kasai : Simple Analysis of High Strength
Bolted Tube Flange Joints, Proc. of the 4th East Asia-Pacific Conference on Structural Engineering
and Construction, Sep. 1993, pp. 861-866.
8) T.Yamaguchi, E.Watanabe, K.Sugiura, S.Kasai : Joint Rigidity of High Strength Bolted Tube
Flange Joints, Proc. of the 48th Annual conference of JSCE, Sep. 1993(in Japanese).
9) Opening Characteristics of High Strength Bolted Flange Joints and its Tensile Rigidity, Journal of
Structural Engineering, JSCE, Vol. 40A, Mar. 1994, pp. 153-l62(in Japanese).
10) T.Yamaguchi, E.Watanabe, K.Sugiura. S.Kasai, T.Mitamura : 30 FEM Analys1s on
Contact/Operation Behavior of Split Tee Flange Joints, Proc. of the 49th Annual conference of JSCE,
Sep. 1994(in Japanese).
11) E.Watanabe, K.Sugiura, T.Yamaguchi, S.Kasai : Contact/Opening Behavior of Split Tee Flange
Joints and its tensile rigidity, Journal of Construction Steel, Vol 2, JSSC. Nov. 1994, pp. 93-LOO(in
Japanese).
12) l11e Assessment of 2-dimensional Analysis on Split Tee Flange Joints. Journal of Structural
Engineering, JSCE. Vol. -llA, Mar. 1995, pp. 95-l 02(in Japanese)
13) E.Watanabe, K.Sugiura. T. Utsunomiya, T. Yamaguchi, S. Kasai ct al. : Fatigue Strength of High
Strength Bolted Tensile Joints, Proc. of the 5th East Asia-Pacific Conference on Structural
107
 Enginecnng and Construction, July 1995, pp.
14) T.Yamaguchi, E. Watanabe, K.Sugiura, S.Kasai, et al. : Fatigue Strength on Split Tee Flange
Tensile Jotnts, Proc. of the 50th Annual conference of JSCE, Sep. l995(in Japanese).
15) K.Fujitani, E.Watanabe, K.Sugiura, T.Yamaguchi, S.Kasai : Effective stress-strain relation of
high strength bolts considering threads, Journal of Construction Steel, Vol. 3, JSSC, Nov. 1995, pp.
28l -2 88(in Japanese).

208