Behavior and Modeling of Semi - Rigid Steel Beam
Behavior and Modeling of Semi - Rigid Steel Beam
Behavior and Modeling of Semi - Rigid Steel Beam
By Temesgen Wondimu
A Thesis Submitted to the Graduate School of the Addis Ababa University in Partial Fulfillment of the Requirements for the Degree of Master of Science in Civil Engineering
By Temesgen Wondimu
A Thesis Submitted to the Graduate School of the Addis Ababa University in Partial Fulfillment of the Requirements for the Degree of Master of Science in Civil Engineering Approved by Board of Examiners
Dr. Sjifferaw Taye Advisor Dr.-Ing. Bedilu Habte External Examiner ____________ Signature ____________ Signature ____________ Date _____________ Date
ACKNOWLEDGMENT
I gratefully acknowledge my supervisor Dr. Shifferaw Taye for his continuous assistance through the entire work. His advice on technical as well as non technical matters is greatly appreciated. More over, I would like to acknowledge the financial support of the Office of Research and Graduate studies, Addis Ababa University. I would also like to thank my family, especially my wife Meskerem, for their continuous encouragement and assistance during the thesis work. The continual material and moral supports of all my friends is also thankfully appreciated.
TABLE OF CONTENTS
LIST OF FIGURES LIST OF TABLES LIST OF APPENDICES LIST OF SYMBOLS ABSTRACT CHAPTER 1. INTRODUCTION 1.1 Introduction 1.2 Scope and Objective of the Thesis
v vi vii viii x 1 3
CHAPTER 2. MOMENT ROTATION RELATIONSHIP OF SEMI RIGID CONNECTIONS 2.1.Connection Classification 2.2.Modeling of Semi Rigid Connections 2.2.1. Based on Available Experimental Test Data 2.2.2. Finite Element Modeling of Semi Rigid Connections 2.2.3. An Effective Component Based Method of Modeling Semi Rigid Connections 2.2.4. Mechanical Semi Rigid Joint Modeling 25 4 7 7 13 17
CHAPTER 3. BEHAVIOR AND MODELING OF SEMI RIGID FRAMES 3.1.Introduction 3.2.Proposed Simplified Effective Component Method 3.3.Modeling of Semi Rigid Frame for Static analysis 3.3.1. The Investigated Structure 3.3.2. Finite Element Analysis of the Frame 3.4.Dynamic Analysis and Response of Semi Rigid Frames 3.4.1. Introduction 3.4.2. Method of Analysis and Analysis Results 27 28 31 31 33 37 37 37
3.4.2.1.Modal Analysis 3.4.2.2.Spectral Analysis 3.4.2.2.1. Single Point Response Spectrum 3.4.2.2.2. Analysis Results
38 41 42 49
APPENDIX A: ANSYS MATRIX 27 ELEMENT APPENDIXB: STIFFNESS COEFFICIENTS OF VARIOUS COMPONENTS OF THE END PLATE JOINT APPENDIX C: REAL CONSTANTS / STIFFNESS COEFFICIENTS/ FOR THE JOINTS REFERENCES
LIST OF FIGURES Chapter Fig 2 Fig. 2.1 Fig. 2.2 Fig.2.3 Fig.2.4 Fig.2.5 Fig.2.6 Fig.2.7 Fig.2.8 Fig.2.9 Fig.2.10 Fig.2.11 Fig.2.12
3
Fig.3.1 Fig.3.2 Fig.3.3 Fig.3.4 Fig.3.5 Fig.3.6 Fig.3.7 Fig.3.8 Fig. A.1
Name Classification of connections according to Bjorhorde et al Semi rigid connection types and size parameters Three parameter power model Moment rotation relationship for a fourparameter model of semi rigid connections Finite Element modeling semi-rigid connection Beam member with rotational springs General Models for steel & composite joints Spring model of the semi rigid joint Degrees of freedom of the base element and complete element Deformation modes corresponding to shear, compression and bending of the panel zone Degrees of freedom of the interior joint Mechanical semi rigid joint model proposed by Tschemmernegg et.al Position and simplified effective modeling of semi- rigid connections Stiffness derivation process for node I of the basic connection element Reference frame The investigated frame Static analysis results of semi rigid and rigid frames The first four modal shape of semi rigid frame The first four modal shape of rigid frame Moment and shear diagram of semi rigid and rigid frames subjected to an earthquake respectively MATRIX27 Schematic
Page 6 8 11 12 13 13 19 19 21 22 24 26 28 29 31 32 36 51 52 55 58
LIST OF TABLES Chapter Table 2 Table 2.1 Name Curve fitting constants and standardization constants for the Frye-Morris polynomial model
3
Page 9 32 33 34 43 50 53
Table 3.1
Comparison of results obtained by the proposed model, EC3 and the model of Bayo et al. Characteristics of the semi rigid connection used for the investigated frame Comparisons of moments at the joints and beam mid spans for semi rigid and rigid connections Types of spectrum loading Modal analysis results, natural period and frequency of rigid and semi rigid frames Spectral analysis results: story shear, drift, joint, moment and span moment
Table 3.2
Table 3.3
Table 3.6
LIST OF APPENDICES APPENDIX A APPENDIX B ANSYS MATRIX27 ELEMENT STIFFNESS COEFFICIENTS OF VARIOUS COMPONENTS OF THE END PLATE JOINT APPENDIX C REAL CONSTANTS / STIFFNESS COEFFICIENTS/ FOR THE JOINTS 66 58 63
LIST OF SYMBOLS
relative deformation angle of the connection coupling coefficient damping ratio interaction factor mode shape vector natural frequency non dimensional connection rotation plastic rotation capacity of the connection participation factor for mode i the ith mode shape
nodal displacements in the x, y and z directions respectively scaling factor for the purpose of numerical stability non-dimensional parameters used in the classification of connections vector describing the excitation direction mass matrix mode coefficient for the ith mode correction matrix accounting for semi rigidity effects width of the column beam depth to which the connection is attached. curve fitting parameter Youngs modulus lever arm Heavisides step function moment of inertia of the beam total stiffness matrix for a semi-rigidly connected member elastic element stiffness matrix of the original fully rigidly connected member:
ij
p
i i
u, w,
m
[D] [M] Ai C d d Dk E h H[ ] Ib K K0
stiffness of the panel zone (column web) under shear compression stiffness of the panel zone (column web) under shear resultant of the stiffness coming from column- web in tension, column right flange in bending, end plate in bending, angles, bolts in tension and reinforcement in case of composite structures
geometrical stiffness matrix resultant initial stiffness of all components bending moment full plastic moment capacity of the beam shape parameter total number of expanded modes fixity factor total modal response instantaneous connection stiffness initial connection stiffness connection strain-hardening stiffness transformation matrix
10
ABSTRACT
In most steel frame designs the beam to column connections are assumed to be rigid or pinned. Rigid joints, where no relative rotation occurs between the connected members, transfer not only substantial bending moments, but also shear and axial forces. On the other extreme, pinned joints are characterized by almost free rotation movement between the connected elements that prevent the transmission of bending moments. Despite these facts, the great majority of joints doesnt exhibit such idealized behavior. A substantial effort has been made in recent years to characterize the behavior of semi rigid connections. Most design codes included methods and formulas to determine both their resistance and stiffness. EC3 and EC4, for instance, allow the use of springs attached to the end of the beams at both sides of the joints. In order to account for the panel shear deformation the code allows the use of interaction parameter, called the factor. But since the definition of the factor implies an approximation of internal forces at the joint, it requires an iterative process at the time of the global analysis of the structure. In order to avoid this iterative process various researchers proposed new elements accounting for various deformation modes of the connections. Although the new proposed element by Bayo et.al accurately characterizes the behavior of semi-rigid connections based on the EC component method it is not appropriate to use in commonly available software. In this thesis a new component-based two-node-connection element is proposed. By using the proposed two-node element the static and dynamic response behaviors of a semi-rigid frame as compared to a rigid frame of the same geometry and cross section are studied using a general purpose finite element package ANSYS. The study showed that connection flexibility tends to increase vibration periods and hence reduces the internal stresses due to a given earthquake action in the frame elements as compared to those in rigid frames.
11
CHAPTER ONE
1.1. INTRODUCTION
Traditionally, steel frame design assumes that beam-to-column joints are rigid or pinned. Rigid joints, where no relative rotations occur between the connected members, transfer not only substantial bending moments, but also shear and axial forces. On the other extreme, pinned joints are characterized by almost free rotation movement between the connected elements that prevents the transmission of bending moments. Despite these facts, it is largely recognized that the great majority of joints doesnt exhibit such idealized behavior. This is explained by the fact that in semi rigid frames the internal force distributions, lateral displacement magnitudes, collapse modes are functions of joint flexibility. Extensive studies have been carried out over the past twenty five years to estimate the actual behavior of such joints. Innumerable studies have been produced on composite and steel semi-rigid connections, covering the state of the art [1], numerical studies and experimental tests [2 - 8]. The fundamental results of these investigations led to code specifications that provided structural engineers with adequate procedures to evaluate the moment rotation characteristics of semi rigid connections. A good example of this new design trend is available in Eurocode3 [17]. Despite the substantial increase in structural design knowledge, the semi-rigid connection design is still facing resistance from structural engineers. This is explained by: 1. Lack of detailed information on the advantages of semi-rigid design philosophy. Semi-rigid frame design has many advantages. These include:
a. Economy
Investigations showed that four a four story semi-rigid floor system up to 15% economy in terms of steel weight was reached even when compared to the traditional most
12
economical solution. As these structures are used in large scale construction the achieved 15% weight reduction can generate an even greater economy [10].
b. Ductility
The components of partial strength joints can deform in a ductile manner in the case of strong earthquake. In the common case where beam sizes are governed by drift or alike design criteria, rather than flexural strength requirements, ductile partial strength connections allow the formation of a desirable beam hinging global frame mechanism. The possibility of using partial strength connections as the main energy dissipative mechanism for the seismic resistance of frames has now been recognized in modern design codes [12, 13, and 19]. 2. Complexity and lack of effective tools for global analysis of semi-rigid frames The non linear characteristics of a steel beam-to-column connection play a very important role for frame global analysis. Although modern structural steel and composite codes [Eurocode3 &4] include procedures and formulas to define both the stiffness and resistance of semi rigid connections there is no explicit method to model and analyze frames with these connections. 3. Lack of sufficient experimental data The design and construction of steel, as compared to others, is not yet at its outset in our country which leads to lack of experience. Besides the behavior of semi rigid connections is very dependent on experimental investigations which is not possible in our country due to the cost associated. Hence structural engineers usually face the problem of experimentally validated connection data bases which, in consequence, hinders the planning of structural steel frames of semi rigid nature.
13
14
CHAPTER TWO
MOMENTROTATION RELATIONSHIP OF SEMI-RIGID CONNECTIONS
15
3. Type III or semi rigid framing. This construction assumes that the connections
can transfer vertical shear and also have adequate stiffness and capacity to transfer some moment. The AISC-LRFD specifications (1986) designate two types of constructions in their provisions: Type FR (fully restrained) and Type PR (partially restrained). Type FR corresponds to EBSC-3 Type I and Type PR includes EBCS-3 Type 2 and 3. To be pertinent, the rigidity of the connection should be defined with respect to the rigidity of the connecting member. For general application to a wide range of beam-to-column connection, Bjorhorde et al. [2] introduced a non-dimensional system of classification that compares the connection stiffness to the beam stiffness. In defining the beam stiffness, a reference beam length 5d is used, where d is the beam depth to which the connection is attached. The non-dimensional parameters used in the classification of connections are:
m= where:
M , = r Mp p
(2.1)
p =
Mp (EI b /5d)
M p - full plastic moment capacity of the beam. The classification is based on the strength and stiffness of the connections with the boundary regions shown in Fig 2.1. The different regions in Fig 2.1 are defined as: 1. Rigid connection
0.7 m 2.5 in terms of strength (2.2) in terms of stiffness
16
(2.3)
3. Flexible connection
0.2 m 0.5
Bjorhorde et al. [2] also have proposed an expression for calculating the required rotation capacity of the connection based on a reference beam length and by curve fitting with test data. This simplified expression is written as:
5.4 2 m= 3
(2.5)
According to this equation, the required rotation capacity of the beam-to-column connection depends on the ratio of the ultimate moment capacity of the connection to the fully plastic moment of the beam, and it is inversely proportional to the initial connection stiffness. In other words the smaller the initial connection stiffness, the larger the necessary rotation capacity. Rotation capacity
1.0
5.4 2 m= 3
Rigid
0.8 0.6
m
0.4
Semi rigid
0.2
Flexible
0 0.4 0.8 1.2 1.6 2.0 2.4 2.7
17
Although relatively better investigations were made on bare steel connections most standards give little guidance on the property and design of composite connections. Composite connections can be defined as a connection between a composite member and any other member in which reinforcement is intended to contribute to the resistance of the connection. This chapter is devoted to previous researches undertaken in order to provide a complete set of behavior, analysis and design rules of semi- rigid connections.
2.2. MODELING OF SEMI REGID CONNECTIONS 2.2.1 Based on Available Experimental Test Data
The most commonly used approach to describe M- r curve is to curve fit the experimental data with simple expressions. Several analytical models have been developed to represent connection flexibility using available experimental test data. Early models used the initial connection stiffness as the key parameter in a linear M- r model [1, 2]. Although the linear model is very easy to use, it has a serious disadvantage. It is suitable for only a small range of the initial relative rotation. A closer approximation of true connection behavior can be obtained by using either a bilinear model or a piecewise linear model. In these models, the abrupt changes in connection stiffness at the transition points make their practical use difficult. Jones et al. [1] proposed a cubic -B- spline model to obtain a more suitable function. However, this model requires a large number of sampling data during the formulation process. Frye and Morris [2] have reported a polynomial model to evaluate the behavior of the several types of connections. In this model, the M- r behavior is represented by an odd power polynomial. Some of the importance of connection models will be discussed in more detail in the following section.
a. Frye- Morris polynomial Model
The most popular model for structural analysis is the polynomial function proposed by Frye and Morris (1975). The Frye-Morris model was developed based on a procedure formulated
18
by Sommer (1969). They used the method of least square to determine the constants of the polynomial. This model has the general form shown in Eq. (2.9).
r = C1 ( KM )1 + C 2 ( KM ) 3 + C 3 ( KM ) 5
and C1, C2, and C3 are curve fitting constants.
(2.9)
where K is the standardization parameter dependent upon the connection type and geometry,
This model represents the M- r behavior reasonably well. The main drawback is that the nature of the polynomial is to peak and trough within a certain range. Then, the first derivative of this function, which indicated the tangent and connection stiffness, may become negative at some value of connection moment M. This is physically unacceptable. This negative stiffness makes structural analysis difficult if the analysis scheme with tangent connection stiffness is used. Relative moment-rotation curves of extensively used semi-rigid connections are shown in Fig 2.2 The curve fitting constants C1, C2, and C3 connection type are summarized in Table 2.1 and the standardization constant K for each
19
Table 2.1: Curve fitting Constants and Standardization constants for the Frye-Morris Polynomial Model
The Chen-Lui exponential Model has been refined to accommodate linear components by Kishi and Chen [2] and is referred to herein as the modified exponential model. This model is represented by a function of the following form
j=1 r k =1 M = M o + C j 1 exp 2j + D k ( r k )H[ r k m n
(2.10)
where: Mo
the initial connection moment scaling factor for the purpose of numerical stability curve fitting parameters, the starting rotation of the kth linear component given from the experimental M- r curve
Cj, Dk
k
H[ ]
20
Using the linear interpolation technique for the original M- r data, the weight function for each M- r datum is nearly equal. The constants Cj and Dk for the exponential and linear terms of the function are determined by a linear regression analysis. The instantaneous connection stiffness Rk at an arbitrary relative rotation r can be evaluated by differentiating Eq.(2.10) with respect to r . When the connection is loaded, we have
R k = R kt = dM d r =
r = r m j=1
r k =1 + D k H[ r k exp 2j 2j n Cj
(2.11)
Cj 2 j
+ Dk H [ k
k =1
(2.12)
This model has the following merits: the formulation is relatively simple it can deal with connection loading and unloading for the full range of relative rotation in a second order structural analysis with secant connection stiffness the abrupt changing of the connection stiffness among the sampling data is only general from inherent experimental characteristics.
21
2K
n = 1.7472 + 0.0012 K Mp
0.012
0.014 0.016
Mp/K
(a) Generalized moment rotation curve Fig. 2.3. Three parameter power model The generalized form of this model is M= R ki r 1 + r 0
n
1/n
(2.13)
where Rki 0 n
Eq. 2.13 has the shape shown in Fig. 2.3. From this figure, it is recognized that the larger the power index n, the steeper the curve. The shape parameter n, can be determined by using the method of least squares for the differences between the predicted moments and the experimental test data
22
M=
R l r 1 + r 0 M Mu
n
1/n
+ R kp r
(2.14)
where: connection moment connection ultimate moment capacity Rkp connection strain-hardening stiffness
One of the salient merits of this model is that the associated tangent stiffness of a semi- rigid connection can be readily obtained through differentiation of Eq. 2.14 with respect to r :
Rt = Rl 1 + r 0
n
(n +1)/n
+ R kp
(2.15)
Rotation Fig.2.4 Moment rotation relationship for a four-parameter model of semi rigid connections
23
If the initial tangent stiffness, Rki, is specified as a single independent parameter, the other three parameters needed in Eq. (2.14) are determined by Rkp= 0.0282Rki n= shape factor determined from Fig.2.3 Mu=0.00338Rki which were provided by Hsieh and Deierlein [15] through a condensation process of collected experimental data in order to eliminate scale variations.
The finite element representation of semi rigidly connected structural members is proposed by Xu [15]. The member is assumed to be consisting of the originally fully rigidly connected member and zero length spring elements (with spring constant equal to the appropriate secant stiffness of the semi-rigid connection it simulates) attached to each member end. R1 @ R2 @
E, I, L, A
Fig. 2.5 Finite Element modeling semi-rigid connections Consider the beam member shown below with the end springs. ra and rb represent the relative spring rotations of both ends and ka and kb expressed as: ka= Ma/ ra ; kb=Mb/ rb (2.18a) are the corresponding spring stiffness
MA P @ A
rA A B rB
L
MB @ B P
24
The relationship between end-moments and end-rotations of a beam can be written by replacing the end rotations A and B by ( A rA ) and ( B rB ) respectively, as follows [4]
MA EI 4 A kA L M + 2 B B k B
MA =
(2.18b)
MB =
M MB EI + 2 A A 4 B kA kB L
(2.18c)
The elastic element stiffness matrix of a semi-rigidly connected member, denoted by K , can then be formulated as [6] K = K oC member:
EA L 0 12EI L3 0 6EI L2 4EI L EA L 0 0 EA L symm 0 12EI L3 6EI L2 0 12EI L3 0 6EI L2 2EI L 0 6EI L2 4EI L
(2.19)
where Ko is the elastic element stiffness matrix of the original fully rigidly connected
(2.20)
with E, L, I, A as elasticity modulus, length, cross-sectional moment of inertia, and cross sectional area of a given member, and C, a correction matrix accounting for semi rigidity effects:
25
C=
1 0 0 0 0 0
0 2Lr1 (1 r2 ) 4 r1 r2 3r1 (2 r2 ) 4 r1 r2 0 0 0
0 0 0 1 0 0
rj =
1 , 3EI 1+ R jL
j = 1,2
(2.22)
where Rj is the appropriate secant stiffness, under current load level, of the semi-rigid connection at jth end. Physically speaking, the fixity factor indicates how stiff a semi- rigid connection is relative to the structural member to which it attaches. Two extreme cases are r=1 for fully rigid connections and r=0 for perfectly pinned connections. Obviously a realistic semi-rigid connection will have a value of 0<r<1. Note that the end moments need also to be modified due to the presence of semi rigid connections. For example, a structural member with semi-rigid connections at both ends and subjected to uniformly distributed transverse load of w has the following end moments [15]:
26
wL2 M1 = 12
(2.23)
A geometrical stiffness matrix derived by Xu [15], which is specifically suited for semi rigidly connected members, takes the following form:
where: 0 0 G 22 0 G 23 G 33 0 0 0 0 0 G 25 G 35 0 G 55 0 G 26 G 36 0 G 56 G 66
(2.24)
symm.
G 22 = G 23 G 33 G 36 G 26 G 66 G 55
2N 2 2 2 2 2 2 3r1 r2 + r1 r2 + r1 r2 + 8r1 + 8r2 34r1 r + 40 5L(4 r1 r2 ) 2 N 2 2 2 2 2 = r r 12r1 r2 + 16r1 r2 + 32r1 28r1 r2 2 1 2 10(4 r1 r2 ) 2NL 2 2 2 2 = 2r1 r2 7r1 r2 + 8r1 2 5(4 r1 r2 ) NL 2 2 2 2 7r1 r2 16r1 r2 16r1 r2 + 28r1 r2 = 2 10(4 r1 r2 ) N 2 2 2 2 2 r r + 16r1 r2 12r1 r2 + 32r2 28r1 r2 = 2 1 2 10(4 r1 r2 ) 2NL 2 2 2 2 2r1 r2 7r1 r2 + 8r2 = 5(4 r1 r2 ) 2 G 35 = G 23 , G 56 = G 26 = G 25 = G 22 ,
(2.25)
Note that the geometrical stiffness matrix, KG , for a fully rigidly connected member can be easily obtained by letting r1=r2=1;
27
0 KG =
non-linearity is then: K = K o C + KG
0 6N 5L
0 N 10 2NL 15
0 0 0 0
0 6N 5L
N 10 0 6N 5L
0 N 10 NL 30 0 N 10 2NL 15
(2.26)
The total stiffness matrix for a semi-rigidly connected member considering geometrical
(2.27)
The main problem associated with this modeling approach is unless an analysis program is developed or the stiffness matrices accounting semi - rigid behavior is incorporated in the software used for structural analysis it is not easy to use. Besides it will not account for the panel shear deformation.
28
EC3 and EC4 [17, 18] take in to account the deformation of the panel zone as one of the components of the joint. Further more, EC3 provides a sophisticated and elaborated tool based on the so called component method that allows one to obtain the strength and stiffness characteristics of all the different components of the joint, including the panel zone, shown in Fig. 2.7. The component method has been preferred over the finite element method due to the ability of the former to better model local phenomena [3]. Under EC3 and EC4 the different components (springs) are assembled to form a resultant elastic-plastic spring element that models the connections at each side of the joint to be attached at the end of each beam.
EC3 proposes a mechanical model for the semi- rigid joint in which each component is modeled by an equivalent linear spring. Fig 2.7 shows a condensed model of the right side connections in which Keq2 components: column- web in tension, column right flange in bending, end plate in bending, angles, bolts in tension and reinforcement in case of composite structures. The springs Kcws and Kcwc represent the stiffness of the panel zone (column web) under shear and compression respectively. These springs are assembled to form a single bi-linear (elastic-plastic) rotational spring that models the connection and is attached at the intersection between beam and column for global analysis. The bilinear model is characterized by an initial stiffness Kj,ini ( resultant of all the components) and moment resistance Mj,Rd which are defined by the following expressions:
K j,ini = h2 1 1 1 + + K cws K cwc K eq
(2.28)
M j,Rd = Flt,i,Rd h i where h is the lever arm and FLt,i,Rd is the resistance of each component[17].
(2.29)
29
Keq2
(b) Complete joint Fig. 2.8.Spring model of the right side and complete semi-rigid joint
30
where M1 and M2 are the bending moments acting at the left and right side of the joint, and Vc1 and Vc2, are the shear forces in the upper and lower columns, respectively. The expressions in brackets of Eqns (2.30a) and (b) constitute the interaction factors that are used in EC3 to define the shear forces acting on the panel zone. EC3 further simplifies Eq. 2.30 by omitting the beneficial effect of the shear forces Vc1 and Vc2 [17], and by limiting to positive values less than 2, as in the following equation:
1 = 1 + M2 M 2 and 2 = 1 + 1 2 M1 M2 ( 2.31)
These factors affect the values of the joint stiffness Kj,ini and moment resistance Mj,Rd. However, the bending moments M1 and M2 and shear forces Vc1 and Vc2 coming from different elements (beams and columns) are not known in advance, and are only known after a global analysis. Therefore, an initial guess of these values and a subsequent iteration procedure are needed to update the joint characteristics and obtain the precise final internal forces and moments.
b. Using Finite Dimensioned Four Node Elastic-Plastic Joint Element
E. Bayo et.al, 2006 [3] proposed an EC3 and EC4 component based finite dimensioned elastic-plastic 4-node joint element which takes into account the actual size of the joint, its deformation characteristics (components) including those of the panel zone, local phenomenon and all the internal forces that concur at the joint. As a consequence, this new element avoids the use of the -factor and the inherent initial guessing and iterative process
31
that it requires. Since the element is attached to the adjacent beams and columns the eccentricities of the internal forces coming from them are also taken into account.
Interior joint
The model of an interior joint starts from the scheme of the components shown in Fig. 2.8. The springs Keq1 and Keq2 model the left and right connection stiffness respectively. The springs Kcws and Kcwc represent the stiffness of the panel zone (column web) under shear and compression respectively, which are usually the most critical components of the joint. All these values may be obtained by means of formulas described in EC3 and EC4. Fig. 2.9 shows the new base element with 4 nodes and 9 degrees of freedom, proposed by Bayo. et.al, in which the rigid body modes have been restrained.
5 6 C 4 9 8 7
1 A B
2 3 2 h 3 1
5 4 6
8 7 9 d 12
11 10
(a) Base element Fig. 2.9 Degrees of freedom of the interior joint
The dimension d is equal to the width of the column, and the height h corresponds to the lever arm defined in EC3, Part 1.8 [17]. This base element will be used to obtain the basic flexibility and stiffness matrices. Afterwards, the contribution of the rigid body modes will be added to obtain the complete stiffness matrix in terms of the 12 degrees of freedom shown in Fig. 2.9b.
32
It should be noticed that all the forces and moments coming from the adjacent beams and columns concur at the joint at the points A, B, C and D. Therefore the complete force field in the panel zone is known with no need of the factor. Also, since the real dimensions of the joints are being considered the eccentric moments are automatically taken into account. In order to develop the flexibility matrix the following deformation modes are considered (some of these models are depicted in Fig 2.10):
Fig. 2.10. Deformation modes corresponding to shear, compression and bending of the panel zone.
Shear deformation of the panel zone Tension and compression of the panel zone Column flange and end-plate under bending Elasticity and configuration of the bolts Bending deformation of the column Axial deformation of the column
Due to the stiffening effect of the column flanges the segment AB (see Fig. 2.9) is supposed to be rigid under the shear forces coming from the beams. Thus, considering the 9 degrees of freedom of element of Fig 2.9 (left) and the deformation modes, the resulting non-zero elements of the upper triangular part of the flexibility matrix are derived and shown in Eq. 2.32. The bilinear behavior of the springs is defined according to the values given by EC3. Once the flexibility matrix has been obtained the basic stiffness matrix may be calculated through matrix inversion:
K b = F 1
33
Following standard procedures in matrix structural analysis [25] the rigid body modes may be added by means of a transformation matrix T, which is shown in Eq 2.33.
F11 =
1 h2
F13 =
1 h2
1 K cws
F14 =
1 1 2h K cws
F17 =
1 1 2h K cws
F22 =
F24 =
F27 =
F33 =
F47 =
F55 =
F66 =
h 2EI c h 2EA c
F77 =
F88 =
F99 =
where Ac and Ic represent the area and moment of inertia of the column respectively.
34
0 1 d 1 0 1 d T = 0 0 0 0 0 0 0
1 0 0 1 0 0 0 0 0 0 0 0
0 1 d 0 0 1 d 1 0 0 0 0 0 0
1 h 2d 0 0 h 2d 0 1 0 0 0 0 0
0 1 2 0 0 1 2 0 0 1 0 0 0 0
0 1 d 0 0 1 d 0 0 0 1 0 0 0
1 h 2d 0 0 h 2d 0 0 0 0 1 0 0
0 1 2 0 0 1 2 0 0 0 0 0 1 0
0 1 d 0 0 1 d 0 0 0 0 0 0 1
(2.33)
Thus, the basic stiffness Kb may be transformed to obtain the general stiffness matrix in terms of the 12 degrees of freedom (Fig 2.10 right) as follows:
K = TK b T T
(2.34)
External joints
The external and corner joints will only have 9 and 6 degrees of freedom, respectively, as shown in Fig. 2.11
5 2 6 1 3 5 6 4 2 1 3 9 7 10 4
35
Their stiffness matrices may be obtained from the 12 degrees of freedom stiffness matrix, K, by means of the static condensation algorithm that helps eliminating those degrees of freedom that are not necessary. The general 12 degrees of freedom stiffness matrix may be partitioned as follows:
K dd K= K ed K de K ee
(2.35)
where the sub-index d stands for desired, and sub-index e for eliminated degree of freedom, respectively. Then the condensed stiffness matrix may be written as [3]:
-1 K * = K dd - K de K ee K ed dd
(2.36)
Deformation caused by pure bending Deformation caused by pure shear Local deformations of the connection and the column regions opposite to the beam flanges
36
With this model, the bending and shear deformations of the joint zone are obtained simply by defining the correct cross section inertia moment and shear area for the members Ci-Cm and Cm-Cs, for the members Bi-Bm , Bm-Bs , Di-Dm, Dm-Ds, and also for the members Bm-Cm, Cm-Dm, the cross section inertia moment and shear area should be infinite. The axial flexibility of the members Bi-Ci and Ci-Di, must correspond to the sum of the local flexibilities of the connection and of the column, on the regions opposite to the lower flanges of the left and right beams, respectively. The same is valid for members Bs-Cs and Cs-Ds, but with the upper instead of the lower flanges of the beams. The members Bm-Cm and Cm-Dm must transmit only the beam shear forces to the column (the left and right beam bending moments and axial forces are transmitted by Bi-Ci/Bs-Cs and Ci-Di/Cs-Ds, respectively); to provide for this, the link between members Bm-Cm and Bi-Bs, at Bm, and the link between members Cm-Dm and Di-Ds, at Dm, only transmit vertical force.
Bs Bm Bi
Cs Cm Ci
Ds M Dm Di
Fig 2.12 Mechanical semi rigid joint model proposed by Tschemmernegg et.al
37
CHAPTER THREE
BEHAVIOR AND MODELING OF SEMI RIGID FRAMES.
3.1 INTRODUCTION
The effect of semi rigid connections on the behavior of steel frames and their potential economical benefit is well recognized. The beam to column connection rotational behaviors directly affect the frame stability as it increases drift of the frame and causes a decrease in effective stiffness of the member which is connected to the joint. An increase in frame drift will multiply the second order (P- ) effect of beam-column members and thus will affect the over all stability of the frame. Hence, the non linear features of the beam to column connections have important function in structural steel design. In order to incorporate the non linear behavior of semi rigid connections in semi rigid frame analysis various modeling approaches have been proposed, as discussed in chapter two. Most of the approaches are based on experimental results. Various design codes, EC3 and EC4 for example, recently included in their edition methods to include the effect of semi rigid connections in the global analysis of partial strength frames. The lack of a large and parameterized experimental data base doesnt allow for generating standardized functions. Thus, there is a need to be able to analytically generate a reliable moment rotation response of semi rigid connections which can be incorporated in the analysis and design of partial strength frames. Lately E. Bayo et.al proposed a new four node element based on Eurocode component based method to model internal as well as external joints for the global analysis of steel and composite frames. This new modeling scheme is not suitable to use in the standard available structural software. A two node simplified element whose stiffness matrix can be derived subsequently has been adopted in this study to investigate the static and dynamic behavior of partial strength frames.
38
ANSYS has an inbuilt element defined in its library called MATRIX27 characterized by six degrees of freedom (Ux, Uy, Uz, x , y , z ) at each node. The stiffness coefficients of the two node element proposed in this study has been used as an input, appropriately, as a real constant for MATRIX27. A reference frame is analyzed by using the proposed two node element and the result is compared with the four node element proposed by Bayo et.al and the EC component based method - factor. The proposed model gave an adequate result.
Vi, wi
h
V j, wj
I
Mi, i Ni, ui
Kcws
Mj, j
J
Nj, uj
Kcwc
Fig. 3.1. Position and simplified effective modeling of semi rigid connections In order to derive the stiffness matrix of the connection element, the simplest case of the element with only three springs is considered, as shown in Fig. 3.1 (b). As in the original
39
component method it is assumed that a joint will primarily deform in its plane, and all other out of plane and torsional degrees of freedom (DOF) are assumed to be rigidly connected. This reduces the problem to a two node element with three degrees of freedom per node, suitable to use the stiffness matrix element MATRIX27 available in the ANSYS library. If now each of these DOF is moved individually, as shown in Fig. 2.14, and the resulting spring forces are calculated, it is possible, with the help of some geometrical considerations and the spring characteristics, to derive the stiffness matrix of the basic connection element. The resulting stiffness matrix for the basic two dimensional connection element, is shown in equation (3.1) below.
eq2 Feq2=Keq2eq2,i Keq2 Fcws=Kcwscws,i ui=1
I
Keq2
Kcws
Kcws wi=1
I
Kcwc
cwc
Fcwc=Kcwccwc,i
(a) ui=1
eq2,i
Feq2=Keq2eq2,i
Kcwc
(b) wi=1
Keq2 Kcwc
Fcws
(c) i=1
cwc,i
Fcwc=Kcwccwc,i
Fig.3.2. Stiffness derivation process for node I of the basic connection element.
40
(K
cwc
+ K eq 2 ) 0
0 K cwc 0
h (K eq 2 K cwc ) 2
(K cwc + K eq 2 ) 0 h (K eq 2 K cwc ) 2
0 K cwc 0
h (K eq 2 K cwc ) 2
(K cwc + K eq 2 ) 0 h (K eq 2 K cwc ) 2
0 K cwc 0
(K
cwc
+ K eq 2 ) 0
0 K cwc 0
h (K eq 2 K cwc ) 2
h (K eq 2 K cwc ) 2
(3.1)
41
J5
400cm 0.25kN/m
20kN J1 B1 J2 J3 B2 J4
400cm
600cm
300cm
42
Table 3.1.Comparison of results obtained by the proposed model, EC3 and the model of Bayo et al. Bayo et al. (1) [kNm] J1 16.37 J2 78.11 J3 16.87 J4 18.67 J5 17.79 J6 63.77 J7 28.72 J8 8.33 B1 57.88 B2 6.72 B3 64.35 5.98 B4 * wrt with respect to Location Proposed model (2) [kNm] 22.24 75.28 8.92 23.34 24.80 58.37 18.45 13.15 63.79 12.01 71.21 12.34 using of EC3 (3) [kNm] 18.81 71.78 9.38 19.83 19.71 57.18 21.00 9.26 59.83 9.90 66.68 9.37 Rigid [kNm] 27.54 92.04 18.56 23.69 28.41 74.35 32.42 11.10 52.71 7.01 61.12 13.00
wrt* (3) 3.43 3.50 0.46 3.51 5.09 1.19 2.55 3.89 3.96 2.11 4.53 2.97
wrt* (1) 5.87 2.83 7.47 4.67 7.01 5.40 10.14 4.82 5.91 5.29 6.86 6.36
Fig. 3.4 shows the frame to be investigated in the subsequent sections of this thesis. The static loads shown are arbitrarily chosen. The member cross sections and joint characteristics are the same as the reference frame of Fig. 3.3.
0.25kN/m 10kN IPE300 IPE200
400cm
B5 J9 C3 B3 IPE200
400cm
B1 J1 C1
HEB180 HEB160
600cm
300cm
43
Table 3.2 Characteristics of the semi rigid connection used for frame in Figs 3.2 and 3.3. Connection Characteristics Joint J1 J2 J3 J4 J5 J6 J7 J8 J9 J10 J11 J12 Kcws [kN/cm] 4852 5583 8432 7330 4852 5583 8432 7330 4852 5583 8432 7330 Kcwc [kN/cm] 20875 19420 29290 20465 20875 19420 29290 20465 20875 19420 29290 20465 Keq [kN/cm] 10790 10760 10130 10160 10790 10760 10130 10160 10790 10760 10130 10160 Resistance [kN.cm] 7342 8306 5413 4824 7342 8306 5413 4824 7342 8306 5413 4824 h [cm] 29.35 29.35 19.70 19.70 29.35 29.35 19.70 19.70 29.35 29.35 19.70 19.70 d [cm] 16 18 18 16 16 18 18 16 16 18 18 16
44
Table 3.3 Comparisons of moments (kNm) at the joints and beam mid spans for semirigid and rigid connections. Location J1 J2 J3 J4 J5 J6 J7 J8 J9 J10 J11 J12 B1 B2 B3 B4 B5 B6 max (cm) Rigid [kNm] 17.00 102.26 12.05 31.72 37.69 89.96 15.92 24.51 30.62 74.46 31.99 12.10 52.87 6.24 48.68 7.91 59.96 6.08 3.90 Semi Rigid [kNm] 13.71 86.21 4.65 29.32 30.74 75.61 7.90 23.27 27.46 59.44 18.72 13.44 62.54 11.14 59.33 12.54 69.05 12.04 4.81 [kNm] 3.29 16.05 7.40 2.40 6.95 14.35 8.02 1.24 3.16 15.02 13.27 -1.34 9.67 4.90 10.65 4.63 9.09 5.96 0.91 % increase 19.35 15.69 61.41 7.57 18.44 15.95 50.38 5.06 10.32 20.17 41.48 11.07 18.29 78.53 21.88 58.53 15.16 98.03 23.18
45
46
47
Fig. 3.5 Static analysis results of semi rigid and rigid frames
3.4.1. Introduction
The 1994 Northridge earthquake (USA) has shown the vulnerability of rigid connections in steel moment resisting frames subject to severe earthquake ground motions. Since then, as an alternative connection type, semi rigid connections are considered for the retrofit and new design of steel moment resisting frames in high seismic regions [20]. In seismic design, all steel frames are expected to suffer large inelastic deformation to dissipate energy during strong ground motion [12, 25]. In the current design codes no analysis or design guidance is given for semi rigid frames although the possibility of using them as the main energy dissipative mechanism has been recognized. In this thesis the analysis and response of a typical semi rigid frame depicted in Fig3.4 is carried out using a general finite element package, ANSYS. For comparison purpose the same frame but rigidly connected is studied. The frames are subjected to the same earthquake and founded on soil class B. The peak ground acceleration is 0.1g (seismic hazard zone 4 as per EBCS-8) and the normalized elastic acceleration response- spectrum corresponding to the stated soil condition is used [21].
48
[K ]{ i } = i 2 [M ]{ i }
where :
( 3.2)
[K ]
{ i }
i
stiffness matrix mode shape vector (eigen vector) of mode i natural frequency of mode i ( i is the eigen value) mass matrix
2
[M ]
Many numerical methods are available to solve the above equation. ANSYS offers these methods:
Block Lanczos method Subspace method Power Dynamics method Damped method
The damped, unsymmetric, and QR damped methods are not available in the ANSYS Professional program.
49
b. Subspace method
The subspace method uses the subspace iteration technique, which internally uses the generalized Jacobi iteration algorithm. It is highly accurate because it uses the full K and M matrices. For the same reason, however, the subspace method is slower than the reduced method. This method is typically used in cases where high accuracy is required or where selecting master DOF is not practical.
d. Reduced method
The reduced method uses the HBI algorithm (Householder-Bisection-Inverse iteration) to calculate the eigen values and eigenvectors. It is relatively fast because it works with a small
50
subset of degrees of freedom called master DOF. Using master DOF leads to an exact K matrix but an approximate M matrix (usually with some loss in mass). The accuracy of the results, therefore, depends on how well M is approximated, which in turn depends on the number and location of masters.
e. Unsymmetrical method
The unsymmetric method, which also uses the full K and M matrices, is meant for problems where the stiffness and mass matrices are unsymmetric (for example, acoustic fluid-structure interaction problems). It uses the Lanczos algorithm which calculates complex eigenvalues and eigenvectors if the system is non-conservative (for example, a shaft mounted on bearings). The real part of the eigenvalue represents the natural frequency and the imaginary part is a measure of the stability of the system - a negative value means the system is stable, whereas a positive value means the system is unstable. Sturm sequence checking is not available for this method. Therefore, missed modes are a possibility at the higher end of the frequencies extracted.
f. Damped method
The damped method is meant for problems where damping cannot be ignored, such as rotor dynamics applications. It uses full matrices (K, M, and the damping matrix C). It uses the Lanczos algorithm and calculates complex eigen values and eigenvectors (as described below). Sturm sequence checking is not available for this method. Therefore, missed modes are a possibility at the higher end of the frequencies extracted.
51
of the eigen value will be zero. The eigen value results reported by ANSYS are actually divided by 2 . This gives the frequency in Hz (cycles/second). In other words: Imaginary part of eigen value, as reported =
In a damped system, the response at different nodes can be out of phase. At any given node, the amplitude will be the vector sum of the real and imaginary components of the eigenvector.
i. QR damped method
The QR damped method combines the advantages of the Block Lanczos method with the complex Hessenberg method. The key concept is to approximately represent the first few complex damped eigen values by modal transformation using a small number of eigenvectors of the undamped system. After the undamped mode shapes are evaluated by using the real eigen solution (Block Lanczos method), the equations of motion are transformed to these modal coordinates. Using the QR algorithm, a smaller eigenvalue problem is then solved in the modal subspace. This approach gives good results for lightly damped systems and can also be applicable to any arbitrary damping type (proportional or non-proportional symmetric damping or nonsymmetrical gyroscopic damping matrix). This approach also supports nonsymmetrical stiffness if present in the model. The sub space method is used in this study for its accuracy.
3.4.2.2. Spectral analysis
Two types of spectrum analyses are supported by ANSYS: the deterministic response spectrum method and the nondeterministic random vibration method. Both excitation at the
52
support and excitation away from the support are allowed. The three response spectrum methods are the single-point, multiple-point and dynamic design analysis method. The random vibration method uses the power spectral density (PSD) approach.
a. Assumptions and restrictions
1. The structure is linear. 2. For single-point response spectrum analysis and dynamic design analysis method, the structure is excited by a spectrum of known direction and frequency components, acting uniformly on all support points or on specified unsupported master degrees of freedom (DOFs). 3. For multi-point response spectrum and power spectral density analyses, the structure may be excited by different input spectra at different support points or unsupported nodes. Up to ten different simultaneous input spectra are allowed.
b. Description of analysis
The spectrum analysis capability is a separate analysis type and it must be preceded by a modal analysis. If mode combinations are needed, the required modes must also be expanded, as described in modal analysis. The four options available are the single-point response spectrum method (SPRS), the dynamic design analysis method (DDAM essentially a response spectrum analysis in which the spectrum is obtained from a series of empirical equations and shock design tables provided in the U.S. Naval Research Laboratory Report NRL-1396.), the random vibration method (PSD statistical measure defined as the limiting mean-square value of a random variable. It is used in random vibration analyses in which the instantaneous magnitudes of the response can be specified only by probability distribution functions that show the probability of the magnitude taking a particular value.) and the multiple-point response spectrum method (MPRS similar to PSD but with no correlation between spectra). Each option is discussed in detail in ANSYS Release 9.0 documentation [23]. The only method available in the ANSYS Professional program is the single-point response spectrum, discussed in detail subsequently.
53
Both excitation at the support (base excitation) and excitation away from the support (force excitation) are allowed for the single-point response spectrum analysis (SPRS). The table below summarizes these options as well as the input associated with each. Table 3.4 Types of Spectrum Loading
Excitation Option Excitation at support Excitation away from support
Response spectrum table Direction vector Constant on all support points Velocity Acceleration
Amplitude multiplier table x, y, z direction at each node Amplitude in x, y, or z directions Displacement Force
=
i
i + c + 2
+ im
(3.3)
s j
54
where : ,i i c Nm j Ej =
s m
effective damping ratio for mode i beta damping undamped natural circular frequency of the ith mode damping ratio (input as RATIO, DMPRAT command) number of materials damping constant stiffness matrix multiplier for material j
1 { i }T K j { i } = strain energy 2 displacement vector for mode i stiffness matrix of part of structure of material j modal damping ratio of mode i
[ ]
{ i }
[K ]
j
The participation factors for the given excitation directions are defined as:
i = {}i [M ]{D}
T
for the base excitation option for the force excitation option
(3.4a) (3.4b)
i = {}i {F}
T
55
where : i participat ion factor for the ith mode normalized eigenvecto r vector describing the excitation direction input force vector
{}i
[D ]
{F}
{D} = [T ]{e}
where:
(3.4c)
{D} = [D1a
a j
Da 2
a D3
D a ... 4
D = excitation at DOF j in direction a; a may be either x, y, z, or rotations about one of these axes 1 0 0 [T ] = 0 0 0 where: 0 1 0 0 0 0 0 0 1 0 0 0 0
(- z + z 0 ) (y y 0 )
1 0 0
(z z 0 )
0
(- x + x 0 )
0 1 0
(- y + y 0 ) (x x 0 )
0 0 0 1 (3.4d )
x, y, z global Cartesian coordinates of a point on the geometry xo, yo, zo global Cartesian coordinates of point about which rotations are
56
We can calculate the statically equivalent actions at j due to rigid-body displacements of the reference point using the concept of translation of axes [T]. For spectrum analysis, the determination of Da values may be obtained accurately from [23].
Complete Quadratic Combination Method (CQC) Grouping Method (GRP) Double Sum Method (DSUM) SRSS Method NRL-SUM Method (NRL SUM)
These methods generate coefficients for the combination of mode shapes. This combination is done by a generalization of the method of the square root of the sum of the squares which has the form:
57
N N 2 R a = ij R i R j i =1 j=1
(3.5)
where: Ra total modal response N ij total number of expanded modes coupling coefficient.
The value of ij = 0.0 implies modes i and j are independent and approaches 1.0 as the dependency increases Ri = Aii modal response in the ith mode Rj = Ajj modal response in the jth mode Ai mode coefficient for the ith mode Aj mode coefficient for the jth mode i the ith mode shape j the jth mode shape
i and j may be the DOF response, reactions, or stresses. The DOF response, reactions, or stresses may be displacement, velocity or acceleration.
58
( 3.6)
(1 r
2 2
) ( + r )r ) + 4 r(1 + r ) + 4(
8
' i ' i 1 ' 2 j ' i ' j ' j 2
3 2 '2 i 2
+ 'j r 2
r=
j i
Grouping Method
This method is from the NRC Regulatory Guide. For this case, equation 3.4 specializes to:
N N R a = ij R i R j i =1 j=1 where : j i 0.1 1.0 if i ij = 0.0 if j i > 0.1 i 2
1
(3.7)
Closely spaced modes are divided into groups that include all modes having frequencies lying between the lowest frequency in the group and a frequency 10% higher. No one frequency is to be in more than one group.
59
The Double Sum Method also is from the NRC Regulatory Guide. For this case, equation 3.4 specializes to: N N R a = ij R i R j i =1 j=1 where : 2
1
(3.8a )
ij =
i' = damped natural circular frequency of the ith mode i = undamped natural circular frequency of the ith mode i'' = modified damping ratio of the ith mode
The damped natural frequency is computed as : = i 1 i
' i '
( ))
1 2 2
The modified damping ratio i'' is defined to account for the earthquake duration time:
i'' = i' +
where:
t d i
(3.8b)
SRSS Method
The SRSS (Square Root of the Sum of the Squares) Method is from the NRC Regulatory Guide. For this case, Eq. 3.4 reduces to:
60
N 2 R a = (R i ) i =1
1 2
( 3.9)
NRL-SUM Method
The NRL-SUM (Naval Research Laboratory Sum) method (O'Hara and Belsheim) calculates the maximum modal response as: N 2 R a = R a1 + (R ai ) i=2 where: |Rai| absolute value of the largest modal displacement, stress or reaction at the point Rai displacement, stress or reaction contributions of the same point from other modes.
3.4.2.2.2 Analysis results
1 2
(3.10)
The main objective of the thesis is to investigate how connection flexibility influences the dynamic response of semi rigid frames. As shown in the previous sections the frames under investigation are three storey frames of the same geometry, cross section and material properties but one rigidly connected and the other being flexible. Modal analysis is carried out using the subspace method and the results are shown in Table 3.5. As can be seen readily from Table 3.5, connection flexibility increases the natural period of vibration of the frame, especially in the lower modes, while it tends to decrease the frequency. Thus semi rigid frames will be subjected to lower lateral seismic loads as compared to rigid ones.
61
Table 3.5: Modal Analysis results, natural period and frequency of rigid and semirigid frames. Rigid Frame Mode 1 2 3 4 5 where: natural frequency Tn natural period i (rad/sec) 3.641 11.594 19.886 31.206 35.944 Tn (sec) 1.726 0.542 0.316 0.201 0.175 Semi-rigid Frame i (rad/sec) 3.195 10.608 19.263 29.979 33.440 Tn (sec) 1.966 0.592 0.326 0.210 0.188 Increase in period 0.241 0.050 0.010 0.008 0.013 % increase 13.945 9.295 3.234 4.093 7.488
62
63
64
Earthquake responses
In order to study the earthquake response of the frames using elastic design spectrum of EBCS-8 soil class B and seismic hazard zone of 4 is taken. Both the semi rigid and rigid frames are subjected to the same earthquake and the responses are tabulated as shown in table below. As can be seen from the table the shear forces and internal moments in rigid frames are more than the semi rigid frame for the same earthquake. The higher the natural period of the frame the lower the action effect of the earthquake, which resulted in the lower shear force and bending moments. Table 3:6 Spectral analysis results: story shear, drift, joint moment and span moment Story Shear Story level 0 1 2 3 Semi rigid [kN] 26.14 21.86 12.24 12.24 Rigid [kN] 53.73 43.29 21.39 21.39 Inter story drift Semi rigid [cm] 0.00 3.14 3.89 2.18 Rigid [cm] 0.00 3.37 3.72 1.93
65
Table 3.6 (contd) Moment Semi- Rigid [kNm] 53.47 55.37 26.08 32.94 41.43 42.93 19.90 24.57 16.06 16.85 7.92 10.11 19.14 13.28 14.85 9.75 5.90 4.10 42.44 28.09 12.01 64.97 42.57 17.43 37.76 16.62 4.97 [kNm] 19.14 16.17 17.47 16.44 11.35 9.39 10.88 9.60 1.77 1.09 2.53 1.78 4.40 5.13 2.48 2.78 0.15 0.34 8.00 7.48 3.94 13.32 13.83 7.09 7.57 5.86 3.47
Rigid [kNm] 72.61 71.54 43.55 49.38 52.78 52.32 30.78 34.17 17.83 17.94 10.45 11.89 23.54 18.41 17.33 12.53 6.05 4.44 50.44 35.57 15.95 78.31 56.4 24.52 45.33 22.48 8.44
% increase 35.80 29.20 66.99 49.91 27.40 21.87 54.67 39.07 11.02 6.47 31.94 17.61 22.99 38.63 16.70 28.51 2.54 8.29 18.85 26.63 32.81 20.50 32.49 40.68 20.05 35.26 69.82
66
67
Fig.3.8 Moment and shear diagram of semi rigid and rigid frames subjected to an earthquake respectively
CHAPTER FOUR
CONCLUSION AND RECOMMENDATIONS 4.1. CONCLUSIONS
In this study a two node simplified connection element is proposed. The panel shear deformation of the joint has been taken care of by allowing the joint region of the column to undergo shear deformation. The proposed model has given a good result as compared to the EC3 component method, using the interaction parameter factor. Though there are some discrepancies between the results obtained by Bayo et.al[3] and the proposed model at some locations, it will be suitable for design purposes. The static analysis response of the frames has indicated that a reduction in the joint moment is accompanied by an increase in the span moments. Reducing joint moments is advantageous as detailing, modeling and design of joints is the most cumbersome part of steel frame design. In composite construction beams are usually laterally restrained and have sufficient strength to sustain design loads in their span than connection region. This will make semi rigid connections an economical design solution. From the dynamic analysis response of the two frames it can be concluded that joint flexibility will increase the natural period, and hence decrease the natural frequency, of the frame especially in the fundamental mode. Correspondingly the stresses produced due to an earthquake excitation in the semi rigid frame elements are significantly lower than the rigid frame elements. This indicates that semi rigid frame design will give a good solution in earthquake prone areas. The large deformation capability of semi rigid frames will increase their energy dissipating capacity, provided the lateral deformation is within the limits.
68
4.2. RECOMMENDATIONS
In this thesis the planar static and dynamic response behavior of semi rigid frame is studied using a simplified two node connection modeling based on the EC component method. The model is easy and suitable for use in day to day semi rigid steel frame design. In the model the real dimension of the connection elements has not been taken into account. It is clear that further and extensive research work needs to be done in order to accurately incorporate the behavior of semi rigid connections in the analysis and design of semi rigid frames. The following are among the areas of semi rigid connections which need further research. 1. Efficient semi rigid connection modeling in 3-D 2. Development of computer program incorporating accurately the behavior of semi rigid connections. 3. Extensive investigation of semi rigid connections in order to provide design methods, and detailing and fabrication criteria for their use in seismic structures. 4. Quantification of typical rotation demand on composite semi rigid connections.
69
APPENDIX A
MATRIX27 represents an arbitrary element whose geometry is undefined but whose elastic kinematic response can be specified by stiffness, damping, or mass coefficients. The matrix is assumed to relate two nodes, each with six degrees of freedom per node: translations in the nodal x, y, and z directions and rotations about the nodal x, y, and z axes. See MATRIX27 in the ANSYS, Inc. Theory Reference for more details about this element. Other similar, but less general, elements are the spring-damper element (COMBIN14), and the mass element (MASS21).
I
y
[K ], [C], or [M ]
The node locations and the coordinate system for this element are shown in Fig. A.1: "MATRIX27 Schematic". The element is defined by two nodes and the matrix coefficients. The stiffness, damping, or mass matrix constants are input as real constants. The units of the stiffness constants are Force/Length or Force Length/Radian and the damping constants, Force Time/Length and Force Length Time/Radian. The mass constants should have units of Force Time2/Length or Force Time2 Length/Radian.
70
All matrices generated by this element are 12 by 12. The degrees of freedom are ordered as
Ux, Uy, Uz, ROTx, ROTy, ROTz for node I followed by the same for node J. If one node is not
used, simply let all rows and columns relating to that node default to zero. The matrix constants should be input according to the matrix diagrams shown in "MATRIX27 Output Data". For example, if a simple spring of stiffness K in the nodal x direction is desired, the input constants would be C1 = C58 = K and C7 = -K for KEYOPT (2) = 0 and KEYOPT (3) = 4. A summary of the element input is given below.
MATRIX27 Input Summary
Nodes
I, J
Degrees of Freedom
DAMP
Surface Loads
None
Body Loads
None
Special Features
71
KEYOPT (2)
Matrix formulation:
0 --
Symmetric matrices
1 --
Unsymmetric matrices
KEYOPT (3)
The solution output associated with the element consists of node displacements included in the overall nodal solution. There is no element solution output associated with the element.
72
KEYOPT (4) = 1 causes the element matrix to be printed (for the first sub-step of the first load step only).
For KEYOPT(2) = 0, the symmetric matrix has the form:
C3 C14 C 24 C 25
C4 C15 C 25 C 34
C5 C 43
C6
C7
C8
C9
C10
C11 C 22
C 23
C 33
C 42
C 51 C 57
C 58 C 63
C 64 C 68
C 69 C 72
C 73 C 75
C 76 C 77
C 50
C12 C 23 C 33 C 42 C 50 C 57 C 63 C 68 C 72 C 75 C 77 C 78
C3 C14 C 24 C84
C4 C 34 C88
C5 C 43 C 93
C6
C7
C8
C9
C10
C11 C 22
C136
C 51 C 99 C139
C 58 C106 C140
C 64
C14 C141
C 69 C123 C132 C142
C 73 C133 C143
C 76 C144
C135
C137
C138
73
Nodes may be coincident or no coincident. Since element matrices should normally not be negative definite, a note is printed for those cases where this can be easily detected. With a lumped mass matrix all off-diagonal terms must be zero. The matrix terms are associated with the nodal degrees of freedom and are assumed to act in the nodal coordinate directions
When used in the product(s) listed below, the stated product-specific restrictions apply to this element in addition to the general assumptions and restrictions given in the previous section.
Damping and unsymmetric matrices are not allowed. Real constants C79 through C144, for unsymmetrical matrices, are not applicable. The birth and death special feature is not allowed. KEYOPT (2) can only be set to 0 (default). KEYOPT (3) = 5 is not allowed. The DAMP material property is not allowed.
74
APPENDIX B
STIFFNESS COEFFICIENTS OF VARIOUS COMPONENTS OF THE END PLATE JOINT [1, 17] Table B1
Table B2
75
76
TableB2 (contd)
77
0 4852 0 0 - 4852 0
0 - 4852 0 0 4852 0
J2
0 5583 0 0 - 5583 0
0 - 5583 0 0 5583 0
J3
39420 0 188726 - 39420 0 - 188726 0 8432 0 0 - 8432 0 188726 0 1543385 -188726 0 -1543385 -39420 0 -188726 39420 0 188726 0 - 8432 0 0 8432 0 -188726 0 -1543385 188726 0 1543385
78
J4
30625 0 101504.3 - 30625 0 - 101504 0 7330 0 0 - 7330 0 101504.3 0 1367888 -101504 0 -1367888 -30625 0 -101504 30625 0 101504.3 0 - 7330 0 0 7330 0 -101504 0 -1367888 101504.3 0 1367888
79
REFERENCES
1. R. Bjorhovde, A. Colson, R. Zandonini Bpcwheatons Ltd, 1996 2. W.F.Chen And S.Toma, Advanced Analysis of Steel Frames, CRC Press , Inc, 1994 3. E.Bayo, J.M.Cabrero, B.Gil. An Effective Component Based Method to Model SemiConnections in Steel Structures III,
Rigid Connections for the Global Analysis of Steel and Composite Structures.
Engineering Structures 28 (2006) 97-108 4. D.C.Keulen, D.A.Nethercot, H.H.Snijder, M.C.Bakker. Frame Analysis Incorporating
Semi Rigid Joint Action: Applicability of the Half Initial Secant Stiffness Approach.
Journal of Construction Steel Research 59(2003) 1083-1100 5. E.Bayo,J.Gracia, J.M.Cabrero, B.Gil. Advanced Global Member Stability Analysis of
Semi Rigid Frames. Stability and Ductility of Steel Structures, Lisbon, Portugal,
September 6-8,2006 6. S.O.Degertekin and M.S.Hayalioglu. Design of Non Linear Semi Rigid Steel Frames
With Semi Rigid Column Bases. Electronic Journal of Structural Engineering, 4 (2004) 7. P.Premthamkorn and P.Chomchuen. Modeling of Semi Rigid Steel Connections by Multi linear Equations
8. A.A. Del Savio, L.F. Martha, P.C.G Da Silva Vellasco, S.A.Lopes De Andrade.
of Steel and Composite Portal Frames With Semi Rigid Connections. Engineering
Structures 28(2006) 543-556 11. W.K.Yu, K.F.Chung,M.F.Wong. Analysis of Bolted Moment Connections in Cold
80
Partial Strength Beam to- Column Composite Joints for Seismic Applications.
Earthquake Engineering and Structural Dynamics 2007;36:142-161 13. A.U.Ozturk and H.H. Catal. Dynamic Analysis of Semi Rigid Frames. Mathematical and Computational Applications:10:1:(2005):1-8, 14. F.Block, I.Burgess, B.Davison and R.Plank. The Development of a Component Based
Connection Element for End Plate Connections in Fire. University of Sheffield, civil
and structural engineering department 15. Min Liuand Scott A. Burns. Multiple Fully Stressed Designs of Steel Frame Structures
Rigid and Semi Rigid Frames. Turkish Journal of Engineering and Environmental
Science 27(2003), 253-264 20. E.M.Lui and A.Lopes. Dynamic Analysis and Response of Semi Rigid Frames. Engineering Structures:19(1997):8: 644-654. 21. EBCS 8 (1995): Design of Structures for Earthquake Resistance, Ministry of Works and Urban Development, Addis Ababa, Ethiopia 22. EBCS 3 (1995): Design of Steel Structures, Ministry of Works and Urban Development, Addis Ababa, Ethiopia 23. EBCS 4 (1995): Design of Composite Steel and Concrete Structures, Ministry of Works and Urban Development, Addis Ababa, Ethiopia. 24. ANSYS Release 9.0 Documentation, ANSYS Inc.,2004
81
25. W. Weaver, Jr. James M.Gere. Matrix Analysis of Framed Structures, D.Van Nostrand Company, 2nd ed., 1980.
82
CANDIDATES DECLARATION
I hereby declare that the work which is being presented in this thesis entitles Behavior and Modeling of Semi-regid Steel Beamto-column Connections is original work of my own, has not been presented for a degree in any other university and that all sources of material used for the thesis have been duly acknowledged.
Date
This is so certify that the above declaration made by the candidates is correct to the best of my knowledge.
Date
83