Engineering Structures 30 (2008) 31143118

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Analysis of the steel frames with the semi-rigid beam-to-beam and beam-to-column knee joints under bending and axial forces
Alfonsas Daniunas , Kestutis Urbonas
Department of Steel and Timber Structures, Vilnius Gediminas Technical University, Sauletekio al. 11, LT-10223 Vilnius, Lithuania

article

info

a b s t r a c t
This paper presents an analysis of framework structures with semi-rigid joints. The semi-rigid end-plate bolted joints are subjected to bending and tension or a compression axial force. Usually the influence of an axial force on joint rotational stiffness is neglected. In some cases, the level of axial forces in the joints of structures can be significant and has a profound influence on characteristics of semi-rigid joints. One of the most popular practical methods permitting the determination of rotational stiffness and moment resistance of a joint is the so-called component method. The extension of the component method for evaluating the influence of bending moment and axial force on the rotational stiffness and moment resistance of a joint and an application to the analysis of framework structures are presented in the paper. The numerical results of calculations of steel frameworks are presented in this paper as well. The study shows that the estimation of axial force influence to rotation stiffness characteristics of the joint has significant influence on the distribution of internal forces and displacements of steel framework. 2008 Elsevier Ltd. All rights reserved.

Article history: Received 18 January 2008 Received in revised form 11 April 2008 Accepted 21 April 2008 Available online 29 May 2008 Keywords: Steel semi-rigid joints Knee joints Bending moment Axial force Component method Joints characteristics Steel frameworks analysis

1. Introduction One of the possibilities to improve the accuracy of steel framework analysis can be an evaluation of the actual behavior of joints. For such purposes, the semi-rigid joint conception was introduced. The most relevant influence on characteristics of the joint is produced by bending moments. In this case, the behavior of a semirigid joint is described by momentrotation M curve of the joint. This conception is adopted by most researchers [16] and design code EC3 [7]. Joints are also influenced by axial and shear forces. These forces are not usually evaluated in calculating the rotational stiffness and moment resistance of the joints. There are many different types of joints in which the axial force may be significant and it has a profound influence on joint rotational stiffness and moment resistance. The influence of an axial force on characteristics of the joint has not been analysed widely and it has not been completely investigated yet [810]. Various methods may be used for the establishment of momentrotation M curves of semi-rigid joints [3]. The most common method applied in practice is the so-called mechanical component method. It estimates the mechanical properties of the material and geometrical properties of the joint and allows

calculating the rotational stiffness and moment resistance of the joint [3,7,11]. The component method is developed for evaluation of the bending moment influence on momentrotation characteristics of joint. This paper presents extension of the component method to semi-rigid beam-to-beam end-plate bolted and beam-to-column end-plate bolted knee joints under bending and axial force. The main focus is made on an estimation of the axial force influence in addition to bending moment influence on the joints rigidity and strength. Estimation of real behavior of the joints has significant influence to the behavior results of the steel frameworks [12 15]. In this article, the iterative procedure for the calculation of steel frameworks with semi-rigid joints is presented. In each iteration, the steel frame is analysed using joint momentrotation characteristics, which are calculated using the bending moments and axial forces, obtained in the previous iteration. 2. The evaluation of bending moment and axial force in joints by component method The joints of steelworks are acted by bending moments as well as axial and shear forces. The influence of these forces to the behavior of joint depends mostly on the type of the joint. From the practical point of view, the investigation of total bending moment and axial force influence on joint characteristics is important for end-plate beam-to-beam knee, beam-to-column knee and

Corresponding author. Tel.: +370 5 274 50 03; fax: +370 5 270 01 12.
E-mail address: [email protected] (A. Daniunas). 0141-0296/$ see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.engstruct.2008.04.027

 A. Daniunas, K. Urbonas / Engineering Structures 30 (2008) 31143118

3115

Notation
E Fc Fc,i,Rd Fc,max Fc,max,i,Rd

Ft Ft,max Ft,max,r,Rd

Ft,max Ft,min,i,Rd

Ft,r,Rd M Mj,Rd N Sj,ini V hr keq zc zeq zt zt

t,eq t,max,eq t,min,i

modulus of elasticity compression force in components resistance of the weakest compressed component i maximal compression force in components resistance of the weakest compressed component i of the joint, when both sides of the joint are under compression forces tension force in components maximal tension force in components effective tension resistance of bolt row r in tension part of the joint, where the maximal tension force is acting minimal tension force in components resistance of the weakest component i in tension part of the joint, where the minimal tension force is acting effective tension resistance of bolt row r bending moment moment resistance axial force initial rotational stiffness of the joint shear force distance from bolt row r to the centre of compression equivalent stiffness coefficient of two or more boltrows in tension distance between the centre of the connection and the centre of compression equivalent lever arm distance between the centre of the connection and the centre of tension distance between centre of connection and bolt row in the lower part of connection rotation of joint equivalent tension deformation equivalent tension deformation of components under maximal tension force tension deformation of components under minimal tension force

Mechanical model of the joint depends on values and directions of acting forces [17,18]. There are three possible cases: (a) components are under compression or under tension when one part of the joint is in tension and the other one is in compression Fig. 2(b); (b) no components are under tension when both sides of the joint are in compression Fig. 2(c); (c) components are under tension when both sides of the joint are in tension Fig. 2(d). Tension and compression force in components, when one part of the joint is under tension and the other under compression force Fig. 2(b) can be expressed by the formula:
Ft = Fc = M zeq M zeq

(N cos V sin ) zc
zeq

;
(1)

(N cos V sin ) zt
zeq

where M is accepted as always positive;(N cos V sin ) expresses value of joint axial force which is perpendicular to end-plate; N is axial force in connected element and accepted as positive for tension case; means + or and according to this proposition normal to the end-plate of shear force (V sin ) accepted as positive if it causes the tension of the joint. Equivalent deformation of components under tension can be found by the formula:
t,eq =
M zeq

(Ncos V sin )zc

E keq

M + (N cos V sin ) zc E keq zeq

(2)

where keq is equivalent stiffness coefficient of two or more boltrows in tension and zeq is equivalent lever arm. The rotation of the end-plate joint depends on the deformation of all tensed components i:
= t,eq
zeq

(3)

The initial rotational stiffness of the joint is expressed by the ratio between bending moment and rotation of the joint:
Sj,ini = M

M zeq

t,eq

(4)

Maximal compression force in components, when both sides of the joint with symmetric section elements are under compression forces Fig. 2(c), can be expressed by:
Fc,max = M

column-to-column joints. Some of them are presented in Fig. 1. In some framework structures with these types of joints axial forces may be significant and may strongly influence joint rotational stiffness and moment resistance. In this paper the procedure, which evaluates the behavior of these joints under the bending moment and axial force is based on the component method. The evaluation of the behavior of joint in the component method consists of three main steps: identification of the basic component, evaluation of the mechanical properties of the components, and assembling active components to one mechanical model [10]. According to the component method, the characteristics of the joint must be determined by the basic components of the joint [16]. The component method allows to characterise all components of the joint independent from the type of loading applied to the joint. An elastic stiffness coefficient and resistance that belongs only to mechanical and geometrical data of the joint represent each component. Axial force increases or decreases the forces in components from the bending moment. This property allows component method suggested to evaluate joint behavior under the bending moment to extend for evaluation joint behavior under bending moment and axial force.

2zc

(N cos V sin )
2

(5)

In such case, there are no components in tension and the joint according to EC3 [7] in such type of loading behaves as ideally rigid (see formulas (3) and (4)). Maximal and minimal tension forces in components, when both sides of the joint are under tension forces Fig. 2(d) can be expressed by:
Ft,max = Ft,min

(N cos V sin ) zt ; (zt + zt ) M (N cos V sin ) zt = + . (zt + zt ) (zt + zt )

(zt + zt )

(6)

Deformation of components under maximal and minimal tension forces can be found the following way:
t,max,eq =
(zt +zt )
M

(Ncos V sin )zt (zt +zt )

E keq
t

;
(7)

t,min,i =

M (zt + + z )

(Ncos V sin )zt (zt +zt )

E kt , i

3116

A. Daniunas, K. Urbonas / Engineering Structures 30 (2008) 31143118

Fig. 1. Semi-rigid end-plate joints: (a) column-to-beam knee joint; (b) beam-to-beam knee joint; (c) beam-to-beam joint.

Fig. 2. Loading schemes and mechanical models of joint: (a) beam-to-beam knee joint with positive direction of internal forces; (b) upper part of joint in tension and lower part in compression; (c) both sides of joint in compression; (d) both sides of joint in tension.

The initial rotational stiffness of the joint for this case of loading is expressed by:
Sj,ini = M

M(zt + zt )

t,max,eq

t,min,i

(8)

The above presented procedure for stiffness calculation is only applicable in the case where the behavior of the structure is elastic. The moment resistance of the joint loaded by bending and axial forces can be expressed as a function of axial force (the force that is perpendicular to end-plate) and resistance of the components. When one part of the joint is under tension and the other one is under compression force Fig. 2(b), the moment resistance can be expressed by the formula:
Mj,Rd = min

Fig. 3. Flowchart of framework analysis with semi-rigid joints under bending and axial forces.

3. Analysis of steel frames with semi-rigid joints including influence of bending moments and axial forces In this chapter, the procedure for calculation of steel frames with semi-rigid joints suggested by the authors is presented. This procedure can be used as part of an iterative design process for steel frames. The calculation of steel frames according to this procedure can be performed with the software used in steelwork design. The evaluation of the rotational stiffness of joints is the minimum requirement for the software to be used. When the influence of bending moment and axial forces is taken into account the characteristics of joint and magnitude of acting forces are interdependent. In this case characteristics of joints as well as frame forces and displacements have to be calculated by an iterative procedure until the changes of joint characteristics and forces in structure between two iterations become very small (Fig. 3). When the influence of the axial forces is disregarded, then no iteration procedure is needed for determination of characteristics of the joints and internal forces of the frame.

(Ft,r,Rd hr ) (N cos V sin ) zc Fc,i,Rd zeq + (N cos V sin ) zt

(9)

where hr distance from bolt row r to the centre of compression; indicate the sum according to r, where r is the number of bolt rows in tension part of joint. When both sides of the joint are under compression forces Fig. 2(c), the moment resistance can be expressed by the formula:
Mj,Rd = zc (2Fc,max,i,Rd + (N cos V sin )) .

(10)

When both sides of the joint are in tension forces Fig. 2(d), the moment resistance can be expressed by the formula:
Mj,Rd = min

(Ft,max,r,Rd (hr zc + zt )) (N cos V sin )

zt (N cos V sin ) zt Ft,min,i,Rd (zt + zt ) .

where Ft,max,r,Rd effective tension resistance of bolt row r in tension part of the joint, where the maximal tension force is acting; Ft,min,i,Rd resistance of the weakest component i in tension part of the joint, where the minimal tension force is acting.

 A. Daniunas, K. Urbonas / Engineering Structures 30 (2008) 31143118

3117

Fig. 4. Portal frame. Fig. 6. Vertical displacement of rooftop joint B (Fig. 4) when the influence of axial forces is taken into account (M + N) and disregarded (only M).

Fig. 7. Roofing structure.

Fig. 5. Horizontal displacement of joint A of the left column (Fig. 4) when the influence of axial forces is taken into account (M + N) and disregarded (only M).

4. Numerical examples In order to investigate the effect of semi-rigid joints on the behavior of framed structures the portal frame and roofing structure are examined. The joint stiffness and strength characteristics under bending and axial forces are obtained and analysis of framed structures is performed according to the procedure given in previous sections. Example 1. The investigated structure is presented in Fig. 4. The structure was investigated for the four cases of inclination angle (0, 15, 30 and 45 degrees). It was loaded by uniformly distributed vertical and horizontal loads and they are applied simultaneously. The steel grade of all members is S235 and Youngs modulus is 210 GPa. Beams and columns in joints A and B are connected by the bolted end-plate knee joint with four M24 8.8 classes bolts. Four cases of end-plates with the thickness of 12, 16, 20 and 26 mm were analysed. The calculations were performed when the characteristics of the joints A and B were calculated taking into account the influence of axial force or disregarding it. The iterative procedure is applied for problem solving. The initial rotational stiffness of joints for different end-plate thickness at the final iteration is presented in Table 1. Figs. 5 and 6 show the influence on frame displacements when the rotational stiffness of the joint is obtained by taking into account bending moment and axial force and when only bending moment is taken into account. Example 2. Investigated roofing structure is presented in Fig. 7 [18]. Beams were loaded by a uniformly distributed vertical load. The steel grade of all members is S355 and Youngs modulus is 210 GPa. Beams were connected by the bolted end-plate knee joint A with four M24 8.8 classes bolts. Four cases of end-plates with the thickness of 12, 16, 20 and 26 mm were analysed. The calculations were performed when the rotational stiffness of the joint A is calculated by taking into account the influence

Fig. 8. Influence of the thickness of the end-plate to the bending moment in the joint A (Fig. 7) when the influence of axial forces is taken into account (M + N) and disregarded (only M).

Fig. 9. Influence of the thickness of the end-plate to the axial force in the joint A (Fig. 7) when the influence of axial forces is taken into account (M + N) and disregarded (only M).

of the axial force or disregarding it. This example shows the case when acting tension axial force in beams decreases the rigidity of the joint A and changes the behavioral results of the whole roofing structure. The iterative procedure is applied for problem solving. Some obtained results (Table 2, Figs. 8 and 9) of the analysed roofing structures behavior have shown a significant influence of the axial force to the internal forces of the structure. In addition, internal forces and displacements of the structure are significantly influenced by the thickness of the end-plates of the joint A.

3118

A. Daniunas, K. Urbonas / Engineering Structures 30 (2008) 31143118

Table 1 The initial stiffness SA,ini (joint A) and SB,ini (joint B) (Fig. 4) with the different thickness of end-plate and angle of inclination (in brackets is given initial stiffness, when axial force is disregarded) Initial stiffness (kNm/rad) Thickness of the end-plate (mm) 26 20 16 12

= 0
SA,ini SB,ini SA,ini SB,ini SA,ini SB,ini SA,ini SB,ini

238 000 (215 000) 96 700 (88 900)

170 700 (153 700) 69 300 (63 700) 138 000 (116 000) 80 830 (69 800) 119 000 (93 000) 126 800 (93 000) 111 348 (78 600) 3200 000 (153 700)

114 000 (102 800) 46 350 (42 600) 92 600 (77 500) 54 040 (46 700) 79 740 (62 000) 84 300 (62 000) 75 160 (52 500) 2200 000 (102 800)

59 000 (53 200) 24 000 (22 000) 48 240 (40 150) 28 040 (24 200) 42 175 (32 000) 42 610 (32 000) 40 150 (27 200) 1300 000 (53 200)

= 15
192 000 (161 000) 112 800 (97 400)

= 30
164 700 (130 000) 176 960 (130 000)

= 45
152 000 (110 000) 4500 000 (215 000)

Table 2 The initial stiffness SA,ini of joint A and vertical displacements uv of joint A and point B (Fig. 7) with the different thickness of end-plates in joint A Thickness of the end-plate (mm) 26 A B 20 A B 16 A B 12 A B

The influence of the axial force is estimated

53 661 6.97 97 440 7.05

12.07

37 449 6.90 69 840 7.01

12.44

23 905 6.80 46 700 6.94

12.99

10 871 6.58 24 200 6.80

14.17

The influence of the axial force is disregarded 11.64 11.86 12.21 12.98

5. Conclusions In the article, the suggested extension of the well-known component method for estimation of the total influence of axial force and bending moment allows one to obtain more accurate stiffness and moment resistance characteristics of the various joints. The performed investigations have shown that axial force influences the rotational stiffness and moment resistance of the joints. Tension axial force decreases the rotational stiffness of the joints and compression axial force increases it. Analysis of steel frameworks has shown that calculation results differ when total influence of bending moment and axial force on characteristics of the joints was estimated and when only influence of bending moment on characteristics of the joint was estimated. Suggested simple iterative procedure for the analysis of framed structures with semi-rigid joints with estimation bending moments and axial forces allow to use the framework analysis software, which enable to evaluate only rotational stiffness of the joints. Therefore, joints in which significant axial forces occur have to be designed by estimating the total influence of bending moment and axial force. References
[1] Hasan R, Kishi N, Chen WF. A new nonlinear connections classification system. J Construct Steel Res 1998;47:11940. [2] Jaspart JP. General report: Session on connections. J Construct Steel Res 2000; 55:6989. [3] Faella C, Piluso V, Rizzano G. Structural steel semirigid connections: Theory, design and software. Boca Raton: CRC Press LLC; 2000.

[4] Goto Y, Miyashita S. Classification system for rigid and semirigid connections. J Struct Eng 1998;124:7507. [5] van Keulen DC, Nethercot DA, Snijder HH, Bakker MCM. Frames analysis incorporating semi-rigid joint action: Applicability of the half initial Secant stiffness approach. J Construct Steel Res 2003;59:1083100. [6] Chen WF, Goto Y, Liew JYR. Stability design of semi-rigid frames. John Wiley & Sons; 1996. [7] Eurocode 3, EN 1993-1-8. Design of steel structuresPart 18: Design of joints, CEN Brussels. Belgium: 2005. [8] de Lima LRO, da Silva LS, da S Vellasco PCG, de Andrade SAL. Experimental analysis of extended end-plate beam-to-column joints under bending and axial force. Eurosteel, Coimbra 2002;112130. [9] Cerfontaine F, Jaspart JP. Analytical study of the interaction between bending and axial force in bolted joints. Eurosteel, Coimbra 2002;9971006. [10] Sokol Z, Wald F, Delabre V, Muzeau JP, varc M. Design of end plate joints subject to moment and normal force. Eurosteel, Coimbra 2002;121928. [11] Jaspart JP, Weynand K. Extension of the component method to joints in tubular construction. In: Proceedings of the Ninth International Symposium and Euroconference on tubular structures. Dsseldorf: 2001. p. 51723. [12] Ashraf M, Nethercot DA, Ahmed B. Sway of semi-rigid steel frames, part 2: Irregular frames. Eng Struct 2007;29:185463. [13] Aristizabal-Ochoa JD. Large deflection and postbuckling behavior of Timoshenko beamcolumns with semi-rigid connections including shear and axial effects. Eng Struct 2007;29:9911003. [14] Djebbar N, Chikh N-E. Limit period based on approximate analytical methods estimating inelastic displacement demands of buildings. J Civil Eng Management 2007;13:2839. [15] Gizejowski MA, Barszcz AM, Branicki CJ, Uzoegbo HC. Review of analysis methods for inelastic design of steel semi-continuous frames. J Construct Steel Res 2006;62:8192. [16] Weynand K, Jaspart JP, Steenhuis M. The Stiffness Model of revised Annex J of Eurocode 3. In Proceedings of the third international workshop on connections in steel structures behavior, strength and design. Trento: 1995. p. 44152. [17] Urbonas K, Daniunas A. Component method extension to steel beam-to-beam and beam-to-column knee joints under bending and axial forces. J Civil Eng Management 2005;XI(3):21724. Vilnius: Technika. [18] Urbonas K, Daniunas A. Behaviour of semi-rigid beam-to-beam joints under bending and axial forces. J Construct Steel Res 2006;62:12449.