Sie sind auf Seite 1von 8

DESIGN OF MEMBERS UNDER BENDING AND AXIAL COMPRESSION WITH INTERMEDIATE LATERAL RESTRAINTS

Numerical Simulations and Application of EC 3

Aswandy Institute for Steel, Timber and Shell Structures Graz University of Technology E-mail: aswandy@steel.tu-graz.ac.at

Richard Greiner Institute for Steel, Timber and Shell Structures Graz University of Technology Email: greiner@steel.tu-graz.ac.at

ABSTRACT In this paper, the application of the member buckling rules of EUROCODE3 (EC3 / EN19931-1) is studied for members with intermediate lateral restraints under bending and axial compression. Different kinds of restraints, either just lateral or lateral & torsional restraint have been considered in the middle or in the third-points of the span. Numerical simulations on the basis of Geometrically and Materially Nonlinear Analysis with Imperfection (GMNIA) have been performed for the reasons of verification. Members with I-, H-, and rectangular hollow section have been used, but just results of IPE 500 and HEB300 are presented. The comparison of the numerical simulations and the EC3-interaction formulae showed that the design formulae can give sufficiently good approximation of the capacity of restrained members. 1 INTRODUCTION

Members in real steel structures usually have lateral restraints within the span, which result from transversal beams or stiffeners. This stiffening is a basic measure of structural steelwork design and needs to be taken into account in member design for economical reasons. Research about members with lateral restraints have been done by several researchers [1,2,3], but they did not directly connect it to code-regulations. EUROCODE 3 [4] presents equations for buckling design of members, which are subjected to bending and axial compression. These equations are applied to members with intermediate restraints in the following. The present paper deals with the effect of lateral restraints on the buckling behaviour of single span beam columns. The work has been carried out by numerical simulations on the basis of a Geometrically and Materially Nonlinear Analysis with Imperfections (GMNIA). The ABAQUS program [5] has been used for this purpose, taking into account appropriate geometrical imperfections and residual stresses. Different loading conditions such as distributed load and varying bending moment diagrams have been studied in connection with axial compression. Both lateral flexural and lateral torsional buckling have been considered. The main parameter of the study is the location of the lateral and /or torsional restraints. The numerical investigations have been supplemented by a comparative study using existing EC3formulae Method 2, in order to show the efficiency of these formulae for the given design situation.

ASSUMPTIONS AND SCOPE OF THE STUDY

In this work, single-span members of IPE 500, HEB 300 and RHS 280/100/10 sections with end fork conditions and lateral restraints are studied. The restraints are located in mid-span or in the third-points of the member length. They may be just lateral supports (L-Restraints) or lateral & torsional supports (L&T-Restraints) (see Fig. 1). For pure L-restraints, two positions in the section depth have been considered; -either at the upper flange or in the center point of the section. The study comprises a number of different load-cases, which consist of axial compression plus main axis bending. The bending moments are distributed in different forms, i.e. uniform, parabolic, triangular and bi-triangular along the member-length. For the numerical GMNIA-calculations the relevant bow imperfections had to be determined, i.e. those shapes producing the lowest member capacity. Several shape combinations consisting of different types of lateral, in plane and torsional imperfections- had to be investigated in order to find the relevant imperfection shapes of the different restraint condition. The residual stresses accounted for are illustrated in Fig. 1. The steel grade was assumed as S235 with fy = 235 N/mm2, E=21000 N/mm2, and a bilinear elasto-plastic material law was considered.
L-Restraints L/2 L/2 d
L/3 L/3 L/3

+ + 0.5fy 0.3fy IPE 500 HEB 300

L&T-Restraints L/2 L/2


L/3 L/3 L/3

+0.5fy -0.167fy + +

z = 2.0

RHS 280/100/10

(a) Model of structures and restraints

(b) Sections and residual stresses

Fig. 1: Structural model and sections

DESIGN FORMULAE OF EC3, PART 1 1

For the application of the EC3- design formulae of Method 2 the specific parameters of laterally restrained buckling modes have to be determined. The Method 2 as defined in EC3 presently- is related to members with intermediate restraints, acting both laterally & torsionally. The LT-buckling effect of different moment distributions between the restraints can be taken into account by specific CmLT-factors and the critical buckling moment Mcr can be determined for the segment between the restraints like for single span members (see Fig. 2). Accordingly, the buckling behaviour of such members can be described by the method consistently. However, since in the present study also restraints were considered, which act just laterally, the critical forces Ncr and Mcr must be found for the specific kind of restraint. For this, elastic bifurcation calculations were performed by ABAQUS. These values were then introduced into the design-formulae and comparisons with the GMNIA-results were made to illustrate the validity of the formulae. The rules for member buckling of EC3, Method 2, for I-sections of class 1 and 2 are as follows:

Buckling mode y-y :

M y,Ed N Ed + k yy 1 M y,Rk N Rk y LT M1 M1 M y,Ed N Ed 1 + k zy N Rk M y,Rk z LT M1 M1

(1)

Buckling mode z-z :

(2)

The auxiliary terms: n y =


nz =

N Ed y N Rk M 1
N Ed z N Rk M 1

(3) (4)

k yy = C my 1 + y 0.2 n y C my 1 + 0.8 n y

( (

) )

(5) (6) (7) (8)

0 .1 z 0 .1 k zy = nz 1 nz 1 (C mLT 0.25 ) (C mLT 0.25 )

Cmy = 0.6 + 0.4 0.4 CmLT = 0.6 + 0.4 0.4

CmLT My,Ed NEd c L Cmy

My,Ed
NEd

y for L z for c LT for c


Cmy for CmLT for

Fig. 2: Member with L&T-Restraints

DETERMINATION OF Ncr AND Mcr

4.1 Critical Buckling Load Ncr for Members with Lateral Restraints While in case of L&T-Restraints the critical axial load of lateral buckling modes is always Ncr,z for the section shapes considered here, it may be the torsional or torsional flexural load Ncr,T, or Ncr,TF in case of L-Restraints. Fig. 3 illustrates the critical values for an IPE 500 with different locations of the L-Restraints. It is evident that the torsional buckling modes are intersected by the lateral flexural modes at certain slenderness.

d 1,0 (h 2)
0,5 0,0

N
TFB 1.0 2.0 FBz 3.0 4.0 FBz z=5.0

N
L 1-restraint 2-restraints 3-restraints 9-restraints

d h
N cr ,real
2 EI z
L2

C=

-0,5

-1,0 0 5

IPE 500
C
10 15

Fig. 3: Influence of the L-Restraints position for members subjected to axial compression

4.2 Critical Buckling Moment Mcr for Members with Lateral Restraints

The critical moments for restraints at mid-span are presented in Fig. 4a for different moment diagrams. It illustrates that for constant moment the results are identical for L&T-Restraints and L-Restraints at the top or the center of the section. This is again shown in Fig. 4b more in detail, where we can find that a L-Restraint acts like a L&T-Restraint if it is located in the compression area or even lower when the slenderness increases. This typical behaviour has also been shown in [1] exemplarily. For non-uniform moments Fig. 4a demonstrates the effects of the different restraints. It shows that a L-Restraint at the top or center of the section has little influence in case of bi-triangular moments.

Mcr IPE 500 200000


z = 2.0

1,0

a b

a) b) c)

M1 L/2 M1

M1

d (h 2)
M1

M
L/2 L/2

dh

IPE 500

0,5

150000

L/2 L/2

c
100000

M1 L/2 L/2

M1

C1 =
0,0

Mcr,real Mcr

(L ) 2

d e f

d)

M1 L/2 L/2

M1

50000

e) M1
L/2 L/2

M1

-0,5

z = 0.5

1.0 2.0 3.0 4.0 5.0

f)
0 1.0 0.5 0.0 -0.5

M1 L/2 L/2

M1

-1.0

-1,0 0,0 0,2 0,4 0,6 0,8

7.07
1,0

C1
(b) The influence of the restraint positions

(a) The critical moments

Fig 4: Critical moments and influence of the restraint positions

MEMBERS WITH RESTRAINTS AT MID-SPAN

Out of the study just the figures for selected cases of profiles IPE 500 and HEB 300 with z = 2 are presented here. These figures contain the results of the numerical simulation of GMNIA as well as the results of the design formulae of Method 2 of EC3. The given cases present L&T-Restraints as well as L-Restraints at the flange and at the center. The three cases in Fig 5 for axial compression and uniform bending show that differences of L&T-Restraint and L-Restraint at the center just occur for pure or dominating axial force due to the difference of Ncr,z and Ncr,T. The design formulae give results of good quality; the difference for pure bending is caused by the definition of the LT values in EC3. In case of LRestraint at the top flange (in compression) the GMNIA-curve is identical to the previous ones as long as the bending moment is dominating. When the axial compression increases the curve changes and approaches the Ncr,TF for pure axial load. It can be well understood that in this case the positive bending moment acts stabilizing for the unrestrained flange for high axial compression. The Method 2 formula cannot describe this beneficial effect of the bending moment, but gives a conservative result in this region.
0,8

M N
z = 2.0 IPE 500
L/2 L/2

0,8

M N
L/2 L/2

0,8

M N
z = 2.0 IPE 500
L/2 L/2

N
0,6

N
0,6

0,6

z = 2.0 IPE 500

My/Mpl,y

My/Mpl,y

0,4

0,4

My/Mpl,y
GMNIA EC3 M2

0,4

0,2
GMNIA EC3 M2

0,2

0,2
GMNIA EC3 M2

0,0 0,0

0,0
0,6 0,8

0,0

0,2

0,4

0,0

0,2

0,4

0,6

0,8

0,0

0,2

0,4

0,6

0,8

N/Npl

N/Npl

N/Npl

Fig. 5: Interaction diagrams for members subjected to uniform moment and axial compression

The buckling behaviour under triangular bending moment is illustrated by the three cases of Fig. 6. The different restraints produce different interaction curves here. The L&T-Restraints lead to results, which can be well described by the Method 2-formulae, because the difference between the real Mcr and that calculated in simplified way for the governing segment of the member with the length L/2 and =0.5 is rather small. In case of a L-Restraint in the center the real Mcr has to be used for the Method 2-formula in order to describe the behaviour with good quality. Thereby, the CmLT-values should be based on the moment diagram of the member-segment. In case of L-Restraints at the top flange (in compression) the interaction curve is almost the same as that of L&T-Restraint as long as the bending moment governs. For increasing axial compression the curve turns to approach the Ncr,TF value. Again the Method 2-formula can just conservatively describe this behaviour. Also here the real Mcr-value has to be used.

1,0

M N
z = 2.0 IPE 500
L/2 L/2

1,0

M N
L/2 L/2

1,0

M N
L/2 L/2

0,8

0,8

N
My / Mpl,y

0,8

z = 2.0 IPE 500

z = 2.0 IPE 500


0,6

My /Mpl,y

My/Mpl,y

0,6

0,6

0,4
GMNIA EC3 M2 0,2 Real Mcr EC3 M2 Mcr->simplified 0,0

0,4

0,4

0,2

0,0 0,4 0,6 0,8 0,0

GMNIA EC3 M2 Real Mcr

0,2

GMNIA EC3 M2 Real Mcr

0,0 0,6 0,8 0,0 0,2 0,4 0,6 0,8

0,0

0,2

0,2

0,4

N/Npl

N/Npl

N/Npl

Fig. 6: Interaction diagrams for members subjected to triangular moment and axial compression

The three cases for axial compression and bi-triangular moments are given in Fig. 7. The case of L&T-Restraint leads to a buckling moment resulting in LT=1.0, so that the difference between the real Mcr and simplified one has no effect. The EC3-formula gives a sufficiently good quality. For L-Restraint at the center the capacity is lower under bending moment, because the restraint does not improve the Mcr here. For the Method 2-formula the real Mcr has to be used and the CmLT-value must be related to the moment diagram of the segment of the member. For a L-Restraint at the top flange the capacity is just a little bit increased for high bending moments, which is connected with the higher Mcr according to Fig. 4. The Method 2-formula used with the real Mcr and CmLT of the segment leads to good results.
1,0

M
L/2 L/2

1,0

M
0,8

M
L/2 L/2

1,0

M
0,8

M
L/2 L/2

0,8

My/Mpl,y

My/Mpl,y

My/Mpl,y

0,6

z = 2.0 IPE 500

0,6

z = 2.0 IPE 500

0,6

z = 2.0 IPE 500

0,4
GMNIA EC3 M2 0,2 Real Mcr EC3 M2 Mcr->simplified 0,0

0,4

0,4

0,2

0,0 0,4 0,6 0,8 0,0

GMNIA EC3 M2 Real Mcr

0,2 GMNIA EC3 M2 Real Mcr 0,0


0,6 0,8

0,0

0,2

0,2

0,4

0,0

0,2

0,4

0,6

0,8

N/Npl

N/Npl

N/Npl

Fig. 7: Interaction diagrams for members subjected to bi-triangular moment and axial compression

As a further step the buckling behaviour of an HEB 300 profile with L-Restraint at the center is presented. This is to illustrate, that the more stocky cross section now causes both in-plane and out-of-plane LT-buckling modes, while for the examples of the IPE 500-sections above just out-of-plane LT-buckling modes occurred. This means that for in-plane buckling the equation (1) is relevant and for out-of-plane the equation (2). Fig. 8 indicates those parts of the interaction curves by that equation-number which is relevant there. It was found that for pure axial force in-plane flexural buckling occurs and that with increasing bending moment this buckling mode remained to be governing. Just for high bending moment the buckling mode turned to out-of-plane LT-buckling. The method 2-formulae- used in analogous way as before -lead to sufficient approximation.
1,0 1,0

M
L/2 L/2 z,member = 2.0 z,segment = 1.0 y,member = 1.19 HEB 300

1,0

0,8

0,8

(2)

L/2 L/2 z,member = 2.0 z,segment = 1.0 y,member = 1.19 HEB 300

0,8

GMNIA EC3 M2 Real Mcr

(2) My/Mpl,y
0,6

My/Mpl,y

0,4

(1)

My/Mpl,y

0,6

0,6

0,4

0,4

0,2
GMNIA EC3 M2

0,2

0,0 0,0 0,2 0,4 0,6

0,0 0,0

GMNIA EC3 M2 Real Mcr

(1)

L/2 L/2 0,2 z,member = 2.0 z,segment = 1.0 y,member = 1.19 HEB 300 0,0

(1)

0,2

0,4

0,6

0,0

0,2

0,4

0,6

N/Npl

N/Npl

N/Npl

Fig. 8: Interaction diagrams of members with HEB 300 section

MEMBERS WITH TWO RESTRAINTS AT THIRDS-POINTS

In extension of the above the behaviour of members with two equally spaced restraints are presented in Fig. 9. The GMNIA-results determined with the relevant set of imperfection shapes are compared with the results of the method 2-formulae which have been applied analogously as described before. This means, that for L-Restraints the real buckling moment Mcr is to be used and the CmLT-factor is related to the segment between the restraints. The figures on the one hand show some differences for pure bending according to different buckling moments Mcr for the three cases, and on the other hand differences for pure axial compression according to different buckling modes for the axial force. While for L&TRestraints Ncr,z is governing, it is Ncr,T for L-Restraint in the center and Ncr,TF for L-Restraint at the top. The comparison with the design formulae shows, that in particular in the range of high axial compression some conservative results exist.

1,0

M N
L/3 L/3 L/3

1,0

M N
L/3 L/3 L/3

1,0

M N
L/3 L/3 L/3

0,8

N
z =3.0 IPE 500

0,8

N
z =3.0 IPE 500

0,8

N
z =3.0 IPE 500

My/Mpl,y

My/Mpl,y

My/Mpl,y
GMNIA EC3 M2 Real Mcr

0,6

0,6

0,6

0,4
GMNIA 0,2 EC3 M2 Real Mcr EC3 M2 Mcr->simplified 0,0

0,4

0,4

0,2

0,2
GMNIA EC3 M2 Real Mcr

0,0
0,4 0,6 0,8

0,0 0,4 0,6 0,8 0,0 0,2 0,4

0,0

0,2

0,0

0,2

0,6

0,8

N/Npl

N/Npl

N/Npl

Fig. 9: Interaction diagrams for members with two restraints at thirds-point

CONCLUSION

The present studies are to confirm the efficiency of the buckling design formulae of EC3 Method 2 also for members with intermediate restraints. The code so far covers explicitly just L&T-Restraints and the results given above for these cases confirm the validity of the design formulae. The extension to just L-Restraints in two different positions at center or top of the section leads to the necessity to determine the relevant buckling modes for Ncr and Mcr of the specific restraint condition. The application of the design formulae by using the real Ncr and Mcr showed that good accordance is again reached by the existing code-formulae. The main aim for use in practice is, therefore, to derive approximate tables or diagrams of Ncr and Mcr for practically relevant restraint positions.
REFERENCES

[1] Biljaard, F.S.K., Steenbergen, H.M.G.M., Lateral Torsional Stability of Members with Lateral Restraints at Various Locations between The Supports, Proceeding of Third European Conference on Steel Structure, vol. 1, Portugal, 2002, pp 191-200. [2] Nethercot, D.A., Buckling of Laterally or Torsionally Restrained Beams, Journal of the Engineering Division, ASCE, Vol. 99, No. EM4, 1973, pp. 773-791. [3] Gosowski, B., Einflu von Punktversteifungen auf die Einfachsymmetrischer Stbe, Stahlbau 68, Heft 5, 1999, S.374-381 Biegedrillknicklast

[4] Eurocode 3 : Design of Steel Structures. Part 1-1 : General rules and rules for buildings. EN 1993-1-1. 2003 [5] ABAQUS Standard Users Manual version 6.3. Hibbit, Kalrsson and Sorensen, Inc., 2002.
KEYWORDS

Flexural buckling, lateral torsional buckling, EC3, lateral restraint, lateral & torsional restraint, restraint position, GMNIA

Das könnte Ihnen auch gefallen