BEAM-COLUMNS

SUMMARY:

Structural members subjected to axial compression and bending are known as beam columns. The interaction of normal force and bending may be treated elastically or plastically using equilibrium for the classification of cross-section. The behaviour and design of beam-columns are presented within the context of members subjected to uniaxial bending, i.e. deformation takes place only in the plane of the applied moments. In the case of beam-columns which are susceptible to lateral-torsional buckling, the out-of-plane flexural buckling of the column has to be combined with the lateral-torsional buckling of the beam using the relevant interaction formulae. For beam-columns with biaxial bending, the interaction formula is expanded by an additional term.

OBJECTIVES:

Evaluate the in-plane bending and axial compression force for beam-columns. Calculate the lateral-torsional buckling of beam-columns. Calculate the biaxial bending and axial compression force for beam-columns.

REFERENCES:
Eurocode 3: Design of steel structures Part 1.1 General rules and rules for buildings. Chen W F and Atsuta T: Theory of Beam-Columns Vols. 1 & 2, McGraw-Hill, 1976. Trahair N. & Bradford M: Behaviour and Design of Steel Structures, 2nd edt., Chapman & Hall, 1988. Dowling P J, Owens G W & Knowles P: Structural Steel Design, Butterworths, 1988. Nethercot D A: Limit State Design of Structural Steelwork, 2nd edition, Chapman and Hall, 1991.

CONTENTS:
1. Introduction. 2. In-plane behaviour of beam-columns. 2.1 Cross-sectional behaviour. 2.2 Overall stability. 2.3 Treatment in Eurocode 3. 2.4 The role of ky. 3. Lateral-torsional behaviour of beam-columns. 3.1 Lateral-torsional buckling. 3.2 The design process in Eurocode 3. 3.3 The role of kLT. 4. Biaxial bending of beam-columns. 4.1 Design for biaxial bending and compression. 4.2 Cross-section checks. 5. Verification methods for isolated members and whole frames. 6. Concluding summary.

1. INTRODUCTION.
Beam-columns are defined as members subject to combined bending and compression. In principle, all members in frame structures are actually beam-columns, with the particular cases of beams (N = 0) and columns (M = 0) simply being the two extremes. Depending upon the exact way in which the applied loading is transferred into the member, the form of support provided and the member's cross-sectional shape, different forms of response will be possible. The simplest of these involves bending applied about one principal axis only, with the member responding by bending solely in the plane of the applied moment.

2. IN-PLANE BEHAVIOUR OF BEAM-COLUMNS.

When a beam-column is subjected to in plane bending (figure 1a), its behaviour shows an interaction between beam bending and compression member buckling, as indicated in figure 1b.

Figure 1 In-plane behaviour of beam-columns. Curve 1 shows the beam elastic linear behaviour. Curve 6 shows the limiting behaviour of a rigid-plastic beam at the full plastic moment Mpl. Curve 2 shows the transition of real elastic-plastic beams from curve 1 to curve 6. The elastic buckling load of a concentrically loaded compression member, Ncr is shown in curve 4. Curve 3 shows the interaction between bending and buckling in elastic members, and allows for the traditional moment N v exerted by the axial load. Curve 7 shows the interaction between bending moment and axial force causing the member to become fully plastic. This curve allows for the reduction from the full plastic moment Mpl to Mpr caused by the axial load, and for the additional moment Nv. The actual behaviour of a beam-column is shown by curve 5 which provides a transition from curve 3 for elastic members to curve 7 for full plasticity.

2.1 CROSS-SECTIONAL BEHAVIOUR.

2.1.1 Bending and axial force for class 1 and 2 cross-sections. If full plasticity is allowed to occur, then the failure condition will be as shown in figure 2 and the combination of axial load and moment giving this condition will be:

Figure 2 Full plasticity under axial load and moment. a. For yn (h t f ) / 2 neutral axis in web

N M = 2 f y t w yn
h 2t f M N = f y bt f (h t f ) + f y 2 b. For y n > (h t f ) / 2
2 2 y n t w

(1)

neutral axis in flange

h N M = f y t w (h 2t f ) + 2b t f + y n 2 h M N = f y b y n (h y n )t f 2

(2)

Figure 3 compares Eqs. (1) and (2) with the approximation used in Eurocode 3 of:

M Ny. Rd = M pl . y (1 n) /(1 0,5a) but M Ny. Rd M ply.Rd

(3)

Eurocode 3 5.4.8.1 (5.25) or eq. 6.36

in which n = N Sd / N pl .Rd is the ratio of axial load to squash load (fy A), and a = ( A 2bt f ) / A 0,5 For cross-sections without bolt holes, the following approximations may be used for z axis moments: for n > a : M Nz. Rd = M pl . z .Rd for n > a : M Nz . Rd = M pl . z .Rd

n a 2 1 1 a

Eurocode 3 5.4.8.1 (5.26) or eq. 6.37 and 6.38

where n = N Sd / N pl .Rd and a = ( A 2bt f ) / A but a 0,5 .

Figure 3 Full plasticity interaction major axis bending of HEA 450 section.

Eurocode 3

Further simplifications for a range of common cross-sectional shapes are provided in Table 1. Table 1 Expressions for reduced plastic moment resistance MN (Notation: n = NSd / Npl.Rd). Cross-section Shape Expression for MN 5.4.8.1 Rolled I or H
M N , y = 1,11M pl . y (1 n)

(5.27)

M N , z = 1,56 M pl.z (1 n)(0,6 + n)

(5.28) (5.31) (5.32) (5.33) (5.34)

Square hollow section

M N , y = 1,26 M pl (1 n)

M N , y = 1,33M pl . y (1 n)

Rectangular hollow section

M N , y = M pl .z

1 n ht 0,5 + A

Circular hollow section

M N , y = 1,04 M pl (1 n1,7 )

In all cases the value of MN should, of course, not exceed that of Mpl.
2.1.2 Bending and axial force for Class 3 cross-sections.

Figure 4 shows a point somewhere along the length of an H-shape column where the applied compression and moment about the y axis produce the uniform and varying stress distribution shown in figures 4a and 4b.

Figure 4 Elastic behaviour of cross-section in compression and bending. For elastic behaviour the principle of superposition may be used to simply add the two stress distributions as shown in figure 4c. First yield will therefore develop at the edge where the maximum compressive bending stress occurs and will correspond to the condition: fy = c +b where: fy is the material yield stress, h is the overall depth of section and I is the second moment of area about the y axis.

 c = N / A is the stress due to the compressive load N Mh / 2 b = is the maximum compressive stress due to the moment M. I Class 3 cross-sections will be satisfactory if the maximum longitudinal stress x.Ed satisfies the criterion:

x.Ed f yd ; f yd = f y / M 0
2.1.3 Bending and axial force for class 4 cross-sections.

Eurocode 3 5.4.8.2 (5.31) or eq. 6.42

Class 4 cross-sections will be satisfactory if the maximum longitudinal stress x.Ed calculated using the effective widths of the compression elements (5.3.2.(2) of EC3) satisfies the criterion:

x.Ed f yd ; f yd = f y / M 0

Eurocode 3 5.4.8.32 (5.39) or eq. 6.43

2.2 OVERALL STABILITY.

The treatment of cross-sectional behaviour in the previous section took no account of the exact way in which the moment M at the particular cross-section under consideration was generated. Figure 5 shows a beam-column undergoing lateral deflection as a result of the combination of compression and equal and opposite moments applied at the ends.

Figure 5 Primary and secondary moments.

The moment at any point within the length may conveniently be regarded as being composed of: primary moment M secondary moment N v. Using elastic strut theory gives the maximum deflection at the centre (Trahair & Bradford, 1988) as:

vmax = where PEy =

M sec N 2

N 1 PEy

(4)

2 EI y
L2

is the Euler critical load for major axis buckling, and the maximum moment is:

M max = M sec

N PEy

(5)

In both equations the secant term may be replaced by noting that the first order deflection (due only to the end moments) and the first order moment (ordinary beam theory) are approximately amplified by:
1 1 N / PEy

(6)

as shown in figure 6.

Figure 6 Maximum deflection and moment in beam-columns with equal moments. Thus:
vmax = ML2 1 8 EI y 1 N / PEy
1 1 N / PEy

(7) (8)

M max = M

Since the maximum elastic stress will be:

max = c + b
Eq. (9) may be rewritten as:

M max M

(9)

c
fy

b
f y (1 N / PEy )

= 1,0

(10)

Eq. (10) may be solved for values of c and b that just cause yield, taking different values of PEy (which is dependent on the slenderness L/ry). This gives a series of curves as shown in figure 7, which indicate that as b 0 , c tends to the value of material strength fy.

(a)

(b)

(c)

Figure 7 Form of Eq. (10) effect of: (a) slenderness (b) cross sectional shape (c) moment gradient. Eq. (10) does not recognise the possibility of buckling under pure axial load at a stress Ey given by:

Ey =

PEy A

2 EI y
AL2

2E 2 y

(11)

Use of both Eq. (10) and Eq. (11) ensures that both conditions are covered as shown in figure 8.

Figure 8 Combination of Eqs. (10) and (11)

2.3 TREATMENT IN EUROCODE 3.

Eqs. (10) and (11) are written in terms of stresses and originate from the concept of failure being defined as either the attainment of first yield or elastic buckling of the perfect member. Limit state design codes, as Eurocode 3, normally take ultimate load as the design criterion when considering resistance under static loading.

Thus these equations must be re-written in terms of forces and moments. In doing this it is also necessary to make some allowance for those effects present in real structures that have not so far been explicitly allowed for, i.e. initial lack of straightness, residual stresses, etc. For consistency in design it is essential that the interaction equation for combined loading reduces down to the column and beam design procedures as moment and axial load respectively reduce to zero.
2.3.1 Members with class 1 and 2 cross-sections.

The approach taken in Eurocode 3 (assuming bending about the y axis) is to use:

k y M y .Sd N Sd + 1 y Af y W pl . y f y in which y is the reduction factor for column buckling, and ky = 1

(12)

Eurocode 3 5.5.4(1) (5.51) or eq. 6.61 & 6.62

y N Sd y Af y

but k y 1,5

where k y is a modification factor discussed in section 2.3.4., and

y = y (2 My 4) +

W pl , y Wel , y

1 but y 0,90

where My is an equivalent uniform moment factor accounting for the non-uniformity of the moment diagram, see table 2 (moment diagram about y axis and restraints in the z direction).
2.3.2 Members with class 3 cross-sections.

Members with class 3 cross-sections subject to bending and axial load shall satisfy: k y M y .Sd N Sd + 1 y Af y Wel . y f y where ky and y is as in Eq. (12) with (13) Eurocode 3 5.5.4(3) (5.53) or eq. 6.61 & 6.62

y = y ( 2 My 4) but y 0,90
Table 2 Equivalent uniform moment factors M. (Cm) MOMENT DIAGRAM EQUIVALENT UNIFORM MOMENT FACTOR M End moments M , = 1,8 0,7 Moments due to in-plane lateral loads For uniformly distributed load: M ,Q = 1,3

For concentrated load: M ,Q = 1,4

Moments due to in-plane lateral loads plus end moments

M = M , +

MQ M

( M ,Q M , )

where:
M Q = max M

due to lateral load only and

M = max M for moment diagram without change of sign M = max M + min M where sign of moment diagram changes

2.3.3 Members with class 4 cross-sections.

Members with class 3 cross-sections subject to bending and axial load shall satisfy: k y ( M y.Sd + N Sd eN . z ) N Sd + 1 y Aeff f y Weff . y f y (14) Eurocode 3 5.5.4(3) (5.56) or eq. 6.61 & 6.62

Where: ky and y is as in Eq. (12) with y as in Eq. (13) Aeff.y is the effective cross-sectional area for pure compression Weff.y is the effective cross-sectional modulus for pure bending eN.z is the shift in neutral axis comparing the full cross-section with the effective cross-section (calculated assuming pure compression) used to account for local buckling
2.3.4 The role of ky.

The value of ky, as shown by the equations explaining Eq. (12), depends in a rather complex way on: N Sd Level of axial load as measured by the ratio y Af y Member slenderness y Margin between the cross-sections plastic & elastic moduli Wpl & Wel (for class 1 & 2 only) Pattern of primary moments. When all of this combine in the most severe way the safe value of ky is 1,5. The role of ky is to allow for the secondary bending effect described earlier plus the effects of nonuniform moment and spread of yield. Figure 5 showed how, for the particular cases of equal and opposite beam moments, the primary moments are amplified due to the effect of the axial load N acting through the lateral displacements v.

When the pattern of primary moments is different, the two effects will not be so directly additive since maximum primary and secondary moments will not necessarily occur at the same location. Figure 9 illustrates the situation for end moments M and M, where can adopt values between +1 (uniform single curvature) and 1 (double curvature).

Figure 9 Non-uniform moment case. The particular case shown corresponds to a value 0,5. For the case illustrated the maximum moment still occurs within the member length but the situation is clearly less severe than that of figure 5 assuming all conditions to be identical apart from the value of . It is customary to recognise this in design by reducing the contribution of the moment term to the interaction relationship. Thus in Eurocode 3 ky in Eq. (12) depends upon the ratio . Since the case of uniform single curvature moment is the most severe, it follows that a safe simplification is always to use the procedure for = 1,0. Returning to figure 9, it is possible for the point of maximum moment to be at the end at which the larger primary moment is applied. This would usually occur if the axial load was small and/or slenderness was low so that secondary bending effects were relatively slight. In such cases design will be controlled by ensuring adequate cross-sectional resistance at this end. The formula, table 2, for the particular shape of cross-section being used, should therefore be employed. In cases where only the uniform moment ( = 1.0) arrangement is being considered, the overall buckling check of Eq. (12) will always be more severe than (or in the limit equal to) the cross-sectional check which, and therefore this latter check need not be performed separately.

3. LATERAL-TORSIONAL BEHAVIOUR OF BEAM-COLUMNS.

When an unrestrained beam-column is bent about its major axis (figure 10a), it may buckle by deflecting laterally & twisting at a load significantly less than the maximum load predicted by an inplane analysis. The most general situation is illustrated in Figure 10b. When bending is applied about both principal axes the member's response will be 3-dimensional in nature, involving biaxial bending and twisting.

(a)

(b)

Figure 10 Lateral-torsional behaviour.

This lateral-torsional buckling may occur while the member is still elastic (curve 1 of figure 11), or after some yielding (curve 2) due to in-plane bending and compression has occurred.

Figure 11 Lateral-torsional buckling of beam-columns.

3.1 LATERAL-TORSIONAL BUCKLING.

Considering the lateral-torsional behaviour of an unrestrained I section beam-column bent about its major axis it can be assumed the elastic behaviour and the arrangement of applied loading and support conditions given in figure 12.

Figure 12 Basic case for lateral-torsional buckling. The critical combinations of N and M may be obtained from the solution of (Chen & Atsuta, 1976):
M2 N N 1 1 = 2 i0 PEz PE 0 PEz PE 0

(15)

in which

i0 =

is the polar radius of gyration A 2 EI z PEz = is the minor axis critical load L2 GI t 2 EI w is the torsional buckling load. 1+ PE 0 = 2 2 i0 GI L t

Iy + Iz

Eq. (15) reduces to the buckling of a beam when N0 and to the buckling of a column in either flexure (PEz) or torsion (PE0) as M 0. In the first case the critical value of M will be given by:

M cr =

EI z GI t 1 +

2 EI w
L2 GI t

(16)

in which: EIz is the minor axis flexural rigidity GIt is the torsional rigidity EIw is the warping rigidity.
In deriving Eq. (15) no allowance was made for the amplification of the in-plane moments M by the axial load acting through the in-plane deflections. This may be approximated as
M . Eq. (15) can, therefore, be modified to: 1 N / PEy

M2 N 1 N 1 N = 1 2 i0 PEz PE 0 PEy PEz PE 0

(17)

Noting the relative magnitudes of PEy, PEz and PE0, and re-arranging gives the following approximation: 1 N M + =1 PEz 1 N / PEy i0 PEz PE 0 or
1 N M + =1 PEz 1 N / PEy M cr

(18)

(19)

3.2 THE DESIGN PROCESS IN EUROCODE 3.

For design purposes it is necessary to make suitable allowances for effects such as initial lack of straightness, partial yielding, residual stresses, etc., as has been fully discussed in earlier lectures in the context of columns and beams. Thus some modification to Eq. (19) is necessary to make it suitable for design. In particular, the end points (corresponding to the cases of M = 0 and N = 0) must conform to the established procedures for columns and beams
3.2.1 Members with class 1 and 2 cross-sections.

Eurocode 3 uses the interaction equation: k LT M y .Sd N Sd + 1 z Af y LT W pl . y f y (20) Eurocode 3 5.5.4(2) (5.52) or eq. 6.61 & 6.62

in which z is the reduction factor for column buckling around the minor axis, LT is the reduction factor for lateral-torsional beam buckling, and

k LT = 1

LT N Sd z Af y

but k LT 1,0

with

LT = 0,15( z 2 M , LT 1) but LT 0,90

where M,LT is a factor accounting for the non-uniformity of the moment diagram, see Table 2 (moment diagram about y axis and restrains in the y direction).
3.2.2 Members with class 3 cross-sections.

Members with class 3 cross-section should satisfy the following criterion: k LT M y .Sd N Sd + 1 z Af y LT Wel . y f y (21) Eurocode 3 5.5.4(4) (5.54) or eq. 6.61 & 6.62

3.2.3 Members with class 4 cross-sections.

Members with class 3 cross-section should satisfy the following criterion: k LT M y .Sd + N Sd eN , z N Sd + 1 z Af y LT Weff . y f y (22) Eurocode 3 5.5.4(5) (5.57) or eq. 6.61 & 6.62

3.3 The role of kLT.

The value of kLT, as shown by the equations explaining Eq. (20), depends on: the level of axial load as measured by the ratio the member slenderness z the pattern of primary moments. For the most severe combination kLT adopts the value of unity, corresponding to a linear combination of the compressive and bending terms. This reflects the reduced scope for amplification effects in this case, since the value of NSd cannot exceed z A fy, which will, in turn, be significantly less than the elastic critical load for in-plane buckling PEy. It is, of course, also necessary to ensure against the possibility of in-plane failure by excessive deflection in the plane of the web at a lower load than that given by Eq. (20). This might occur, for example, in situations where different bracing and/or support conditions are provided in the xy and xz planes as illustrated in figure 13.
N Sd z Af y

Figure 13 Column with different support conditions in xy and xz planes. Such cases should be treated by checking, in addition to Eq. (20), an in-plane equation of the form: k y M y.Sd N Sd + 1 (23) min Af y W pl . y f y in which min depends on the in-plane conditions. Usually, however, Eq. (20) will govern.

4. BIAXIAL BENDING OF BEAM-COLUMNS.

Analysis for the full three-dimensional case, even for the simple elastic version, is extremely complex and closed-form solutions are not available. Rather than starting analytically it is more convenient to approach the question of a suitable design approach from considerations of behaviour and the use of the methods already derived for the simpler cases of figure 14.

Figure 14 Biaxial bending. Figure 15 presents a diagrammatic version of the design requirement.

Figure 15 Interaction diagram for biaxial bending. The N-Mz and N-My axes correspond to the two uniaxial cases already examined. Interaction between the two moments Mz and My corresponds to the horizontal plane. When all the three load components N, My and Mz are present the resulting interaction plots somewhere in the three-dimensional space represented by the diagram. Any point falling within the boundary corresponds to a safe combination of loads. Assuming proportional loading, any combination may be regarded as a straight line starting at the origin, the orientation of which depends upon the relative sizes of the three load components. Increasing the loads extends this line from the origin until it just reaches and exceeds the boundary. In each case the axes have been taken as the ratio of the applied component to the members resistance under the load component alone, e.g. NSd/min Afy in the case of the compressive loading. Thus figure 15 actually represents the situation for one particular example with particular values of cross-sectional properties, slenderness and load arrangement.

4.1 DESIGN FOR BIAXIAL BENDING AND COMPRESSION.

Members with class 1 and 2 cross-sections subject to combined biaxial bending and axial compression should satisfy the following criterion: k y M y ,Sd k z M z ,Sd N Sd + + 1 min Af y W pl , y f y W pl , z f y where kz is a factor similar to ky, see Eq. (12). (24) Eurocode 3 5.5.4(1) (5.51) or eq. 6.61 & 6.62

Members with class 1 and 2 cross-sections subject to combined biaxial bending and axial compression where lateral-torsional buckling is relevant, should also satisfy the following criterion:

k LT M y ,Sd kM N Sd + + z z ,Sd 1 z Af y LT W pl , y f y W pl , z f y

(25)

Eurocode 3 5.5.4(2) (5.52) or eq. 6.61 & 6.62

Members with class 3 cross-sections subject to combined biaxial bending and axial compression: k y M y ,Sd k z M z ,Sd N Sd + + 1 min Af y Wel , y f y Wel , z f y (26) Eurocode 3 5.5.4(3) (5.53) or eq. 6.61 & 6.62

Members with class 3 cross-sections subject to combined biaxial bending and axial compression where lateral-torsional buckling is relevant, should also satisfy the following criterion: k M kM N Sd + LT y , Sd + z z , Sd 1 z Af y LTWel , y f y Wel , z f y (27) Eurocode 3 5.5.4(4) (5.54) or eq. 6.61 & 6.62

Members with class 4 cross-sections subject to combined biaxial bending and axial compression: k y ( M y ,Sd + N Sd eNz ) k z ( M z ,Sd + N Sd e Ny ) N Sd + + 1 min Aeff f y Weff , y f y Weff , z f y (28) Eurocode 3 5.5.4(5) (5.56) or eq.6.44, 6.61 & 6.62

Members with class 4 cross-sections subject to combined biaxial bending and axial compression where lateral-torsional buckling is relevant, should also satisfy the following criterion: k LT M y ,Sd + N Sd eNy k z ( M z ,Sd + N Sd eNz ) N Sd + + 1 z Aeff f y LT Weff , y f y Weff , z f y (29) Eurocode 3 5.5.4(6) (5.57) or eq.6.44, 6.61 & 6.62

An important point to note from the definition of Aeff and Weff above is that the calculation of crosssectional properties, and thus also cross-sectional classification, should be undertaken on a separate basis for each of the three load components N, My and Mz. This does, of course, mean that the same member may be classified as (say) class 1 for major axis bending, class 2 for minor axis bending and class 3 for compression. The safe design approach is to check all beam-columns using the least favourable class procedures.

Class 4

4.2 CROSS-SECTION CHECKS.

If allowance has been made when determining the k factors (through the use of M) for the less severe effect of patterns of moment other than the uniform single curvature bending, it is necessary further to

check that the cross-section is everywhere capable of locally restraining the combination of compression and primary moment(s) present at any point. Expressions for checking several types of cross-section under compression plus uniaxial bending were given in section 1.1. For biaxial bending Eurocode 3 uses:
M y.Sd M Ny. Rd + M z .Sd M Nz .Rd

(30)

Eurocode 3 5.5.8.1 (5.35) or eq.6.41

in which the values of and depend upon the type of cross-section as indicated in table 3. Table 3 Values of and for use in Eq. (30) (Notation: n = NSd / Npl.Rd). Type of cross-section I and H sections 2 5n but 1 Circular tubes 2 2 Rectangular hollow sections Solid rectangles and plates
1,66 1 1,33n
2

but 6

1,66 1 1,33n 2

but 6

1,73 + 1,8n3

1,73 + 1,8n3

A simpler but conservative alternative is: M y .Sd N Sd M z .Sd + + 1 N pl .Rd M Ny .Rd M Nz .Rd (31) 5.5.8.1 (5.36)

5. VERIFICATION METHODS FOR ISOLATED MEMBERS AND WHOLE FRAMES.

Normally, the design of an individual member in a frame is done by separating it from the frame and dealing with it as an isolated substructure.

The end conditions of the member should then comply with its deformation conditions, in the spatial frame, in a conservative way, e.g. by assuming a nominally pinned end condition, and the internal action effects, at the ends of the members, should be considered by applying equivalent external end moments and end forces, Figure 16. Methods of verification for these members are given in Section 5.1.

Figure 16 Isolated members A and B from the plane frame analysis. A more general procedure is given in Section 5.2, for the case where members cannot be isolated from the frame structure in the way described above.

5.1. Methods of verification for isolated members.

For the design of beam-columns, with mono-axial bending only, two checks must be carried out:

the in-plane buckling check taking into account the in-plane imperfections. the out-of-plane buckling check, including the lateral-torsional buckling verification that takes account of the out-of-plane imperfections (Figure 17).

Figure 17 Assumptions for member imperfections. It has been found by test calculations that twist imperfections, , of beam-columns that are susceptible to lateral-torsional buckling, can be substituted by flexural imperfections, see Figure 18.

Figure 18 Twist imperfections 0 and flexural imperfection W0y. Members with sufficient torsional stiffness, i.e. hollow section members, need not be verified for lateral-torsional bucking. When the non-dimensional slenderness LT 0,4, the reduction coefficient LT need not be taken into account. This rule may be used for spacing the lateral restraints to resist lateral-torsional buckling.

5.2. Method of verification of whole frames.

Figure 19 gives an example of a portal frame with tapered columns and beams, the external flanges of which are laterally supported by the purlins which, due to their flexural stiffness, also provide torsional restraint; the beams and columns may, however, be subject to distortion of the cross-section, due to the flexibility of the web.

Figure 19 Portal frame with tapered beams and columns, with elastic torsional restraints and displacements restraints by the purlins. The cross-section is susceptible to distortion. An accurate verification of this arrangement should be based on a finite element model which takes the above effects into account. The basic assumptions made regarding the imperfections in this model, would be such that the standard verification given previously would produce equally favourable results since the standard procedure has been calibrated against test results. A more simplified procedure is, therefore, given here which is related to the verification of columns for flexural buckling, and beams for lateral-torsional buckling. The basic principles governing the standard verification of columns for flexural buckling, and beams for lateral-torsional bucking, are as follows: 1. The non-dimensional slenderness is defined by:

FB =

N pl N cr

; LT =

M pl M cr

Where Npl, Mpl are the characteristic values of the elastic/plastic resistances of the column or beam neglecting any out-of-plane effects; and Ncr, Mcr are the critical bifurcation values for the column resistance, or the beam resistance, when considering out-of-plane deflections and hyperelastic behaviour in the equilibrium state. 2. Using the non-dimensional slenderness, , a reduction factor can be determined from the European buckling curves that allows the design value of the resistance of the column or beam to be defined by: Nbd = Npl / M1 for the column Mbd = Mpl / M1 for the beam. In applying this principle to any loaded structure, see Figure 20, the procedure is as follows:

Figure 20 Stepwise verification of a structure, assuming in step 1: elastic-plastic in-plane behaviour and no lateral deflections, and in step 2: hyperelastic behaviour and lateral deflection. 1. As a first step the structure is analysed for a given load case with an elastic or plastic analysis assuming that any out of plane deflections are prevented. By this analysis a multiplier, pl, of the given loads is found that represents the ultimate resistance of the structure. 2. The structure is then checked assuming hyperelastic material behaviour allowing for lateral and torsional deflections. This leads to a multiplier crit of the given loads that represents the critical elastic resistance of the structure to lateral buckling or lateral-torsional bucking. 3. The overall slenderness, , of the structure can then be defined by:

pl crit

And by using the reduction coefficient from the relevant European buckling curve, e.g. curve c, the final safety factor can be derived.: = pl This procedure is analogous to the Merchant-Rankine procedure for the frames non- elastic verification.

In general the procedure described earlier needs a computer program that performs a planar elasticplastic analysis of the frame and determines the elastic bifurcation load of the structure for lateral and torsional deflections, including distortion. Such a program, for calculating the elastic bifurcation loads, can either be based on finite elements or on a grid model where the flanges and stiffeners are considered as beams and the web is represented by an equivalent lattice system that allows for second order effects; such programs are available on PC's.

6. CONCLUDING SUMMARY.
Beam-columns are structural members subjected to axial compression and bending about one or both axes of the cross-section. The behaviour of beam-columns can be understood in three stages: (a) behaviour of the restrained beam-column; (b) uniaxial bending and compression of the unrestrained beam-column; (c) biaxial bending and compression of the unrestrained beam-column. Stage (a) is governed by the behaviour of the cross-section. Stage (b) is governed by an interaction of the cross-section behaviour with in-plane column buckling and/or lateral-torsional buckling. Stage (c) is governed by the same factors as stage (b), but the moment about the other axis must be incorporated into the design equation. For the cross-section, the interaction of normal force and bending may be treated elastically using the principle of superposition or plastically using equilibrium and the concept of stress blocks. When considering the member as a whole, secondary-bending effects must be allowed for. Strut analysis may be used as a basis for examining the role of the main controlling parameters. Design is normally based on the use of an interaction equation, an essential feature of which is the resistance of the component as a beam and as a column. The class of cross-section will affect some of the values used in the interaction equations. The biaxial bending case is the most general and includes the two others as simpler and more restricted component cases. A three dimensional frame may generally be analysed by separating it into plane frames and analysing these on the assumption of no imperfections; the individual members of the frame should then be checked with the imperfection effects taken into account. The isolated members in general represent beam-columns with either in plane or biaxial bending. In certain cases the standard procedure for the verification of a beam- column is not applicable and more accurate models must be used. As a non-linear spatial analysis including the effect of imperfections is difficult, an alternative procedure is provided by which the overall slenderness of a frame is defined; this allows verification of the frame using the European buckling curves which take account of lateral- torsional buckling.

ADDITIONAL READING
1. 2. 3. 4. 5. 6. 7. 8. 9. Narayanan, R., Editor, "Beams and Beam Columns: Stability and Strength", App Sci Pub 1983. Chen, W. F. and Atsuta, T. "Theory of Beam Columns Volume 1 & 2, McGraw Hill 1976. Timoshenko, S. P. and Gere, J. M., "Theory of Elastic Stability" 2nd Edition, McGraw Hill 1961. Bleich, F., "Buckling Strength of Metal Structures", McGraw Hill 1952. Trahair, NS Bradford, MA, "The Behaviour & Design of Steel Structures", Chapman Hall, 1988 Ballio, G. Mazzolani, F. M., "Theory and Design of Steel Structures", Chapman Hall, 1983. Galambos, T. V., "Guide to Stability Deign Criteria for Metal Structures", 4th edition, Wiley . Dowling, P. J., Owens, G. W. and Knowles, P., "Structural Steel Design", Butterworths, 1988. Nethercot, D.A., "Limit State Design of Structural Steelwork", 2nd edition, Chapman Hall, 1991.