6

Buckling of Plates

6-1

Circular Plates

When a thin elastic plate is subjected to compressive in-plane axial loads, in conjunction with small applied lateral loads or imperfections in the plate, the inplane deflections increase gradually with an increase in the applied loads up to a certain critical point. Beyond this point a slight increase in axial loads causes a large and sudden increase in the deflection. This phenomenon, called buckling, is the subject of this chapter for circular and rectangular plates. A more comprehensive treatment of this subject is given by Timoshenko (Timoshenko and Gere, 1961), Bloom (Bloom and Coffin, 2001), Szilard (Szilard, 1974), and Iyangar (Iyengar, 1988). The differential equation for the bending of a circular plate subjected to lateral loads, p, is obtained from Eq. (3-11) as
r2 d 2f df Qr 2 : r f D dr 2 dr

Where, D= E= Q= r= t= f= A= Et3/12(1 A2) modulus of elasticity shear radius of plate thickness angle as shown in Fig. 6-1. Poissons ratio

When in-plane forces Nr are applied as shown in Fig. 6-1, and the lateral loads, p, are reduced to zero, then the corresponding value of Q is
Q Nr f:

Letting
A 2 Nr =D

6-1

167

168

Buckling of Plates

Figure 6-1. In-plane load in a circular plate.

the differential equation becomes

r2 d 2f df r r 2 A 2 1f 0: 2 dr dr

Defining
x Ar and dx A dr

6-2

we get
x2 d 2f df x x 2 1f 0: 2 dx dx

6-3

The solution of this equation is in the form of a Bessel function. From Eq. (B-3) of Appendix B,
f C1 J1 x C2 Y1 x

6-4

at r = 0, Y1(x) approaches infinity. Hence, C2 must be set to zero and Eq. (6-4) becomes
f C1 J1 x:

6-5

For a fixed boundary condition, f = 0 at r = a and a nontrivial solution of Eq. (6-5) is

J1 x 0

or from Table B-1 of Appendix B, x = 3.83, and Eqs. (6-1) and (6-2) give
p Nr =D a 3:83 14:67D : a2

or
Ncr

6-6

Circular Plates

169

For a simply supported plate, the moment at the boundary r = a is zero and Eqs. (10-5) and (3-4) give
Ncr 4:20D : a2

6-7

Equations (6-6) and (6-7) are for the critical buckling load of circular plates with fixed and simply supported boundary conditions, respectively. It is of interest to note that contrary to column buckling where the column is rendered ineffective in carrying any further loads subsequent to buckling, a simply supported circular plate is capable of carrying additional loads in the postbuckling phased due to inplane biaxial membrane stress (Bloom and Coffin 2001). Postbuckling analysis of such plates (Sherbourne 1961) results in an equation of the form
Nu =Ncr 1 0:241w=t2

6-8

Where, Nu = Ncr = t= w= ultimate axial load critical buckling load plate thickness plate deflection

This equation is independent of the ratio r/t and is limited to w/t < 3.0. Values greater than 3.0 result in plastic strains in the plate. Example 6-1 What is the required thickness of a simply supported circular plate subjected to a lateral pressure of 2 psi and in-plane compressive force of 100 lb/inch if a = 29 inches, A = 0.31, E = 30,000 ksi, allowable stress in bending = 10,000 psi, and factor of safety (F.S.) for buckling = 3.0. Solution From Example 3-1,
M Pa 2 3 A 348:0 inch-lb=inch 16 p t 6 348=10; 000 0:46 inch: 6 348 8350 psi 0:5 2

Try t = 0.50 inch.

actual bending stress

actual compressive stress 100=0:5 200 psi D 30; 000 0:5 3 345; 720 inch-lbs: 121 0:312

170
From Eq. (6-7),
jcr

Buckling of Plates

4:20 345; 720 3450 psi: 29 2 0:5

allowable compressive stress = 3450/3 = 1150 psi. Using the interaction equation
actual bending stress actual compressive stress V1:0 allowable bending stress allowable compressive stress

gives
8350 200 0:83 0:17 1:0 o:k 10; 000 1150

Use t = 0.50 inch. The governing differential equation for the buckling of circular plates with a central hole, i.e. annular plate, is derived much the same way as that for solid circular plates. However, the general differential equation (Bloom and Coffin 2001) for an annular plate is more complicated than a solid plate. The equation is similar to Eq.(3-21) for the general bending of a plate with the exception of having applied axial in-plane loads Nr and tangential in-plane loads Nu rather than lateral loads p. Shearing forces Nur are assumed to be zero. The equation is expressed as
  B2 w 1 B2 w 1 B2 w Dj w Nr 2 Nu : Br r Br Bu r 2 Bu2
4

6-9

where, j4 w is given by
j4 w B4 w 2 B3 w 1 B2 w 1 Bw 2 B4 w 2 B3 w 2 3 3 2 Br 4 r Br 3 r Br 2 r Br r Br 2 Bu 2 r Bu 2 Br 4 B2 w 1 B4 w 4 r 4 Bu 2 r Bu4

6-10

The solution of Eq. (6-9) for specific plate geometry with the appropriate boundary conditions results in four simultaneous equations involving Bessel functions. Setting the determinate of the coefficients of these four equations to zero and solving the resultant polynomial provides the eigenvalues associated with buckling of the specific plate and boundary conditions. The critical buckling load is obtained from the lowest eigenvalue. Table 6-1 lists the critical buckling load for an annular plate subjected to in-plane compressive loads at the outside edge. The inner edge is free and the outer edge is either simply supported or fixed.

Rectangular Plates

171

Table 6-1. Critical Buckling Coefficient, k, of an Annular plate with a Free Inside Edge. k values b/a 0.0 0.2 0.4 0.6 0.8
Ncr = kD/a 2 Where, a = outer radius b = inside radius D = Et3 / 12(1 A2) E = modulus of elasticity k = constant Ncr = critical load, force/length t = thickness A = Poissons ratio Note 1: the critical load increases exponentially (Timoshenko and Gere, 1961)

Simply supported outer edge 4.20 3.60 2.76 2.28 2.04

Fixed outer edge 14.64 13.32 17.76 Note 1 Note 1

Problems 6-1 Derive Eq. (6-7). 6-2 What is the required thickness of a fixed circular plate subjected to a lateral pressure of 3 psi and in-plane compressive force of 500 lb/inch if a = 40 inches, A = 0.30, E = 25,000 ksi, allowable stress in bending = 25,000 psi, and factor of safety (F.S.) for buckling = 4.0? 6-3 What is the effect of Nr on the bending moments if it were in tension rather than compression? 6-2 Rectangular Plates

Buckling of rectangular plates is most commonly caused by in-plane shear or inplane axial loads in the x and y directions. In this and the following sections classical as well as numerical methods are presented for solving these loading conditions. Other loading conditions due to temperature gradients and support settlements are beyond the scope of this book. The differential equation for the bending of a rectangular plate with lateral load, p, is given by Eq. (1-26) as
5 4 w p=D:

6-11

172

Buckling of Plates

If the plate is additionally loaded in its plane (Fig. 6-2a) then summation of forces in the x-direction gives
    BNyx BNx Nx dy Nyx dx Nx dx dy Nyx dy dy 0 Bx By BNx BNyx 0: Bx By

or
6-12

Similarly, summation of forces in the y-direction gives

BNy BNxy 0: By Bx

6-13

Summation of forces in the z-direction (Fig. 6-2b) gives the following for Nx:
     Bw BNx Bw B Bw dx Nx dy Nx dx dy Bx Bx Bx Bx Bx

Figure 6-2. In-plane loads in a rectangular plate.

Rectangular Plates
which reduces to
Nx B2 w BNx Bw dx dy dx dy: Bx Bx Bx 2 BNy Bw B2 w dx dy dx dy: 2 By By By

173

6-14

Similarly, for Ny,

Ny

6-15

For Nxy from Fig. 6-3,

     BNxy Bw Bw B2 w dx Nxy dy Nxy dx dy Bx By By Bx By BNxy Bw B2 w dx dy dx dy: Bx By Bx By BNyx Bw B2 w dx dy dx dy: By Bx Bx By

or
Nxy

6-16

Similarly for Nyx

Nyx

6-17

The total sum of Eq. (6-11), which was obtained by summing forces in the zdirection, with Eqs. (6-14), (6-15), (6-16), and (6-17) gives the basic differential equation of a rectangular plate subjected to lateral and in-plane loads.
1 j w D
4

B2 w B2 w B2 w Ny 2Nxy p Nx Bx 2 By 2 Bx By

6-18

It should be noted that Eqs. (6-12) and (6-13), which were obtained by summing forces in the x- and y-directions, were not utilized in Eq. (6-18). They

Figure 6-3. In-plane shear loads.

174

Buckling of Plates

are used to formulate large-deflection theory of plates which is beyond the scope of this book. Another equation that is frequently utilized in buckling problems is the energy equation. It was shown in Chapter 5 that the strain energy expression due to lateral loads, p, is given by Eq. (5-1) as
D U 2 Z (
A

"   2  #) 2 2 B2 w B2 w B2 w B2 w B w dx dy: 21 A Bx 2 By 2 Bx 2 By 2 Bx By

6-19

The strain energy expression for the in-plane loads is derived from Fig. 6-4 which shows the deflection of a unit segment dx. Hence,
s  2 Bw V 2 dx dx dx Bx   1 Bw 2 V ex dx dx dx 2 Bx

or per unit length,

  1 Bw 2 ex : 2 Bx   1 Bw 2 ey : 2 By Bw Bw : Bx By

Similarly,

It can also be shown that

gxy

Figure 6-4. Deflection of a unit segment.

Rectangular Plates
Hence, the strain energy expression for the in-plane forces is given by
Z U
A

175

Nx ex Ny ey Nxy gxy dx dy #  2 Z "  2 Bw Bw Bw Bw Nx Ny 2Nxy dx dy: Bx By Bx By A

1 U 2

6-20

The total strain energy expression for rectangular plates loaded laterally and inplane is the summation of expressions (6-19) and (6-20). Thus,
D U 2 "   2 2 #) 2 B2 w B2 w B2 w B2 w B w dx dy 21 A Bx2 By 2 Bx 2 By 2 Bx By A #  2 Z "   1 Bw 2 Bw Bw Bw Nx Ny 2Nxy dx dy: 2 A Bx By Bx By j U W Z (

6-21

The total potential energy of a system is given by Eq. (5-3) as

6-22

where W is external work. In order for the system to be in equilibrium, Eq. (6-22) must be minimized. Example 6-2 Find the buckling stress of a simply supported rectangular plate, Fig. 6-5, subjected to forces Nx. Solution Let the deflection be expressed as
w
l l X X m1 n1

Amn sin

mkx nky sin ; a b

which satisfies the boundary conditions.

Figure 6-5. Rectangular plate subjected to force Nx.

176

Buckling of Plates

Substituting this expression into Eq. (6-21), and noting that the expression
B2 w B2 w 21 A Bx 2 By 2 "  B2 w Bx By 2 # 0;

gives

D U 2

"  2 2 2 l l X X m k n 2k 2 mkx 2 Amn sin 2 sin 2 2 2 a b a 0 0 m1 n1 Z Z " l l 2 X X 1 b a mkx 2 m k Nx Amn 2 sin 2 sin 2 2 0 0 a a m1 m1 Z
b

# nky dx dy b # nky dx dy b

or,
 2  2 l l l l X X X X k 4 ab m n2 k 2b 2 U Amn 2 m 2 Amn : D Nx a2 b 8 8a m1 n1 m1 n1

Since there are no lateral loads, we can take the external work in the z-direction as zero and Eq. (6-22) becomes
j U:

Solving Eq. (1) for

Bj 0 BAmn

we get

 2 k 2a 2D m 2 n 2 Ncr : a2 b2 m2

The smallest value of Eq. (2) is for n = 1. Also, if we substitute

jcr Ncr =t

and
D Et 3 121 A 2 k 2E 121 A 2 b=t 2  K  m a=b 2 : a=b m

we get
jcr K

where

A plot of Eq. (4) is shown in Fig. 6-6 and shows that the minimum value of K is 4.0.

Rectangular Plates

177

Figure 6-6. Plot of Eq. (4) in Example 6-2.

Equation (3) in Example 6-2 gives the critical elastic buckling stress in a simply supported rectangular plate. The minimum value when k = 4.0 is
jcr k 2E 31 A 2 b=t 2

6-23

This equation can also be written as

jcr KD=tk=b 2

6-24

where, b= D= E= K= t= jcr = short dimension of plate E t3/12(1 A2) modulus of elasticity 4.0 thickness of plate critical buckling stress of plate

Theoretical postbuckling equations (Bloom and Coffin 2001) as well as experimental work has shown that the critical buckling stress in rectangular plates increases significantly from that of Eq. (6-24) as the buckling deflection increases. This is demonstrated in Fig. 6-7 as a function of the ratio of plate deflection, y, to thickness t. Experiments on rectangular plates, Fig. 6-8a, have shown that parts of the plate closer to the edges, Fig. 6-8b, carry significantly more load than at the central portion of the plate. Various investigators developed different expressions for the effective width when ultimate loads are used. Marguerres (Marguerre, 1937) equation is based on the terminology of Fig. 6-8c and is given by
be bjcr =ju 1=3

6-25

178

Buckling of Plates

Figure 6-7. Stress vs. deflection of a rectangular plate.

while Koiter (Koiter 1943) had a different expression in the form of

be b1:2jcr =ju 0:4 0:65jcr =ju 0:8 0:45jcr =ju 1:2

6-26

The ultimate load, Pu, carried by the plate can then be approximated by
Pu ju be t

6-27

Problems 6-4 Derive Eq. (6-15). 6-5 What is the required thickness of the plate in Fig. 6-5 if a = 40 inches, b = 15 inches, Nx = 400 lb/inch, E = 16,000 ksi, and A = 0.33? Use a factor of safety of 4.0. Use increments of 1/16 inch for thickness. 6-3 Rectangular Plates with Various Boundary Conditions

Figure 6-9 shows a plate simply supported on sides x = 0 and x = a and subjected to force Nx. The differential Eq. (6-18) becomes
j 4w N x B2 w : D Bx 2

6-28

Rectangular Plates with Various Boundary Conditions

179

Figure 6-8. Stress distribution across plate.

Let the solution be of the form

w
l X m1

f y sin

mkx : a

6-29

180

Buckling of Plates

Figure 6-9. Rectangular plate with two sides simply supported.

This solution satisfies the two boundary conditions w = Mx = 0 at x = 0 and x = a. Substituting Eq. (6-29) into Eq. (6-28) results in
d 4f d 2f A 2 Bf 0 dy 4 dy

6-30

where
A 2m 2 k 2 a2 m 4 k 4 Nx m 2 k 2 : a4 D a2

The solution of Eq. (6-30) is

f y C1 eay C2 eay C3 cos hy C4 sin hy

6-31

where

s r m 2k 2 Nx m 2 k 2 a 2 a D a2 s r m 2k 2 Nx m 2 k 2 h 2 a D a2

Values of the constants C1 through C4 are obtained from the boundary conditions y = 0 and y = b.

Rectangular Plates with Various Boundary Conditions

Case 1

181

Side y = 0 is fixed and side y = b is free. The four boundary conditions are
for y 0; deflection w 0 Bw rotation 0 By for y b; moment My 0 shear Q 0 B2 w B2 w A 2 By 2 Bx

B3 w B3 w 2 A : By 3 Bx By 2

From the first boundary condition we get

C1 C2 C3 0

and from the second boundary condition we get

aC1 aC2 hC3 0

or,
C1 C3 hC4 2 2a

and
C2 C3 hC4 : 2 2a h sinh ay a

Substituting C1 and C2 into Eq. (6-31) gives

f y C3 cos hy cosh ay C4 sin hy

With this expression and Eq. (6-29), we can solve the last two boundary conditions. This results in two simultaneous equations. The critical value of the compressive force, Nx, is determined by equating the determinant of these equations to zero. This gives
2ghg 2 h 2 cos hb cosh ab 1 a 2 h 2 h 2 g 2 sin hb sinh ab ah m 2k 2 a2

6-32

where
g a2 A

and
h h2 A m 2k 2 : a2

For m = 1, the minimum value of Eq. (6-32) can be expressed in terms of stress as
jcr k 2E 121 A 2 b=t 2 K

6-33

182
where, for A = 0.25,

Buckling of Plates

Kmin 1:328:

6-34

Case 2 Side y = 0 is simply supported and side y = b is free. Again starting with Eq. (6-27) and satisfying the boundary conditions at y = 0 and y = b, we get a solution (Timoshenko and Gere 1961) identical to Eq. (6-33) with K given by
K 0:456 b2 for A 0:25: a2

6-35

Case 3 Sides y = 0 and y = b are fixed. In this case, the minimum value of K in Eq. (6-33) is
K 7:0 for A 0:25:

6-36

Example 6-3 Let the plate in Fig. 6-5 be simply supported at x = 0 and x = a, simply supported at y = 0, and free at y = b. Calculate the required thickness if a = 22 inches, b = 17 inches, Nx = 300 lb/in, factor of safety (F.S.) = 2.0, A = 0.25, jy = 36 ksi, and E = 29,000 ksi. Solution Assume t = 0.25 inch. From Eq. (6-35) the minimum value of K = 0.456 and Eq. (6-33) becomes
jcr k 2 29; 000; 000 120:937517=0:25 2 0:456

2510 psi which is less than the yield stress:

Allowable stress = 2510/2 = 1255 psi. Actual stress = 300/0.25 = 1200 psi. o.k. Problems 6-6 Derive Eq. (6-30). 6-7 Derive Eq. (6-31). 6-8 Derive Eq. (6-32). 6-9 Determine the actual expression of K in Eq. (6-34). Plot K versus a/b for m values of 1, 2, and 3 and a/b values from 1.0 to 5.0.

Finite Difference Equations for Buckling

6-4 Finite Difference Equations for Buckling

183

The finite difference method is a powerful tool for determining the buckling of plates with irregular shapes, complicated boundary conditions, and non-uniform applied loads. The basic differential equation of the buckling of a rectangular plate as given by Eq. (6-18) is
B4 w B4 w B4 w 1 2 2 2 4 4 Bx Bx By By D  B2 w B2 w B2 w p Nx 2 Ny 2 Nxy Bx By Bx By 

6-37

where, Nx, Ny, and Nxy are entered as negative values. This equation can be transformed to a finite difference form by substituting Eqs. (5-14), (5-18), and (5-20) into the left hand side of this equation. Equations (5-12), (5-16), and (5-19) are substituted in the right hand side of Eq. (6-37). The resultant equation, after deleting the expression p, becomes
1=E4 20 wo 8wE wW wN wS 2wNE wSE wSW wNW wNN wSS wEE wWW 1=DE 2 wE 2 wo wW Nx wS 2 wo wN Ny 1=2wNE wSE wNW wSW Nxy

6-38

The following example illustrates the application of this equation in determining the buckling strength of a plate. Example 6-4 Find the buckling strength of a simply supported square plate, Fig. 6-10a, due to applied in-plane force -Nx. Let Ny = Nxy = 0. Assume a) E = 10 inches and, b) E = 5 inches. Solution a) The grid pattern for E = 10 inches is shown in Fig. 6-10b. By inspection, m = n = 1. Applying Eq. (6-38) at node o gives
1=E 4 20 wo 80 20 wo wo wo wo 1=D1=E 2 2 wo Nx

or,
Nx 8D=E 2

Substituting E = a/2 in the above expression gives the critical buckling value
Nx cr 32D=a 2

The theoretical critical value is given by

Nx cr 4k2 D=a2 which is about 19% larger than the calculated value:

184

Buckling of Plates

Figure 6-10. Finite difference layout of a square plate.

b) The grid pattern for E = 5 inches is shown in Fig. 6-10c. Applying Eq. (6-38) at nodes 3, 2, and 1 gives
1=E4 20 w3 84 w2 24 w1 01=D1=E 2 2 w2 2 w3 Nx 1=E 4 20 w2 8w3 2 w1 22 w2 w2 w2 1=D1=E 2 w3 2 w2 Nx 1=E4 20 w1 82 w2 2w3 w1 w1 w1 w1 1=D1=E 2 w2 2 w1 Nx

Other Aspects of Buckling

These three equations can be written as
2 6 6 68 K 6 4 2 20 2K 32 2K 24 2K 16 K 2 3 0 76 7 6 7 76 7 6 7 16 76 w2 7 6 0 7 76 7 6 7 54 5 4 5 0 w1 20 2K 8 32 w3 3

185

where,
K E 2 Nx =D

A non-trivial solution of these three equations is obtained by setting the determinant to zero. This gives
K 3 24K 2 160K 256 0

The roots of this cubic equation are K = 13.6569, 8.000, and 2.3431. Taking the smallest root and substituting it in the expression for K gives
Nx cr 2:3431 D=E 2

or,
Nx cr 37:49 D=a 2 :

This value is more accurate than that obtained in part a) and is about 5% smaller than the theoretical value. For rectangular plates with boundary conditions other than simply supported as well as other rectangular plates, a finer mesh is usually needed in order to improve the accuracy. Also, various values of m and n must be investigated in order to obtain minimum buckling load. 6-5 Other Aspects of Buckling

Temperature gradients are often encountered in many plate structures such as boiler casings, and aircraft wings. The plates in these structures, Fig. 6-11, are assumed to be continuous in the x-direction. They are also assumed to have no thermal gradients in that direction. The plates are also assumed to be supported by stiffeners at y = 0 and y = b. The stiffeners act as a heat sink resulting in a temperature gradient that varies in the y-direction in accordance with the equation T T0 T1 cos 2ky=b 6-39 Where, T0, and T1 are constants that define the temperature variation along the y-axis. It is assumed, Boley (Boley and Weiner, 1997) that the actual stress in the y-direction is zero and the actual stress in the x-direction is given by
jx a E T1 cos 2ky=b

6-40

186

Buckling of Plates

Figure 6-11. Continuous plate.

where, E = modulus of elasticity a = coefficient of thermal expansion. Substituting Eq. (6-40) into Eq. (6-37) and solving the resultant equation for buckling results in
jcr k k 2E t=b2 121 A 2

6-41

where,

n o k 1=2M1 2 M 2 94 4M 2 12 M 2 92 1=2 M 2 92

6-42

and
M mb=a M1 a=mb m 1; 2; . . .

Equation (6-42) is shown in Fig. 6-12. A minimum value of k can be taken as k = 3.848. The critical temperature T1cr is defined as
aE T1cr jcr 2

6-43

Substituting this expression into Eq. (6-41) gives

T1cr k k2 t=b2 6a1 A 2

6-44

Other Aspects of Buckling

187

Figure 6-12. Plot of Eq. (6-42).

Example 6-5 A rectangular plate with lengths a = 60 inches and b = 20 inches has a temperature distribution in accordance with Eq. (6-39). It has a temperature of 800jF at y = 0 and y = 20 inches. The temperature at y = 10 inches is 1000jF. Let E = 30 106 ksi, a = 6.5 10-6 in/in/jF, yield stress = 30 ksi, and A = 0.30. What is the required thickness due to this temperature distribution assuming a factor of safety of 1.3? Solution From Eq. (6-39) at y = 0,
800 To T1

1 2

and at y = b/2,
1000 To T1

Solving Equations (1) and (2) gives To = 900jF and T1 = 100jF. The actual maximum thermal stress in the plate is obtained from Eq. (6-40) as
j 6:5 106 30 106 100 19; 500 psi:

The critical buckling stress can be taken as this number multiplied by the factor of safety of 1.3.
jcr 19; 500 1:3 25; 350 psi:

Since this value is lower than the yield stress of the material, buckling controls. Substituting this value into the buckling Eq. (6-41) gives
25; 350 3:848 k 2 30 106 t=202 : 121 0:3 2

188

Buckling of Plates

solving for thickness results in

t 0:31 inch:

6-6

Application of Buckling Expressions to Design Problems

Many codes utilize the expressions of Section 6-3 to establish buckling criteria for various members. The American Institute of Steel Construction Manual (AISC 1991) assumes the buckling stress of unsupported members in compression not to exceed the yield stress of the material. Thus, Eq. 6-33 can be written as
jy k 2E 121 A 2 b=t 2 s b K : 162 t jy K

6-45

or, for steel members with A = 0.3 and E = 29,000 ksi

6-46

Equation 6-45 is based on the assumption that the interaction between the buckling stress and the yield stress in designated by points ABC in Fig. 6-13. However, due to residual stress in structural members due to forming, the actual interaction curve is given by points ADC and Eq. 6-46 is modified by a factor of 0.7 as
s b K : 114 t jy

6-47

Figure 6-13. Elastic-plastic buckling.

Application of Buckling Expressions to Design Problems 189

Single-Angle Struts Leg b of the angle shown in Fig. 6-14a is assumed free at point B. Point A is assumed simply supported because it can rotate due to deflections. Thus, Eq. 6-35 is applicable with a minimum value of K = 0.456 and Eq. 6-46 becomes
b 76 p : t jy

Double Angles Due to symmetry (Fig. 6-14b) the possibility of rotation is substantially reduced from that of case 1. Thus, the AISC uses the average of the simply supportedfree case, Eq. 6-35, and the average of the simply supported-free and fixed-free, Eq. 6-34, cases.
K 0:456 0:4561:328 2 2 0:674

Figure 6-14. Various structural members.

190

Buckling of Plates

and Eq. 6-47 is approximated by AISC as

b 95 p : t jy

Stems of Ts For this case, point A in Fig. 6-14c is assumed fixed due to the much thicker flange, and point B is taken as free. The K value is taken from Eq. 6-34 and expression 6-47 gives
b 132 p : t jy

The AISC reduces this value further down to

b 127 p : t jy

Flanges of Box Sections Points A and B in Fig. 6-14d are conservatively taken as simply supported. In this case, k = 4.0 and Eq. 6-47 becomes
b 228 p : t jy

The AISC increases this value to match experimental data and it becomes
b 238 p : t jy

Perforated Cover Plates In Fig. 6-14e, the plate between the perforation and edge is assumed fixed because additional rigidity is obtained from the continuous areas between the perforations. The dimension of the perforated plate is assumed to be a/b = 1.0. This ratio results in a K value of about 7.69. This is higher than that given by Eq. 6-36 which is based on the smallest possible value. Equation 6-47 becomes
b t s 317 : jy

Other Compressed Members Other members are assumed to have a K value that varies between 4.0 for simply supported edges and 7.0 for fixed edges. The AISC uses a value of K = 4.90. This gives
b t s 253 : jy

Application of Buckling Expressions to Design Problems 191

Another standard that uses the plate buckling equations to set a criterion is the American Association of State Highway and Transportation Officials (AASHTO 1992). The equations are very similar to those of AISC. AASHTO uses the square root of the actual compressive stress in the b/t equations rather than the square root of the yield stress. Also, limits are set on the maximum b/t values for various strength steels. A theoretical solution of the buckling of rectangular plates due to various loading and boundary conditions is available in numerous references. Two such references are Timoshenko and Gere (1961) and Iyengar (1988). Timoshenko discusses mainly isotropic plates whereas Iyengar handles composite plates. Various NASA publications are also available for the solution of the buckling of rectangular plates with various loading and boundary conditions. NASAs Handbook of Structural Stability consists of five parts and contains numerous theoretical background and design aids. Part I, edited by Gerard and Becker, includes various classical buckling solutions for flat plates. Part II, edited by Becker, is for buckling of composite elements. Part III is for buckling of curved plates and shells and Part IV discusses failure of plates and composite elements. Parts III and IV were edited by Gerard and Becker. Compressive strength of flat stiffened panels is given in Part V which is edited by Gerard.