Vous êtes sur la page 1sur 9

Truss and Cable Elements

The Finite Element Library of the MIDAS Family Programs includes the following types of Truss and Cable
Elements: general truss element, a tension-only truss element (with Hook ability), a compression-only truss
element (with Gap ability), and a cable element (equivalent truss element & elastic element of a cable with sag
catenary cable).
A typical, three dimensional, truss element is shown Fig. 1. The formulation of truss element is based on the
following assumption:
1.

The element is a straight bar of uniform cross-section.

2.

The element is capable of resisting only axial forces.

3.

At the end of element can be member end-offsets along its centroidal axis.

X1

u2
2

Z1
Y1
1

u1

Z
Y
X

Figure 1 Typical 3D Truss Element


The formulation procedure for a general truss element will be illustrated on the basis of 2D truss element. The
nodal degrees of freedom (DOFs) in local axes and the positive sign convention are shown in Fig. 2.

Z1
u1

Y1
1

2 u2

L/2

X1

Figure 2 A 2D Truss Element


The element displacement field is defined in terms of nodal DOFs by the following function:

u=

1
1
(1 ) u1 + (1 + ) u2
2
2

(1)

where

ui

displacement along local element X 1 axis at i th node

natural coordinate (see Fig.1) of the element

( 1 +1 ).

Axial displacement defined by Eq. (1) can be rewritten in terms of local x coordinate as:

u=

1 x
x
u1 + u2
L
L

(2)

In matrix form Eq. (2) can be expressed as

x
u = 1
L

x u1
= fq
L u2

(3)

where

f
q

matrix of shape functions

vector of nodal displacements

The strain-displacement relationships for the truss element consist of one derivative

= x = du =
where

vector of strains

d
B

vector of derivative operators

the strain-displacement matrix

du df
= q = Bq
dx dx

(4)

Hence, the strain-displacement matrix is presented as,

B=

df 1
= [ 1 1]
dx L

(5)

Similarly, stress-strain relationships become merely,

= x = D = E x = DBq

(6)

in which

vector of stresses

elasticity (constitutive) matrix.

Thus, the elasticity matrix is

D=E

(7)

where E is the Young modulus


Then, the element stiffness matrix of a prismatic truss element can be evaluated as follows:

K = BT DBdV =
V

E
L2

L
1
EA 1 1
1 [ 1 1] 0 A dAdx = L 1 1

(8)

The stiffness matrices for a tension-only truss element and a compression-only truss element are the same as
above. But, they can or cannot exhibit the stiffness depending on the condition defined by the sign of the
member forces or relative displacements.

Z1

Y1

X1

X1

Gap

Z1

Y1

Hook

Figure 3 Truss Element with Hook and Gap Propetries

Thus, the tension-only truss element is capable of resisting only an axial tension force. Therefore, if a member
force is positive (tension), it exhibits the stiffness previously described, and if a member force is negative
(compression), its stiffness becomes zero. Hook ability means that the element has a constant initial distance

h , named as hook distance (see Fig.3); so only when the relative displacement is larger than or equal to h ,
the element exhibits its stiffness. That is,

( q 2 q1 ) h

(9)

Conversely, a compression-only truss element transfers only an axial compression force. Therefore, only when
a member force is negative, it exhibits the stiffness of Eq. (8). Similarly, it has Gap ability, so only when the
relative displacement is smaller than or equal to a constant initial distance g , termed as gap distance (see
Fig. 3), it exhibits its stiffness. That is,

( q2 q1 ) g

(10)

A typical, cable-equivalent truss element is shown Fig. 1. Te cable element, however, is capable of resisting
only the axial tension force and is used to model structural behavior of the cable. Basic property of the cable is
known as stress stiffening due the tension force applied to the element. It should be noted, that this element
is formulated in such way, that when it is used in linear analysis it is automatically transformed in an
equivalent truss element, and in geometric nonlinear analysis it acts as an elastic catenary cable element.
The stiffness of the equivalent truss element consists of a general elastic stiffness and the stiffness due to sag.
The cable element has the same nodal DOFs and forces as a general truss element and general elastic stiffness
is the same as in Eq. (8)

K elastic =

EA
L

while the stiffness due to sag is given by,

K sag =

12T 3
w2 L3

(11)

where, T is the tension force in the cable element, and w is the self-weight per unit length of the cable.
If we combine the elastic stiffness and the stiffness due to sag in series, the stiffness of the combined cableequivalent truss becomes,

K comb =

1
=
1 K elastic + 1 K sag

EA
w2 L2 EA
L 1 +

12T 3

(12)

The elastic catenary cable element is based on Lagrangian formulation and accounts for geometric nonlinear
effects caused by large displacement. Basic procedure in the formulation of this element is evaluation of
tangent stiffness matrix. For the implemented cable element the tangent stiffness matrix is obtained as follows.
Lets consider a 2-node cable, located in the fixed coordinate system x, y, z, as shown in Fig. 4. Assume that in

initial equilibrium configuration cable is subjected to six nodal forces F1 , F2 , F3 , F4 , F5 , F6 and his
initial properties are defined as follows: section area A0 , length L0 , and self-weight per unit length (weight
density) w0 .
Assume that cable is subjected to a displacement field defined by the six displacement components: u1, u2 and

u3 at the node i and u4, u5 and u6 at the node j, accompanying the nodal forces F1, F2, F3, F4, F5 and F6.
The Lagrangian (curvilinear) coordinate of an arbitrary point P (x, y, z) on a cable element is s for the
unstretched length, and p for the stretched length (see Fig. 4).

Figure 4 A Catenary (Cable) Element subjected to nodal displacements


In this configuration of the cable element, the geometric constraints and the equilibrium conditions for tension
forces are,

dx dy dz
+ + = 1
dp dp dp

(13)

dx
T = F1
dp
dy
T = F2
dp

(14)

dz
T = F3 ws
dp
Accordingly, the tension force T at point P is given by,

T ( s ) = F12 + F22 + (F3 + ws )

2 1/ 2

(15)

The relationships between the undeformed Lagrangian coordinate s and Cartesian coordinate are,

dx
ds
ds
dy
y ( s ) = ds
ds
dz
z ( s ) = ds
ds

x (s) =

(16)

where,

dx dx dp
F dp
F T
F1
=
= 1
= 1
+ 1 = 1 +
ds dp ds
T ds
T EA0
EA0 F12 + F22 + (F3 + ws )2

1/ 2

(17)

and the boundary conditions are,

x = 0, y = 0, z = 0, p = 0 at s = 0
x = l x , y = l y , z = l z , p = L at s = L0

(18)

Therefore,

[{
(
) } ln{F + (F + F + F ) }]
F
F
s [ln{F + ws + (F + F + (F + ws ) ) } ln{F + (F + F + F ) }]
y (s ) =
w
EA
F
1
F
[{F + F + (F + ws ) } {F + F + F } ]
s
z (s ) =
2 EA w
EA
x(s ) =

F1
F
2
s 1 ln F3 + ws + F12 + F22 + (F3 + ws )
EA0
w
2

2
1

2
2

2
2

1/ 2

2
1

2
2

2 1/ 2
3

2
1

2
2

2 1/ 2
3

2 1/ 2

2
1

2 1/ 2

2
1

2
2

2 1/ 2
3

(19)

and,

[{
(
) } ln{F + (F + F + F ) }]
FL F
l =
[ln{F + wL + (F + F + (F + wL ) ) } ln{F + (F + F + F ) }]
EA
w
FL
wL
1
l =

[{F + F + (F + wL ) } {F + F + F } ]
EA 2 EA w
lx =
y

F1 L0 F1
2
ln F3 + wL0 + F12 + F22 + (F3 + wL0 )
EA0 w
2

2
1

2
2

1/ 2

2 1/ 2

2
1

2
1

2
2

2
2

2 1/ 2
3

2 1/ 2
3

(20)

3 0
0

2
0

2
1

2
2

2 1/ 2

2
1

2 1/ 2
3

2
2

The equilibrium conditions of the nodal forces and the compatibility conditions of nodal displacements are,

F4 = F1
F5 = F2
F6 = F3 w0 L0

(21)

lx = lx0 u1 + u4 = f ( F1 , F2 , F3 )
l y = l y0 u2 + u5 = g ( F1 , F2 , F3 )
lz = lz0 u3 + u6 = h ( F1 , F2 , F3 )

Differentiating both sides of Eq. (21), the relationships between nodal forces and changes of cable lengths can
be presented as:

dlx =

f
f
f
dF1 +
dF2 +
dF3
F1
F2
F3

dl y =

g
g
g
dF1 +
dF2 +
dF3
F1
F2
F3

dlz =

h
h
h
dF1 +
dF2 +
dF3
F1
F2
F3

(22)

or in matrix form

dlx
dF1

dl y = F dF2

dlz
dF3

(22a)

where F is, so called, nodal flexibility matrix defined as follows:

F1
g
F=
F1
h

F1

f
F2
g
F2
h
F2

F3
f11
g
= f 21
F3
f 31
h

F3

f12
f 22
f 32

f13
f 23
f33

(22b)

Rewriting Eq. (22a) as

dF1
dlx


1
dF2 = K dl y K = F
dF
dl
3
z

(22c)

were K is the nodal stiffness matrix.


The components of the flexibility matrix in Eq. (22b) are

L
f
1
= 0 ln { F3 + wL0 + B} ln { F3 + A}
F1
EA0 w

f11 =

1
1
2
2

B + ( F3 + wL0 ) B A + F3 A

1
1
FF
f
f12 =
= 1 2 2
2

w B + ( F3 + wL0 ) B A + F3 A
F2
F12

f13 =

F +A
F F + wL0 + B
f
= 1 2 3
23

w B + ( F3 + wL0 ) B A + F3 A
F3

f 21 =

g
= f12
F1

f 22 =

L
g
1
= 0 ln { F3 + wL0 + B} ln { F3 + A}
F2
EA0 w

f 23 =

F22
w

1
1
2
2

B + ( F3 + wL0 ) B A + F3 A

g F2
f13
=
F3 F1

F 1 1
h
= 1
w B A
F1
h F2
f32 =
f 31
=
F2 F1
f31 =

f33 =

(23)

L
h
1 F + wL0 F3
= 0 3

EA0 w
B
A
F3

where,

A = F12 + F22 + F32

12

, B = F12 + F22 + ( F3 + wL0 )

Thus, the tangent stiffness is obtained as follows:

{dF } = K T {du}

2 12

where

F1
u
1
F2
u
1
F3

u1

KT =
F1

u1
F2

u1
F
3
u1

F1
u2

F1
u3

F1
u4

F1
u5

F2
u2

F2
u3

F2
u4

F2
u5

F3
u2

F3
u3

F3
u4

F3
u5

F
1
u2

F
1
u3

F
1
u4

F
1
u5

F2
u2

F2
u3

F2
u4

F2
u5

F3
u2

F3
u3

F3
u4

F3
u5

F1
u6

F2
u6

F3

u6 K ii
=
F1 K ii

u6
F
2
u6
F
3
u6

K ij
K ij

(24)

However, since dl x = du1 = du4 , dl y = du2 = du5 , dl z = du3 = du6 ,

K ii = K and

K ij = K

(25)

K
K

(26)

and, therefore tangent stiffness matrix can be presented as

K
KT =
K

It should be noted that when using the catenary cable element, along with section area and modulus of
elasticity the weight density of the cable must be specified. This cable properties are used to define geometry
of the cable and cable tension in initial configuration.