Sie sind auf Seite 1von 12

51st AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference<BR>18th

12 - 15 April 2010, Orlando, Florida

AIAA 2010-2607

Finite Element Analysis of Cable-truss Structures


Lin Liao1
AEROS, Montebello, CA, 90640
Baisong Du2
School of Civil Engineering & Architecture, Chongqing Jiaotong University, Chongqing, China, 400074

[Abstract] Finite element techniques are applied to analyze truss structures containing
pretensioned cables. This paper focuses on the study of the influence of cable property and
pretension on trusses and the variation of cable tension in undeformed and deformed
configurations. This work shows that cable pretension, elastic modulus, and cross-sectional
area have significant effect on truss performances. Stability and optimization associated with
structural design of trusses are briefly discussed. The paper presents some case studies of
cable-truss structures, in which commercial FE packages ANSYS and NASTRAN have been
utilized.

Nomenclature
E

A
X
Y
Z
Ux
Uy
Uz

Fx
Fy
Fz
Ci
Ni

Elemi

=
=
=
=
=
=
=
=
=
=
=
=
=
=
=

elastic modulus
Poissons ratio
cross-sectional area of
X coordinate
Y coordinate
Z coordinate
X component of displacement
Y component of displacement
Z component of displacement
applied force in X direction
applied force in Y direction
applied force in Z direction
the ith cable tension force
the ith node
the ith truss element

I. Introduction
Truss structures have attracted tremendous interests due to their extensive applications in the construction of
infrastructures and space structures. Considerable efforts have been placed on the development and analysis of truss
bridges composed of concrete and steel. Research works have been focused on material characteristics, truss joint
design, and processing and construction of structural components in the field of civil engineering. Researchers have
carried out investigations of composite trusses for aerospace applications [1-3]. These space truss structures are
distinctively different from civil structures regarding materials, strength, stiffness, and weight. In recent years, the
overall performances and characteristics of composite truss structures have been studied [4, 5].
Cables serve as essential and important members of truss structures. Cables can only withstand tension forces and
are utilized to maintain stability and strength of truss systems. In this paper, cable is used to denote the structural
members that only support tension loads, which can also be called rope, wire rope, string, chain and etc. Cable
tension has been considered as a critical factor in optimal design of trusses and prediction of truss performances. The
1
2

PhD, Aeronautical Engineer, AEROS, Montebello, CA, AIAA Member.


PhD, Associate Professor, School of Civil Engineering & Architecture, Chongqing Jiaotong University, China.
1
American Institute of Aeronautics and Astronautics

Copyright 2010 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.

arrangement of cable configuration and determination of cable pretension are closely associated with load
conditions, application environment, and structural characteristics. Since cables are tension-only structural members,
nonlinear analysis was applied to study the erection of space trusses by cable tensioning technique [6]. A two-step
optimum method was developed by Zhang and Nai to optimize cable pretension, cross-sectional area, diameter, and
thickness [7]. Minimum cost was chosen as optimal objective, and stability and displacements were applied as
constraints. Multi-objective optimization technique was proposed for a cable-truss deployable space antenna [8]. In
the approach truss thicknesses and cable tensions were design variables, and minimum weight and surface accuracy
were the objective functions.
Besides the aforementioned investigations on cable-truss optimization, the researches primarily concentrated on
cable tension or prestress have been reported in the literature too. The effects of prestressed Kevlar/FRP cables in
composite structural systems were studied [9]. Experimental study and numerical analysis have been carried out on
composite space trusses with prestressed cables made of steel and compression members made of concrete [10, 11].
Zhang and his co-workers conducted research on the impact factors of cable tension of crescent-shaped multi-rib
concrete filled steel tube truss arch bridges [12]. The impact effects were caused by road surface roughness, vehicle
speed, and structural damping ratio. Li developed a method to determine cable prestress of truss-string structures
[13]. Cable prestress was realized by increasing temperature loads in the model of construction state and cable stress
and controlling point displacement were obtained correspondingly for this state. Liu introduced an approach for the
analysis of internal forces of cable-stayed space grids with multi-step-loading [14]. The stayed cables were
simulated as equivalent linearization elements. A method for pretension design of cable-net structures of axissymmetric parabolic antenna was proposed based on force balance equations of cable nodes and structural
characteristics [15].
Based on the authors literature survey, the interaction between cables and truss members and the effects of cable
property and configurations on truss behavior have not been studied. This paper is aimed to conduct comprehensive
study of truss structures containing cables and truss members. First, finite element analysis (FEA) of cable-truss
structural systems is introduced. Hence, the effect of cable property and pretension on deformation of trusses having
diverse configurations will be covered. A variety of truss structures contained pretension cables will be modeled and
analyzed using commercial finite element (FE) packages ANSYS/NASTRAN. The authors discuss design
considerations of cable-truss structures. In particular, the stability and optimization problems will be addressed. This
paper is concluded with a summary in the last section.

II. Finite Element Analysis of Cable-truss Structures


Cable-truss structures having simple configurations and fewer members can be solved analytically. As for
complex truss systems we have to resort to FEA for efficient and accurate solutions. The construction of FE model
for cable-truss structures is similar to conventional approach of FEA. Usually, 1D element is used to model truss
rods and cables since they can be considered as slender structures. Based on the characteristics of structural
members (whether the members can support bending loads), they can be simulated as beams (compression-tensionbending-twisting element), rods (compression-tension element), and cables (tension-only element). Equivalent
material properties could be applied for composite truss beams in the analysis of complex truss systems. Cables
serve as crucial members for generating determined truss structural systems. Cable pretension can be prescribed by
initial strains or prestress and is introduced in different ways in various FE packages. In this study the authors only
discuss the modeling of pretensioned cables in ANSYS and NASTRAN. It is noted that other FE software such as
ABAQUS, COMSOL can also deal with tension cables with specific build-in elements. Different FE codes have
their unique way of handling cable characteristics.
Thermal loads in terms of temperature differential can be applied to cable elements to achieve pretension effect
while truss elements are not subjected to thermal loads. According to the equation of thermal effect, F=EAT,
thermal loads can be converted to cable pretension. Generally, thermal expansion coefficients of cables could be
selected arbitrarily and dont affect the result of simulation. The exclusion of thermal loads of truss elements can be
realized by setting either temperature differential or thermal expansion coefficients zero. CROD element in
NASTRAN can be used to simulate cables. NASTRAN linear solver SOL101 can be employed for this kind of
analyses. However, compressive forces could be generated for cable elements in deformed states using linear solver.
In such cases, a second analysis is needed by deleting cable elements having compressive forces. Alternatively,
NASTRAN nonlinear solver SOL106 is applicable to this problem. Cables are assigned nonlinear material property:
stiffnesses vary in tension and compression range. Resultant cable forces will be equal to or larger than zero using
this solver.

2
American Institute of Aeronautics and Astronautics

The second method is to employ tension-only elements if they are available in existing FE software package.
According to the authors knowledge, ANSYS offers the straightforward solver for cables analysis. ANSYS
LINK10 elements are capable of dealing with only tension or only compression by assigning certain parameters for
cable elements. Since this kind of FE models involves tension-only elements, nonlinear iterative solver of ANSYS is
applied and sub step settings need to be specified for convergent study. This approach does not lead to compressive
forces in cable elements. However, the solution accuracy and convergence need to be prudently checked.
In this paper, both ANSYS and NASTRAN have been implemented into the analysis of cable-truss structures.
There are diverse types of cable-truss structures and the current work is focused on specific structures with the
assumptions listed below
1. Truss members are assumed to be perfectly connected at joints.
2. Cables are connected to truss rods at joints. There is no friction sliding of cables with respect to truss joints.
3. Small strain and displacement are considered. Geometrical nonlinearity is not accounted in this study.
4. Material nonlinearity of trusses is not considered.
5. Loads are applied at joints except gravity loads. Truss members only support tension/compression loads. All
trusses are axially loaded at the ends (truss joints) and have uniform property along the length of each member. It
should be noted that truss members can be designed to beam-type structures, which hold tension, compression, and
bending loads. This type of structure is beyond the scope of this work.

III. Effect of Cable Property on Truss Performance


In this section, the effect of cable property on truss performances will be studied using ANSYS. Parameters used
to specify cables in FE models include elastic modulus, cross-sectional areas (diameters), and density.
An example of 3D Truss A (see Figure 1) is given
below to illustrate the influence of cable material
property. Blue lines represent truss elements and green
lines represent cable elements. Totally, there are nine
truss members and six cables. Nodal coordinates of all
truss joints in this example and subsequent examples are
given in Table 1. Original cable and truss properties are
provided in Table 2. The four nodes 14 of Truss A are
constrained. The forces applied at Node 5 and 6 are Fz=50 lbs. The deformation of Node 5 and Node 6 is the
same for this truss under the prescribed loading. Table 3
presents the variation of total displacement (magnitude of
total displacement, absolute value) of Node 5 or Node 6
with respect to elastic modulus when cable pretension is
Figure 1. Schematic of Truss A.
kept constant (300 lbs). It can be seen that truss maximum
displacement decreases along with the increase of cable
elastic modulus. Under the condition of constant pretension, stiff cables generate less deformation than flexible
cables. Although cables with high stiffness can reduce truss deformation, there are drawbacks of using stiff cables.
Stiff cables tend to lose tension forces in contrast with cables with low elastic modulus, which will be studied in
details in Section V.
Next, we study the effect of cable geometry on truss deformation. Table 4 shows the variation of Node 5 and
Node 6 displacement due to cross-sectional areas when cable pretension (100 lbs) and elastic modulus (1E6 psi) are
constants. The increase of cable cross-sectional areas results in the decrease of displacements. Industrial cables most
often consist of a few strands of small fibers or metal ropes (such as 6, 7, or 8 strands). The bulk area provided in
cable specification is usually larger than the actual cross-sectional area. Real cross-sectional area used in FE models
needs to be calculated by taking out the gaps between strands of ropes. Additionally, nonstructural masses
associated with cable adjustment and installations need to be accounted in the construction of FE models for
structural systems. Some FE programs provide the option of nonstructural masses.
Table 1. Nodal coordinates of Truss A-H.

X (in)
Y (in)

N1
-10
-10

N2
10
-10

N3
10
10

N4
-10
10

N5
0
-10

N6
0
10

N7

3
American Institute of Aeronautics and Astronautics

N8

N9

N10

Z (in)
X (in)
Y (in)
Z (in)
X (in)
Y (in)
Z (in)
X (in)
Y (in)
Z (in)
X (in)
Y (in)
Z (in)
X (in)
Y (in)
Z (in)
X (in)
Y (in)
Z (in)
X (in)
Y (in)
Z (in)
X (in)
Y (in)
Z (in)
X (in)
Y (in)
Z (in)
X (in)
Y (in)
Z (in)
X (in)
Y (in)
Z (in)

0
250
0
90
-100
0
200
-19.02
6.18
0
0
0
0
0
0
0
0
0
0
0
0
0
N11
-375
216.5
0
30
0
30
69.28
0
40
69.28
30
0

0
125
216.5
90
100
0
200
-11.76
-16.18
0
30
0
0
34.64
0
0
34.64
0
0
40
0
0
N12
-375
-216.5
0
60
0
30
73.92
17.32
40
103.92
30
0

0
-125
216.5
90
-100
100
200
11.76
-16.18
0
60
0
0
69.28
0
0
69.28
0
0
40
30
0
N13
0
0
140
90
0
30

0
-250
0
90
100
100
200
19.02
6.18
0
90
0
0
103.92
0
0
103.92
0
0
0
30
0
N14

10
-125
-216.5
90
-100
0
0
0
20
0
120
0
0
13.92
51.96
0
143.92
0
0
-60
0
40
N15

10
125
-216.5
90
100
0
0
-9.51
3.59
8
120
30
0
43.92
34.64
0
34.64
0
20
-40
30
40
N16

90
30
30

60
30
30

30
30
30

143.92
30
0

34.64
30
20

69.28
30
40

103.92
30
60

0
-433
0
-100
100
0
-5.88
-7.59
8
90
30
0
73.92
17.32
0
69.28
0
40
0
0
40

375
-216.5
0
100
100
0
5.88
-7.59
8
60
30
0
103.92
0
60
103.92
0
60
0
30
40

375
216.5
0
-250
0
100
9.51
3.59
8
30
30
0
34.64
0
20
0
30
0
40
0
40

0
433
0
250
0
100
0
10.5
8
0
30
0
43.92
34.64
20
34.64
30
0
40
30
40

Table 2. Material property.

truss
cable

E (psi)
1.5E7
1E6

A (sq in)
1.0E-1
1.0E-2

initial strain
0
1.0E-2

Table 3. Variation of nodal displacements with respect to cable material properties.


E (psi)
Initial strain
Displacement (in)

1E6
3.000E-2
0.4269E-02

3E6
1.000E-2
0.4229E-02

4E6
0.750E-2
0.4209E-02

8E6
0.375E-2
0.4131E-02

12E6
0.250E-2
0.4056E-02

Table 4. Variation of nodal displacements with respect to cable cross-sectional areas.


A (sq in)
Initial strain
Displacement (in)

1E-2
1.000E-2
0.4229E-02

2E-2
0.500E-2
0.4169E-02

4E-2
0.250E-2
0.4056E-02

5E-2
0.200E-2
0.4001E-02

4
American Institute of Aeronautics and Astronautics

8E-2
0.125E-2
0.3847E-02

IV. Cable Pretension and Truss Deformation


Cables are usually loaded with pretension when they are assembled with truss members. The mutual
reinforcement between truss members and tensioned cables enhances overall structural strength. Cable pretension is
closely linked to truss deflection and element forces of truss members. The study of pretension and structural
strength is significant and informative for truss structural design.
A. Effect of cable pretension on deformation
Two numerical examples are given below to study the influence of cable pretension on deflection. The first
example is a 3D dome-shape truss with pretensioned cables (see Figure 2). The properties of truss elements are
E=2.5E7 psi; =0.33; A=4.0 in2; and the properties of cables are
E=3E6 psi; =0.25; A=0.05 in2. An external force of 500 lbs is
applied at Node 7 along positive Z direction. The six bottom
nodes (Node 8Node 13) are constrained.
Table 5. Nodal displacements of Truss B.
Pretension (lbs)
150
300
450
600
750

Displacement Uz (in)
4.094E-3
-9.862E-4
-6.066E-3
-1.115E-2
-1.623E-2

Figure 2. Schematic of Truss B.


Table 5 shows the variation of the Node 7 displacement in Z-direction with respect to cable pretension. When
Truss B is subjected an external load in positive Z direction, a small value of cable pretension 150 lbs causes a
positive displacement Uz at Node 7. The absolute value of displacements decreases when pretension increases from
150 lbs to 300 lbs. The increase of cable pretension up to 450 lbs leads to a negative displacement Uz at Node 7.
When the cable pretension is larger than 450 lbs, the displacement continues to increase. Apparently, Truss B has a
minimum deformation (the absolute value of displacements) when cables are given a pretension force around 300
lbs.
The second example is Truss C shown in Figure 3.
The properties of truss element are E=2.5E7 psi; =0.25;
A=5.0 in2; and the properties of cables are E=3E6 psi;
=0.25; A=0.02 in2. The bottom six nodes (Node 12, 5
6, 910) are restrained. Fy=200 lbs forces are applied to
Nodes 3, 4, 7, 8. The displacements of Node 3 and Node
8 in three directions are listed in Table 6. It can be seen
that the absolute values of Ux & Uz displacements
increase due to the increase of pretension. It is interesting
to know that the Y displacement changes from positive to
negative. It shows that the top portion of truss tends to
Figure 3. Schematic of Truss C.
deflect upward when pretension is much smaller than
applied loads; Node 3 and Node 8 deflects downward when pretension is close to or larger than applied loads.
Appropriate selection of cable pretension can minimize the deformation due to external loads.
Table 6. Displacements of Truss C.
Pretension (lbs)
60
120
180
240

Displacement of Node 3 (in)


Ux
Uy
Uz
-0.5895E-04
0.9042E-04
-0.8556E-04
-0.1363E-03
0.3316E-04
-0.1588E-03
-0.2137E-03 -0.2409E-04 -0.2320E-03
-0.2911E-03 -0.8135E-04 -0.3053E-03

Displacement of Node 8 (in)


Ux
Uy
Uz
0.5895E-04
0.9042E-04
0.8556E-04
0.1363E-03
0.3316E-04
0.1588E-03
0.2137E-03
-0.2409E-04
0.2320E-03
0.2911E-03
-0.8135E-04
0.3053E-03

5
American Institute of Aeronautics and Astronautics

The above examples show that truss deformation is not linearly proportional to cable pretension. Final
configurations of deformed trusses are influenced by cable pretension. To minimize truss deformation, it is
important to select suitable cable pretension. Determination of cable pretension is relevant to loading conditions and
cable property. Both theoretical prediction and practical experiences are used to specify cable pretension. Usually,
skillful workers are able to adjust cable pretension by observing the structural behavior.
B. Effect of cable pretension on truss element forces
Cable pretension not only has effect on truss deformation but also affect truss element forces. The study of the
variation of truss element forces due to cable pretension is useful for design and analysis of truss systems, since
element force is one of the key considerations of truss geometrical and strength design. Table 7 displays the
variation of element forces of Truss A subjected to different cable pretension. The data shows that truss element
forces are significantly increased when cable pretensions become large. We increase cable pretension in order to
reduce structural deformation, and the strength of individual truss members need to taken account at the same time.
Too much pretension might lead to excessive truss element forces, and thus enhances the demand of strong truss
materials.
Table 7. Variation of element forces with respect to cable pretension.
Pretension (lbs)
100
200
300
400

Elem 5 (lbs)
-92.76
-150.22
-207.67
-265.12

Elem 6 (lbs)
-92.76
-150.22
-207.67
-265.12

Elem 7 (lbs)
-92.76
-150.22
-207.67
-265.12

Elem 8 (lbs)
-92.76
-150.22
-207.67
-265.12

Elem 9 (lbs)
-162.37
-324.87
-487.38
-649.88

In this section, influence of pretension on structural performance of cable-truss structures is addressed. Cable
pretension can be favorable or unfavorable for the overall structure, depending on load conditions and structural
configurations. Although increase of pretension helps to reduce truss deformation, it also increase truss member
forces. The results indicate that appropriate and optimal positioning and pretensioned loading of cables have
significant influence on the performance of trusses. At the same time of increasing cable pretension, we should
consider the stiffness of truss members.

V. Variation of Tension Forces in Undeformed and Deformed Configurations


Before the structures are subjected to external loads, each cable is loaded with the same pretension in undeformed
configurations. Constant cable pretension can lead to variable tension forces in cables in deformed truss structures.
Cable tension after deformation might be larger or smaller than
pretension. Cables might lose all preloaded tension in truss deformed
states. A few examples are presented below to study the change of
cable tension subjected to loads.
Truss D (see Figure 4) consists of 15 truss members and 15 cables
(the number of cables is shown by green circles). The properties of
truss element are E=2.5E7 psi; =0.25; A=0.15 in2; and the properties
of cables are E=3E5 psi; =0.25; A=0.04909 in2. A force Fz=-1000 lbs
is applied at each of the five middle nodes (Nodes 6-10). Five bottom
nodes (Nodes 1-5) are constrained. All cables are assigned a pretension
of 147.26 lbs. Table 8 shows that 15 cables are still subjected to tension
loads and cable tensions in deformed configuration vary from
pretension. It can be seen that cable tension forces range from 85.22 lbs
to 196.28 lbs, which is around 58% to 133% of pretension. The applied
loads cause reduction of tension in some cables as well as increase of
Figure 4. Schematic of Truss D.
tension in some cables.
Table 8. Cable tension in deformed configuration of Truss D.
Cable
Tension (lbs)
Cable
Tension (lbs)

C1
176.89
C9
196.28

C2
107.35
C10
85.22

C3
176.89
C11
141.22

C4
107.35
C12
141.22

C5
196.28
C13
142.04

C6
85.22
C14
141.68

6
American Institute of Aeronautics and Astronautics

C7
142.79
C15
141.68

C8
142.79

As displayed in Figure 5, Truss E consists of 30 truss members and 30 cables. Elastic modulus for truss members
and cables are 6E5 psi and 1E5 psi, respectively. Poissons ratio of both materials is 0.33. Cross-sectional areas of
truss rods and cables are 0.8 in2 and 0.1 in2. Truss E is supported at six nodes: Nodes 2-4, 7-9. The prescribed loads
are Fz=-500 lbs at each of the six nodes 11-16, and Fz=500 lbs at Node 1 and Node 10. A pretension of 120 lbs is
applied to each cable. The cable tension in the deformed state is listed in Table 9. It can be seen that substantial
changes of cable tension occur after deformation. Some cables lose tension while some cables gain double value of
pretension. The cables 23, 25, 27, and 29 lose tension and compressive forces are generated for the first run.
NASTRAN input file was modified by deleting these cables and rerun again.
The undeformed and deformed configurations are displayed in Figure 5. Blue lines represent undeformed state
and black lines represent deformed state. The deformation of rectangular frame formed by Nodes 2,3,11,12 causes
Cable 24 gains much larger tension than pretension, and Cable 23 completely lose tension. Similarly, the same
effects occur to the other three pairs of cables: Cables 25-26, Cables 27-28, Cables 29-30. It can be understood that
one of the two cross-cables becomes shorter and the other becomes longer when the rectangular frame was altered to
a parallelogram. To identify which cable in the structure will lose tension under specific load cases is a complicated
problem and need high fidelity prediction of deformed state and element forces. Generally, cables with high
modulus are easier to lose tension than those with low modulus. Stiff cables might completely lose pretension in
deformed configurations. Although Flexible cables are good for maintaining tension, they are are not helpful for
deformation minimization compared with cables with high modulus for the same configuration and load conditions,
as discussed in Section IV.

Figure 5. Schematic of Truss E (left: schematic, right: undeformed and deformed configurations).
Table 9. Cable tension in deformed configuration of Truss E.
Cable
Tension (lbs)
Cable
Tension (lbs)
Cable
Tension (lbs)
Cable
Tension (lbs)

C1
139.75
C9
75.58
C17
110.54
C25
0

C2
139.75
C10
75.58
C18
110.54
C26
344.83

C3
120
C11
91.777
C19
67.77
C27
0

C4
120
C12
91.77
C20
67.77
C28
362.28

C5
120
C13
55.21
C21
68.25
C29
0

C6
120
C14
55.21
C22
68.25
C30
344.83

C7
111.24
C15
106.97
C23
0

C8
111.24
C16
106.97
C24
362.28

Next, Truss F is given as another example (see Figure 6). The properties of truss rods are E=8E5 psi; =0.3;
A=0.8 in2 and those of cables are E=1E6 psi; =0.3; A=0.1 in2. The displacements of Nodes 1, 4, and 5 are
restrained. Fx=-400 lbs is applied to Nodes 9, 10, 11, and 12. The pretension of each cable is 100 lbs. Undeformed
and deformed configurations are shown in Figure 6. The resulting cable tensions in the deformed state are listed in
Table 10. Cable tensions vary from 0 to 359.01lbs. The four cables (5, 8, 10, 15) dont sustain any loads in deformed
truss, which be explained similarly to that of truss E.
Although 4 out of 18 cables are completely released from tension loads for this specific load case, it is not always
like this for other load cases. To investigate the influence of loading conditions on cable behavior, we rerun the
program and change the loads at Nodes 9, 10, 11, and 12 to Fx=-200 lbs, and the same pretension for each cable is
used. The resulting tensions of cables are displayed in Table 11, which are significantly different from the data in
Table 10. When the external loads are decreased, tension forces are produced in all cables. It is reasonable to deduce
7
American Institute of Aeronautics and Astronautics

that cables might regain tension or lose tension if the structure is transferred from one load case to another load case.
Diverse loading conditions lead to considerable variation of cable tension with respect to pretension. As far as
structural design is considered, we need to solve for a variety of load cases and find the maximum possible cable
tension, which will be used in the selection of cables. When we design truss structures sensitive to cable tension, it is
necessary to install devices for cable tension adjustment and structural health monitoring.

Figure 6. Schematic of Truss F (left: schematic, right: undeformed and deformed configuration).
Table 10. Cable tension in deformed configuration of Truss F.
Cable
Tension (lbs)
Cable
Tension (lbs)

C1
26.18
C10
0

C2
147.11
C11
94.04

C3
85.59
C12
48.06

C4
76.44
C13
121.68

C5
0
C14
86.75

C6
214.56
C15
0

C7
248.60
C16
359.01

C8
0
C17
4.14

C9
137.53
C18
255.12

C8
8.51
C17
47.28

C9
97.77
C18
167.55

Table 11. Cable tension in deformed configuration of Truss F.


Cable
Tension (lbs)
Cable
Tension (lbs)

C1
66.61
C10
48.72

C2
110.72
C11
84.41

C3
85.24
C12
65.91

C4
84.23
C13
100.20

C5
7.64
C14
89.11

C6
132.29
C15
29.96

C7
135.24
C16
196.89

In this section, a few numerical examples are presented to show that variation of cable tension in contrast to
pretension. Final tensions of cables are mostly determined by structural deformation and applied loads. Whether
cables are stretched or released pertains to external loads and vary case by case. At one hand the presence of cables
contributes to the strength and integrity of truss structures, while at the other hand cables allow for structural
recovery flexibility: cables accommodates diverse loading by changing tension forces. The breaking strength of
cable materials must meet the requirement of maximum possible tension forces with the consideration of a safety
factor.

VI. Discussion of Design Considerations


The objective of truss design is to achieve stable, strong, and light weight structures. For cable-truss structures,
there are some crucial design considerations associated with unique characteristics of cables. Some examples are
presented to address these factors.
Truss G is shown in Figure 7. The properties of truss rods are E=8E5 psi; =0.3; A=0.8 in2 and those of cables are
E=1E6 psi; =0.3; A=0.1 in2. The displacements of Nodes 1-5, and 9-13 are restrained. Fz=-100 lbs is applied to
Nodes 7, 8, 15, and 16. The pretension of each cable is 100 lbs. Three design architectures are presented in Figure 8.
Design A is not stable due to the lack of connection between truss members along Y direction. If six cables are
included, we can obtain a stable structure: Design B. Four truss members are replaced by two cables in Design C.
The design architectures B & C are stables because the placement of cross cables between two XZ planes in the
truss. Generally speaking, a parallelogram is not stable, and a triangular shape whether formed by truss rods only or
both truss rods and cables generates a stable structural system, and the loads at joints can be transferred loads among
more than three members. The displacements for two design architectures B&C are compared in Table 12.
8
American Institute of Aeronautics and Astronautics

The weight efficiency is always a big concern


of designing space truss structures that are
mainly used in the field of aerospace engineering.
It is known that linear weight density (weight per
unit length) of truss members is usually much
larger than that of cables. To reduce the weight
of overall system, truss rods can be replaced by
cables under the conditions that same
performances can be achieved. Design C (shown
in Figure 8) reduces 4 numbers of truss rods by
using 8 cables. The weights of B and C are 107.9
lbs and 94.3 lbs, respectively. The structural
Figure 7. Schematic of Truss G.
behavior of these two configurations subjected to
same loading conditions is compared in Table 12&13. As shown in Table 12, the displacements of Nodes 6, 7, 8 for
Design C are larger than those of Design B. Truss element forces are provided in Table 13. It can be seen that
weight reduction is achieved by substituting truss rods with tension-only cables with the increase of deformation and
element forces as trade offs.
Substituting more truss rods with cables is beneficial for weight reduction. The key issue is to obtain an optimal
solution with the combined consideration of weight, deformation, and truss strength. The objective is to accomplish
sufficient weight reduction while maintain the acceptable margin of deformation and strength. When we replace
truss members with cables, the problem of cable losing tension should be accounted. Unlike truss rods, cables can
not support compressive loads. Thus, we need to make sure that the replacement of truss rods will not affect
structural integrity for all load cases in application.
Table 12. Comparison of displacements of two design architectures of Truss G (in).
Design
B
C

Node 6
Ux
Uy
Uz
-2.46E-3 1.79E-3 -2.30E-3
SUM: 3.81E-03
2.18E-3 1.74E-3 -4.51E-3
SUM: 5.30E-03

Node 7
Ux
Uy
Uz
-4.05E-3 1.20E-3 -8.08E-3
SUM: 9.12E-03
2.59E-3 1.00E-3 -1.72E-2
SUM: 1.74E-02

Node 8
Ux
Uy
Uz
-4.97E-3 8.87E-4 -1.11E-2
SUM: 1.22E-02
-6.02E-3 8.52E-4 -1.36E-2
SUM: 1.49E-02

Table 13. Comparison of element forces of two design architectures of Truss G (lbs).
Design
B
C
B
C

Elem 1
-52.55
-5.91
Elem 7
-68.23
-90.70

Elem 2
-68.23
-95.56
Elem 8
-68.23
-70.70

Elem 3
-36.64
-90.70
Elem 9
-76.79
-74.09

Elem 4
-57.19
-70.70
Elem 10
-51.38
-42.78

Elem 5
-52.55
-5.91
Elem 11
-37.86
-36.36

Elem 6
-68.23
-95.56
Elem 12
-73.74
-144.35

Figure 8. Schematic of unstable and stable configurations (from left to right: Design A, B, C)
Truss H is used as the last example (Figure 9). The properties of truss rods are E=1E6 psi; =0.3; A=0.8 in2 and
those of cables are E=1E6 psi; =0.3; A=0.1 in2. Fz=-100 lbs is applied to Node 5 and 6. A pretension of 100 lbs is
applied to each cable. All the DOFs of Nodes 1-4 are fixed. The undeformed (yellow color) and deformed (black
color) configurations are shown on the right side of Figure 9. The variation of design architectures for Truss H is
9
American Institute of Aeronautics and Astronautics

shown in Figures 10&11. Design A and B are stable while design C can not hold the prescribed loads. NASTRAN
program shows that singularity occurs at Y translational DOF of Node 5. A portion of the truss structure is free to
move in Y direction. The two parallelogram frames in Design C formed by Nodes 1-4-6-5, Nodes 5-7-8-6 dont
have any cross-cables for interconnection. Either Cables 1-2 or Cables 3-4 provides links for structural members of
these two frames. Figure 11 displays the functions of Cables 5-10 in truss configurations D, E, F. The removal of
Cable 5-8 can still ensure the structure is stable (Design E). However, if all six Cables 5-10 are eliminated (design
F), the truss loses its stability. Usually, triangular configuration can generate steady truss and some kinds of
interconnection is needed for parallelogram configurations. To find out an optimal configuration for cable
positioning and arrangement is not a straightforward direct solution, and it involves a trial and test iterative analysis
procedure.

Figure 9. Schematic of Truss H (left: schematic, right: undeformed and deformed configurations).

Figure 10. Schematic of Truss H (from left to right: Design A, Design B, Design C)

Figure 11. Schematic of Truss H (from left to right: Design D, Design E, Design F)
Next, we study the influence of truss rod materials on structural performances. The model in Figure 9 is solved
for four kinds of truss materials, which have elastic modulus 1E6 psi, 5E6 psi, 1E7 psi, and 5E7 psi, respectively
(denoted by Mat 1-Mat 4 in Tables 14&15). Maximum displacements occur at Node 5. As shown in Table 14, the
three displacement components are reduced due to the increase of elastic modulus. The enhancement of truss rod
stiffness helps to reduce structural deformation. Table 15 shows that axial forces in truss element are increased at the
same time. Most of truss rods are subjected to compression loads, and elements 5, 6 have the maximum axial forces.
Because truss are constrained at Nodes 1-4, elements 1-4 does not support any loads. These four members are
redundant for this specific condition, and might be useful for other load cases. Efficient usage of truss elements and
elimination of redundant members can be a vital design consideration. This study demonstrates that implementation
10
American Institute of Aeronautics and Astronautics

of truss materials with high elastic modulus helps to minimize truss deformation. Alternatively, increasing crosssectional area of truss rods can be a simple way to reduce truss deformation.
Table 14. Comparison of maximum displacements of Truss H (in).
Displacement
Ux
Uy
Uz

Mat 1
9.729E-2
2.105E-2
1.867E-1

Mat 2
8.861E-2
1.595E-2
1.418E-1

Mat 3
8.748E-2
1.524E-2
1.357E-1

Mat 4
8.657E-2
1.465E-2
1.308E-1

Table 15. Comparison of axial forces of Truss H (lbs).


Elem 5
Elem 6
Elem 7
Elem 8
Elem 9
Elem 10
Elem 11
-250.95*
-218.32
-92.58
106.09
-61.08
-107.95
-80.87
-272.59
-234.33
-103.15
112.37
-67.39
-117.29
-97.23
-275.56
-236.50
-104.63
113.21
-68.31
-118.63
-99.57
-277.99
-238.27
-105.84
113.88
-69.06
-119.72
-101.49
Elem 12
Elem 13
Elem 14
Elem 15
Elem 16
Elem 17
0.309
-210.36
-87.43
-197.95
-68.22
-147.97
Mat 1
-5.96
-226.73
-93.57
-214.73
-81.35
-162.57
Mat 2
-6.88
-228.99
-94.44
-217.06
-83.22
-164.63
Mat 3
-7.65
-230.84
-95.14
-218.98
-84.77
-166.32
Mat 4
(* negative sign represents compressive forces, positive sign represents tensile forces)
Mat 1
Mat 2
Mat 3
Mat 4

In this section, the influence of cables on the stability of cable-truss structures is presented. Essentially, cables
play an important role in stability and integrity of truss systems. The issue of truss weight reduction is also
addressed. In order to reduce total weight of truss-cable structures, some truss members can be substituted with a
pair of cross-cables. The unique property of cables that they can only sustain tension loads should be carefully
considered when we replace truss members with cables. The enhancement of truss material property helps to reduce
truss deformation, but results in the increase of internal forces of truss rods.

VII. Summary
In this paper, a variety of truss structures containing pretensioned cables are analyzed. Commercial FE packages
ANSYS and NASTRAN are employed to study the static behavior of cable-truss structures. Nodal displacements,
truss element forces, and cable tension of various trusses are calculated. This paper provides some general guideline
for truss analysis and configuration design. Some findings have been obtained from this work:
1. Cable geometry and property have influence on truss performances. The increase of cable elastic modulus
results in the decrease of truss deformation. Stiff cables generate less deformation than flexible cables for the same
pretension. The increase of cable cross-sectional areas results in the decrease of truss displacements.
2. Cable pretension is of extreme importance for prediction of truss performance. Appropriate setting of cable
pretension is helpful for the minimization of truss deformation.
3. Cable tension of deformed trusses changes significantly in contrast with pretension. Cables losing tension is
raised as an issue in truss configuration design. Cables could completely lose tension in deformed configurations.
4. The paper discusses some design considerations and factors in the last section. The stability of truss structures
associated with cable placement is addressed by analyzing diverse cable positioning and configurations and by
comparing deformations and axial forces. The behavior of cable-truss structures is affected by the stiffness of truss
members.
The paper addresses weight reduction issue in the last section. For general optimization of cable-truss structures,
comprehensive considerations of loading, constraint condition, weight, strength, material, and cost are needed.
Future work can also be directed toward the design of optimal architectures with local reinforcements and variable
geometry and materials.

11
American Institute of Aeronautics and Astronautics

References
1

Oken, S., Skoumal, D. E., Straayer, J. W., and Loy, C. A., Design of a Graphite/Epoxy Metering Truss for the Large Space
Telescope, Proceeding of 16th AIAA, ASME, and SAE, Structures, Structural Dynamics, and Materials Conference, AIAA,
Denver, Colorado, 1975, 12 pp.
2
Kawashima,
T.,
Sakatani,
Y.,
and
Yamamoto,
T.,
Development
of
Graphite/Epoxy
Tube Truss for Satellite, Composite materials: Mechanics, mechanical properties and fabrication; Proceedings of the JapanU.S. Conference, Tokyo, Japan; United Kingdom, 1981, pp. 453-460.
3
Bowles, D. E., and Tenney, D. R., Composite Tubes for the Space Station Truss Structure, SAMPE Journal, Vol. 23, No.
3, 1987, pp. 49-57
4
Larson, G. J., Jensen, D. W., Transverse Loading of Graphite/Epoxy IsoTruss Panel, Proceeding of 47th
AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference, AIAA, Newport, RI, 2006, pp. 1-15.
5
Sun, Y., Yan, S., and Liang, H., Mechanical Behavior Study of a New Type of 3D Braided Composites Iso-lattice Truss,
Yuhang Xuebao/Journal of Astronautics, Vol. 28, No. 4, 2007, pp. 827-830
6
Punniyakotty, N.M., Liew, J.Y.R., and Shanmugam, N.E., Nonlinear Analysis of Self-Erecting Framework by CableTensioning Technique, Journal of Structural Engineering, Vol. 126, No. 3, Mar. 2000, pp. 361-370.
7
Zhang, A., Nai, G., Economical Analysis Based on Structural Optimum of Long Span Plane Truss String Structure,
Beijing Gongye Daxue Xuebao (Journal of Beijing University of Technology), Vol. 35, No. 1, Jan. 2009, pp. 36-41.
8
You, G., Yang, D., Multi-objective Optimization for a Cable-truss Deployable Space Antenna, Ji Xie She Ji Yu Yan Jiu
(Machine Design and Research), Vol. 24, No. 5, Oct. 2008, pp. 112-115.
9
Johansen, G. E., and Roll, F., Prestressed Kevlar/FRP Structural System, Proceeding of First Serviceability and Durability
of Construction Materials Congress, American Society of Civil Engineers,1990, pp. 640-648
10
Han, Q., Ai, J., Pei, B., and etc., Experimental Study on Prestressed Orthogonal Square Pyramid Composite Space Truss,
Jianzhu Jiegou Xuebao/Journal of Building Structures, Vol. 25, No.5, 2004, pp. 55-59
11
Han, Q., Yuan, Z., and etc., Numerical Analysis and Experimental Study of Prestressed Study Diagonal-on-square
Composite Space Truss, Advances in Structural Engineering, Vol. 8, No.4, 2005, pp. 397-409
12
Zhang, H., Zhang, Z., Xie, X., and Gong, R., Theoretical Study on the Impact Factor of Crescent-shaped Multi-rib
Concrete Filled Steel Tube Truss Arch Bridges, Gongcheng Lixue (Engineering Mechanics), Vol. 25, No. 7, July 2008, pp. 118124.
13
Li, W., Shi, J., and Guo, Z., Research on Prestress Stretching Control of a Large-span Truss String Structure, Dongnan
Daxue Xuebao/Journal of Southeast University (Natural Science Edition), Vol. 33, No. 5, Sept. 2003, pp. 593-596.
14
Liu, S., Mechanical Analysis of Cable-stayed Space Grids with Multi-step-loading, Guangzhou Daxue Xuebao (Ziran
Kexue Ban) / Journal of Guangzhou University (Natural Science Edition), Vol. 7, No. 4, Aug. 2008, pp. 79-84.
15
Yang, D., Qiu, Y., and Duan, B., New Method for Prestressed Astromesh Deployable Antenna Design, Yingyong Lixue
Xuebao/Chinese Journal of Applied Mechanics, Vol. 25, No. 4, Oct.-Dec.2008, pp. 617-621.

12
American Institute of Aeronautics and Astronautics

Das könnte Ihnen auch gefallen