DESIGN OF TENSION MEMBERS

5 DESIGN OF TENSION MEMBERS

1.0 INTRODUCTION
Tension members are linear members in which axial forces act so as to elongate (stretch)
the member. A rope, for example, is a tension member. Tension members carry loads
most efficiently, since the entire cross section is subjected to uniform stress. Unlike
compression members, they do not fail by buckling (see chapter on compression
members). Ties of trusses [Fig 1(a)], suspenders of cable stayed and suspension bridges
[Fig.1 (b)], suspenders of buildings systems hung from a central core [Fig.1(c)] (such
buildings are used in earthquake prone zones as a way of minimising inertia forces on the
structure), and sag rods of roof purlins [Fig 1(d)] are other examples of tension members.

Suspenders
Rafter Purlin
Sag rod

Tie (a) Roof Truss

Stay cables Top chord

Stayed bridge (c) Suspended (d) Roof Purlin System

Building
X bracings

Suspenders
Suspension Bridge

(b) Cable Supported Bridges (e) Braced Frame

Fig. 1 Tension Members in Structures

Tension members are also encountered as bracings used for the lateral load resistance. In
X type bracings [Fig.1 (e)] the member which is under tension, due to lateral load acting
in one direction, undergoes compressive force, when the direction of the lateral load is
changed and vice versa. Hence, such members may have to be designed to resist tensile
and compressive forces.

© Copyright reserved

Version II 5-1
 DESIGN OF TENSION MEMBERS

The tension members can have a variety of cross sections. The single angle and double
angle sections [Fig 2(a)] are used in light roof trusses as in industrial buildings. The
tension members in bridge trusses are made of channels or I sections, acting individually
or built-up [Figs. 2(c) and 2(d)]. The circular rods [Fig.2 (d)] are used in bracings
designed to resist loads in tension only. They buckle at very low compression and are not
considered effective. Steel wire ropes [Fig.2 (e)] are used as suspenders in the cable
suspended bridges and as main stays in the cable-stayed bridges.

(b)

(a) (c)

(d) (e)

Fig. 2 Cross Sections of Tension Members

2.0 BEHAVIOUR OF TENSION MEMBERS

Since axially loaded tension members are subjected to uniform tensile stress, their load
deformation behaviour (Fig.3) is similar to the corresponding basic material stress strain
behaviour. Mild steel members (IS: 2062 & IS: 226) exhibit an elastic range (a-b) ending
at yielding (b). This is followed by yield plateau (b-c). In the Yield Plateau the load
remains constant as the elongation increases to nearly ten times the yield strain. Under
further stretching the material shows a smaller increase in tension with elongation (c-d),
compared to the elastic range. This range is referred to as the strain hardening range.
After reaching the ultimate load (d), the loading decreases as the elongation increases (d-
e) until rupture (e). High strength steel tension members do not exhibit a well-defined
yield point and a yield plateau (Fig.3). The 0.2% offset load, T, as shown in Fig. 3 is
usually taken as the yield point in such cases.

T
d
e
b c
a
0.2% 

Fig. 3 Load – Elongation of Tension Members

Version II 5-2
 DESIGN OF TENSION MEMBERS

2.1 Design strength of tension members

Although steel tension members can sustain loads up to the ultimate load without failure,
the elongation of the members at this load would be nearly 10-15% of the original length
and the structure supported by the member would become unserviceable. Hence, in the
design of tension members, the yield load is usually taken as the limiting load. The
corresponding design strength in member under axial tension is given by
Td  f y A /  M0 (1)

where, fy is the yield strength of the material (in MPa), A is the gross area of cross
section in mm2 and M0 is the partial safety factor for failure in tension by yielding. The
value of M0 according to IS: 800 is 1.15.

2.2 Plates under Tension

Frequently plates under tension have bolt holes. The tensile stress in a plate at the cross
section of a hole is not uniformly distributed in the elastic range, but exhibits stress
concentration adjacent to the hole [Fig 4 (a)]. The ratio of the maximum elastic stress
adjacent to the hole to the average stress on the net cross section is referred to as the
Stress Concentration Factor. This factor is in the range of 2 to 3, depending upon the
ratio of the diameter of the hole to the width of the plate normal to the direction of stress.
fy fu
fy

(a) Elastic (b) Elasto-Plastic (c) Plastic (d) Ultimate

Fig. 4 Stress Distribution at a Hole in a Plate under Tension

In statically loaded tension members with a hole, the point adjacent to the hole reaches
yield stress, fy, first. On further loading, the stress at that point remains constant at the
yield stress and the section plastifies progressively away from the hole [Fig.4 (b)], until
the entire net section at the hole reaches the yield stress, fy, [Fig. 4(c)]. Finally, the
rupture (tension failure) of the member occurs when the entire net cross section reaches
the ultimate stress, fu, [Fig. 4(d)]. Since only a small length of the member adjacent to
the smallest cross section at the holes would stretch a lot at the ultimate stress, and the
overall member elongation need not be large, as long as the stresses in the gross section is
below the yield stress. Hence, the design strength as governed by net cross-section at the
hole, Tdn, is given by

Ptn  0.9 fu An /  M 1 (2)

Version II 5-3
 DESIGN OF TENSION MEMBERS

where, fu is the ultimate stress of the material, An is the net area of the cross section after
deductions for the hole [Fig.4(b)] and M1 is the partial safety factor against ultimate
tension failure by rupture (M1 = 1.25). Similarly threaded rods subjected to tension could
fail by rupture at the root of the threaded region and hence net area, An, is the root area of
the threaded section (Fig.5).

droot

dgross
(a) elastic (b) elastic (c) Plastic
-Plastic
Fig. 5 Stress in a threaded Rod

The design tension of the plates with hole or the threaded rod could also be governed by
yielding of the gross cross section beyond the thread (with area equal to Ag) above which
the member deformation becomes large and objectionable and the corresponding design
load is given by
Ptg  f y Ag /  M 0 (3)

where, M0=1.15. The lower value of the design tension capacities, as given by Eqn. 2 and
3, governs the design strength of a plate with holes.

(b) 1
(c)

2
g
3
(a)
p
Fig. 6 Plates with Bolt Holes under Tension
(d) 4

Version II 5-4
 DESIGN OF TENSION MEMBERS

Frequently, plates have more than one hole for the purpose of making connections.
These holes are usually made in a staggered arrangement [Fig.6 (a)]. Let us consider the
two extreme arrangements of two bolt holes in a plate, as shown in Fig.6 (b) & 6(c). In
the case of the arrangement shown in Fig.6 (b), the gross area is reduced by two bolt
holes to obtain the net area. Whereas, in arrangement shown in Fig.6c, deduction of only
one hole is necessary, while evaluating the net area of the cross section.
Obviously the change in the net area from the case shown in Fig.6(c) to Fig.6 (b) has to
be gradual. As the pitch length (the centre to centre distance between holes along the
direction of the stress) p, is decreased, the critical cross section at some stage changes
from straight section [Fig.6(c)] to the staggered section 1-2-3-4 [Fig.6 (d)]. At this stage,
the net area is decreased by two bolt holes along the staggered section, but is increased
due to the inclined leg (2-3) of the staggered section. The net effective area of the
staggered section 1-2-3-4 is given by

An  (b  2d  p 2 / 4 g ) t (4)
where, the variables are as defined in Fig.6(d). In Eqn. 4 the increase of net effective
area due to inclined section is empirical and is based on test results. It can be seen from
Eqn.4, that as the pitch distance, p, increases and the gauge distance, g, decreases, the net
effective area corresponding to the staggered section increases and becomes greater than
the net area corresponding to single bolt hole. This occurs when

p 2 / 4g  d (5)
When multiple holes are arranged in a staggered fashion in a plate as shown in Fig.6 (a),
the net area corresponding to the staggered section in general is given by
 p2 
Anet   b  nd   t (6)
 4 g 

where, n is the number of bolt holes in the staggered section [n = 7 for the zigzag section
in Fig. 6(a)] and the summation over p2/4g is carried over all inclined legs of the section
[equal to n-1 = 6 in Fig.6(a)]. Normally, net area of different staggered and straight
sections have to be evaluated to obtain the minimum net area to be used in calculating the
design strength in tension. An example analysis of a plate with holes under tension is
illustrated in Appendix I.

2.3 ANGLES UNDER TENSION

Angles are extensively used as tension members in trusses and bracings. Angles, if
axially loaded through centroid, could be designed as in the case of plates. However,
usually angles are connected to gusset plates by bolting or welding only one of the two
legs (Fig. 7).
This leads to eccentric tension in the member, causing non-uniform distribution of stress
over the cross section. Further, since the load is applied by connecting only one leg of
the member there is a shear lag locally at the end connections.

Version II 5-5
 DESIGN OF TENSION MEMBERS

Fig. 7 Angles Eccentrically Loaded through Gussets

Kulak and Wu (1997) have reported, based on an experimental study, the results on the
tensile strength of single and double angle members. Summary of their findings is:

 The effect of the gusset thickness, and hence the out of plane stiffness of the end
connection, on the ultimate tensile strength is not significant.
 The thickness of the angle has no significant influence on the member strength.
 The effect of shear lag, and hence the strength reduction, is higher when the ratio of
the area of the outstanding leg to the total area of cross-section increases.
 When the length of the connection (the number of bolts in end connections)
increases, the tensile strength increases up to 4 bolts and the effect of further
increase in the number of bolts, on the tensile strength of the member is not
significant. This is due to the connection restraint to member bending caused by
the end eccentric connection.
 Even double angles connected on opposite sides of a gusset plate experience the
effect of shear lag.

Based on the test results, Kulak and Wu (1997) found that the shear lag due to connection
through one leg only causes at the ultimate stage the stress in the outstanding leg to be
closer only to yield stress even though the stress at the net section of the connected leg
may have reached ultimate stress. They have suggested an equation for evaluating the
tensile strength of angles connected by one leg, which accounts for various factors that
significantly influence the strength. In order to simplify calculations, this formula has
suggested that the stress in the outstanding leg be limited to fy (the yield stress) and the
connected sections having holes to be limited to fu (the ultimate stress). The design tensile
strength, Td, should be the minimum of the following:

Strength as governed by tearing at net section:

Ptn = Anc fu / M1+  Ao fy / M0 (7a)

where, fy and fu are the yield and ultimate stress of the material, respectively. Anc and Ao,
are the net area of the connected leg and the gross area of the outstanding leg,
respectively. The partial safety factors M0 = 1.15 and M1 = 1.25.  accounts for the end
fastener restraint effect and = 1.0, if the number of fasteners is  4,  = 0.75 if the
number of fasteners = 3 and  = 0.5, if number of fasteners = 1 or 2. In case of welded
connection,  = 1.0.

Version II 5-6
 DESIGN OF TENSION MEMBERS

Strength as governed by yielding of gross section:

Ptg = Ag fy/M0 (7 b)

where, Ag is the gross area of the angle section.

Strength as governed by block shear failure:

A tension member may fail along end connection due to block shear as shown in Fig. 8.
The corresponding design strength can be evaluated using the following equations. If the
centroid of bolt pattern is not located between the heel of the angle and the centerline of
the connected leg, the connection shall be checked for block shear strength given by

Block shear plane

Fig. 8 Block Shear Failure

Ptv = (0.62 Avg fy/M0 + Atn fu/M1)

or
Ptv = (0.62 Avn fu/M1 + Atg fy/M0) (7c)

where, Avg and Avn = minimum gross and net area in shear along a line of transmitted
force, respectively, and Atg and Atn = minimum gross and net area in tension from the hole
to the toe of the angle, perpendicular to the line of force, respectively.

The design strength of an angle loaded in tension through a connection in one leg is given
by the smallest of the values obtained from Eqns. 7(a) to 7(c). These equations are valid
for both single angle and double angles in tension, irrespective of whether they are on the
same side or opposite sides of the gusset. A sample design of angle tension member is
given in worked example 2.
The efficiency, , of an angle tension member is calculated as given below:

  Pt /( Ag f y /  M 0 ) (8)

Depending upon the type of end connection and the configuration of the built-up
member, the efficiency may vary between 0.85 and 1.0. The higher value of efficiency is
obtained in the case of double angles on the opposite sides of the gusset connected at the
ends by welding and the lower value is usual in the bolted single angle tension members.
In the case of threaded members the efficiency is around 0.85.

Version II 5-7
 DESIGN OF TENSION MEMBERS

In order to increase the efficiency of the outstanding leg in single angles and to decrease
the length of the end connections, some times a short length angle at the ends are
connected to the gusset and the outstanding leg of the main angle directly, as shown in
Fig. 9. Such angles are referred to as lug angles. The design of such end connections is
discussed in the chapter on connections.

lug angle

Fig. 9 Tension Member with Lug

Angle

3.0 DESIGN OF TENSION MEMBERS

In the design of a tension member, the design tensile force is given and the type of
member and the size of the member have to be arrived at. The type of member is usually
dictated by the location where the member is used. In the case of roof trusses, for
example, angles or pipes are commonly used. Depending upon the span of the truss, the
location of the member in the truss and the force in the member either single angle or
double angles may be used in roof trusses. Single angle is common in the web members
of a roof truss and the double angles are common in rafter and tie members of a roof
truss.
Plate tension members are used to suspend pipes and building floors. Rods are also used
as suspenders and as sag rods of roof purlins. Steel wires are used as suspender cables in
bridges and buildings. Pipes are used in roof trusses on aesthetic considerations, in spite
of fabrication difficulty and the higher cost of such tubular trusses. Built-up members
made of angles, channels and plates are used as heavy tension members, encountered in
bridge trusses.

3.1 Trial and Error Design Process

The design process is iterative, involving choice of a trial section and analysis of its
capacity. This process is discussed in this section. Initially, the net effective area
required is calculated from the design tension and the ultimate strength of the material as
given below.
An  Pt / ( fu /  M1 ) (9)
Using the net area required, the gross area required is calculated, allowing for some
assumed number and size of bolt holes in plates, or assumed efficiency index in the case

Version II 5-8
 DESIGN OF TENSION MEMBERS

of angles and threader rods. The gross area required is also checked against that required
from the yield strength of the gross sections as given below.

Ag  Pt /( f y /  M 0 ) (10)

A suitable trial section is chosen from the steel section handbook to meet the gross area
required. The bolt holes are laid out appropriately in the member and the member is
analysed to obtain the actual design strength of the trial section. The design strength of
the trial section is evaluated using Eqs. 1 to 6 in the case of plates and threaded bars and
using Eqs. 7 in the case of angle ties. If the actual design strength is smaller than or too
large compared to the design force, a new trial section is chosen and the analysis is
repeated until a satisfactory design is obtained.

3.2 Stiffness Requirement

The tension members, in addition to meeting the design strength requirement, frequently
have to be checked for adequate stiffness. This is done to ensure that the member does
not sag too much during service due to self-weight or the eccentricity of end plate
connections. The IS: 800 imposes the following limitations on the slenderness ratio of
members subjected to tension:
(a) In the case of members that are normally under tension but may experience
compression due to stress reversal caused by wind / earthquake loading /r  250.

(b) In the case of members that are designed for tension but may experience stress
reversal for which it is not designed (as in X bracings) /r  350.

(c) In the case of members subjected to tension only /r  400

In the case of rods used as a tension member in X bracings, the slenderness ratio
limitation need not be check for if they are pretensioned by using a turnbuckle or other
such arrangement.

4.0 SUMMARY
The behaviour and design of various types of tension members were discussed. The
important factors to be considered while evaluating the tensile strength are the reduction
in strength due to bolt holes and due to eccentric application of loads through gusset
plates attached to one of the elements. It was shown that the yield strength of the gross
area or the ultimate strength of the net area may govern the tensile strength. The effect of
connecting the end gusset plate to only one of the elements of the cross section was
empirically accounted for by the reduction in the effectiveness of the out standing leg,
while calculating the net effective area. The methods for accounting for these factors in
the design of tension members were discussed. The iterative method of design of tension
members was presented.

Version II 5-9
 DESIGN OF TENSION MEMBERS

5.0 REFERENCES
1. AISC–LRFD. ‘Load and resistance factor design specification for structural steel
buildings’. American Institute of Steel Construction (AISC), Chicago, III, 1993.
2. ASCE Manual No.52. ‘Guide for design of steel transmission towers’ American
Society of Civil Engineers, 1987.
3. BS-5950. ‘Code of practice for design in simple and continuous construction: Hot
rolled sections’ British Standards Institute, London, 1985.
4. CAN3-S16.1-M84. ‘Steel structures for buildings (limit states design)’, Canadian
Standards Assoc., Rexdale, Ontario, Canada, 48, 1984.
5. Eurocode 3. ‘Design of steel structures’, British Standards Institute 1992.
6. IS:800-1984. ‘Code of Practice for General Construction in Steel’ Bureau of Indian
Standards, New Delhi, 1984.
7. Kulak and Wu, ‘Shear Lag in Bolted Angle Tension Members’, ASCE, Journal of
Structural Engineering, Vol.123, No.9, Sept.1997, pp.1144-1152.
8. Mueller, W.H., and Wagner, A. L. ‘Plastic behaviour of steel angle columns’, Res.
Rept., Bonneville Power Admin., Portland, Oreg., 1985, pp 33-82.
9. Murty, Madugula and S. Mohan, ‘Angles In Eccentric Tension’, ASCE, Journal of
Structural Engineering, Vol.114, No.10, October 1988, pp.2387-2396.
10. Nelson, H. M. ‘Angles in Tension’, Publication No.7, British Constructional
Steelwork Assoc., United Kingdom, 1953, pp 8-18.

Version II 5-10
 DESIGN OF TENSION MEMBERS

Job No: Sheet: 1 of 1 Rev:

Structural Steel Job Title: Tension Member Example
Design Project Worked Example: 1
Made by Date: 3-1-2000
SSSR
Checked by Date
VK
Calculation Sheet
PROBLEM 1:
Determine the desig tensile strength of the plate (200 X 10 mm) with the
holes as shown below, if the yield strength and the ultimate strength of the
steel used are 250 MPa and 420 MPa and 20 mm diameter bolts are used.
fy = 250 MPa 1
40
fu = 420 MPa 1 2
30
3 30
200 30
2 30
1 40
1 50 50
Calculation of net area, Anet:

An (section 11)  200  3* 21.5 * 10  1355 mm 2 governs

 2* 50 2 
An (section 1221)   200  4* 21.5  * 10  1557 mm 2

 4* 30 
 4* 50 2 
An (section 12321)   200  5* 21.5  * 10  1758 mm 2

 4* 30 

Pt is lesser of
200* 10* 250/1.15
(i ) Ag f y / M 0   434.8 kN
1000
0.9 * 1342* 420 /1.25
(ii ) 0.9 An f u /γ M 1   405.8 kN
1000

Pt  405.8 kN

Pt 409.8
Efficiency of the plate with holes =   0.93
Ag f y /γ M 0 434.8

Version II 5-11
 DESIGN OF TENSION MEMBERS

Job No: Sheet: 1 of 4 Rev

Structural Steel Job Title: Tension Member Example
Design Project Worked Example: 2
Made by Date 3-1-2000
SSSR
Checked by Date
VK
Calculation Sheet
PROBLEM 2:
Analysis of single angle tension members
A single unequal angle 100 X 75 X 8 mm is connected to a 12 mm thick
gusset plate at the ends with 6 nos. 20 mm diameter bolts to transfer tension.
Determine the design tensile strength of the angle. (a) if the gusset is
connected to the 100 mm leg, (b) if the gusset is connected to the 70 mm leg,
(c) if two such angles are connected to the same side of the gusset through
the 100 mm leg. (d) if two such angles are connected to the opposite sides of
the gusset through 100 mm leg.
12 mm
30 50 * 5

40
ISA 100 X 75 X 8

20 mm  bolts

a) The 100mm leg bolted to the gusset :

Anc = (100 - 8/2 - 21.5) *8 = 596 mm2.
Ao = (75 - 8/2) * 8 = 568.mm2
Ag = ((100-8/2) + (75 – 8/2)) * 8 =1336 mm2
Strength as governed by tearing of net section:

Since the number of bolts = 4;  = 1.0

Pt = Anc fu/M1+  Ao fy/M0

= 596 * 420/1.25 + 1.0 * 568 * 250 / 1.15
= 323734 N (or) 323.7 kN

Version II 5-12
 DESIGN OF TENSION MEMBERS

Job No: Sheet: 2 of 4 Rev

Structural Steel Job Title: Tension Member Example
Design Project Worked Example :2
Made by Date 3-1-2000
SSSR
Checked by Date
VK
Calculation Sheet

Strength as governed by yielding of gross section:

Pt = Ag fy/M0
= 1336 * 250 / 1.15 = 290435 N (or) 290.4 kN
Block shear strength

Pv = (0.62 Avg fy/M0 + Atn fu/M1)

= 0.62 * (5 *50 +30)* 8 * 250/1.15 + (40-21.5/2) * 8 * 420/1.25
= 380537 N = 380.5 kN
or
Pv = (0.62 Avn fu/M1 + Atg fy/M0)
= (0.62 (5 * 50 + 30 –5.5 * 21.5) * 8 * 420 / 1.25
+ 40 * 8 * 250/ 1.15)
= 339131 N = 339.1
The design tensile strength of the member = 290.4 kN
The efficiency of the tension member, is given by

Pt 290.4* 1000
   1.0
Ag f y 100 75  8 * 8* 250/1.15

b) The 75 mm leg is bolted to the gusset:

Anc = (75 - 8/2 - 21.5) * 8 = 396 mm2
Ao = (100 - 8/2) * 8 = 768 mm2

Version II 5-13
 DESIGN OF TENSION MEMBERS

Job No: Sheet: 3 of 4 Rev

Structural Steel Job Title: Tension Member Example
Design Project Worked Example: 2
Made by Date 3-1-2000
SSSR
Checked by Date
Calculation Sheet VK

12 mm

20 mm  bolts ISA 100 X 75 X 88

Strength as governed by tearing of net section:

Since the number of bolts = 6  = 1.0

Pt = Anc fu/M1+  Ao fy/M0
= 396 * 420/1.25 + 1.0 * 768 *250 / 1.15
= 300123 N (or) 300.1 kN
Strength as governed by yielding of gross section:
Pt = Ag fy /M0

= 1336 * 250 / 1.15 = 290435 N (or) 290.4 kN

Block shear strength:
Pv  (0.62 Avg fy/M0 + Atn fu/M1)
= 0.62 * (5 *50 +30)* 8 * 250/1.15
+ (35-21.5/2) * 8 * 420/1.25
= 367097 N = 367.1 kN
Pv  (0.62 Avn fu/M1 + Atg fy/M0)
= (0.62 (5 * 50 + 30 –5.5 *x 21.5) * 8 * 420 / 1.25
+ 35 *8 * 250/ 1.15
= 330435 N = 330.4 kN

The design tensile strength of the member = 290.4 kN

Version II 5-14
 DESIGN OF TENSION MEMBERS

Job No: Sheet: 4 of 4 Rev

Structural Steel Job Title: Tension Member Example
Design Project Worked Example : 2
Made by Date 3-1-2000
SSSR
Checked by Date
VK
Calculation Sheet

Even though the tearing strength of the net section is reduced, the yielding of
the gross section still governs the design strength.
The efficiency of the tension member is as before 1.0
Note: The design tension strength is more some times if the longer leg of an
unequal angle is connected to the gusset (when the tearing strength
of the net section governs the design strength).
An understanding about the range of values for the section efficiency,
, is useful to arrive at the trial size of angle members in design
problems.

(c & d)The double angle strength would be twice single angle strength as
obtained above in case (a)

Pt = 2 * 290.4 = 580.8 kN

Version II 5-15