1

STEEL DESIGN
ECVL 466
Dr. Sana Amir
Assistant Professor of Civil Engineering
American University in Dubai
CHAPTER 3

Tension Members
 Introduction
• What are tension members?

• Members subjected to axial tensile forces are called Tension Members.

• These members tend to elongate on the application of load.

• The stress in an axially loaded P P

tensile member is:
P
P P
f 
A P P

P = tensile force
A = Cross-sectional area
f = axial tensile stress
• Bending due to simultaneous transverse loads and buckling
are significantly reduced and an initially non-straight member
tends to straighten up.

• Typical examples are main members of trusses subjected to

tension.

Truss
• Other examples: bracing for buildings and bridges, cables in suspended
roof systems, and cables in suspension and cable-stayed bridges.
Typical sections used for tension members
Single structural steel sections vs Built-up sections

Built-up sections: Plates, rolled shapes,

or a combination of the two (refer to Ch.1)

In general, the use of single structural shape is more

economical than the built-up sections. However, built-up
members may be required when large loads need to be
resisted.
Importance of connections in tension members

Connection details are

considered in design, as
they generally control
the design of tension
members.
 P
• For uniform cross section f 
A
• If the cross-sectional area of a tension member varies along its length, the
stress is a function of the particular section under consideration.

• The presence of holes in a member will influence the stress at a cross

section through the hole or holes.

• At these locations, the cross-sectional area will be reduced by an amount

equal to the area removed by the holes.

P=T P=T
 Area of a tension member
• Tension members -> Frequently bolted.

• A ½ in × 8 in plate, is connected to a gusset plate [a connection

element whose purpose is to transfer the load from the member to a
support or to another member, example truss on next slide].

• The area of the member at section a–a is (½)(8) = 4 in.2,

• But the area at section b–b is only 4 – (2)(7/8) (½) = 3.13 in.2

Plate ½ x 8
Gusset plate examples
 12

GROSS AREA OF CROSS-SECTION (Ag)

It is the total area of cross-section present throughout the

length of the member. (AISC – D3.1).
The elements, which are discontinued lengthwise, are not
included.
For example, area of lacing elements and spacer plates is
not included in gross area.
The gross area for rolled steel shapes is directly available in
the properties tables.
 13

NET AREA OF CROSS-SECTION (An)

When tension members have holes punched in

them for rivets or bolts, the minimum reduced area
after the holes are taken out is called the net area.

Failure of a tension member always occurs at the

weakest section where area of cross-section is
minimum.
Effective Area
Connections affect performance of a tension
member. It almost always weakens the member,
and the measure of its influence is called the joint
efficiency.

This factor is a function of the ductility of the

material, fastener spacing, stress concentrations at
holes, fabrication procedure, and a phenomenon
known as shear lag.
 15

SHEAR LAG FACTOR (U)

1) Stress concentrations adjacent to localized end connections
are due -> shear lag. In the region near the hole or weld, shear
stresses develop that cause the tensile stresses away from the
hole or weld to be less than adjacent to them.

Local stress concentration and shear lag

2) Similarly, at a connection, if one part of the section is
connected while the other is left free, all the forces have
to pass only through the connected part at the joint.
 17

Away from the joint, these stresses spread to give a uniform

stress distribution.

Efficiency of a joint is defined as how well the stresses are

distributed to transfer the applied forces.

If the joint is not fully efficient, premature failure can occur

reducing the member strength.

This expected reduction is usually applied on the area of cross-

section to get effective net area used to calculate the reduced
member strength.
 18

SHEAR LAG FACTOR (U)

Shear lag factor (U) is the factor by which net area of a
section is reduced for shear lag, stress concentrations and
eccentricity at the joints.

Ae = U An [For bolted connections, effective net area]

Ae = U Ag [For welded connections, effective area,

correction applied on gross area]

where the reduction factor U is given in AISC D3, Table D3.1. The table gives a
general equation that will cover most situations as well as alternative numerical
values for specific cases.
As a rule, when tension load is transmitted through
each of the cross-sectional elements by fasteners
or welds, U = 1.0.
The rules for determining U fall into five categories:
1. A general category for any type of tension member except plates and round
HSS with l ≥ 1.3D (See Figure 3.7e.)

2. Plates

3. Round HSS with l ≥ 1.3 D

4. Alternative values for single and double angles

5. Alternative values for W, M, S, and HP shapes

 21

1. The preferable expression for U for all tension members,

except plates and HSS, where load is not transferred by all
elements of the section, is as follows:

x
U = 1 Equation 3.1

x = distance from centroid of element being connected
eccentrically to plane of load transfer, called connection
eccentricity.

 = length of connection, centre-to-centre of the outer rivet holes

or actual length of weld.

Note: when tension load is transmitted

through each of the cross-sectional
elements by fasteners or welds, U = 1.0.
If a member has two symmetrically
located planes of connection, is
measured from the centroid of the
nearest one-half of the area (fig 3.7 c, d)
 x
U =1 Equation 3.1  = length of connection in the direction of the load,
 centre-to-centre of the outer rivet holes or actual length
of weld.
2) Plates

In general, U = 1.0 for plates, since the cross section has only one
element and it is connected.

There is one exception for welded plates: If the member is connected

with longitudinal welds on each side with no transverse weld (as in Figure
3.9), the following values apply:
 x
U =1 Equation 3.1

• bf  2/
3 d U = 0.90
 (Continued)

• bf < 2/ d U = 0.85
3
(Continued)
 31

Note:

Ae = An. U
and U = 1.0
Limitation on effective area, Ae
There are some limiting values for the effective area:

• For bolted splice plates, Ae = An ≤ 0.85Ag. This limit is

given in a user note and is from a requirement in Chapter
J of the Specification “Design of Connections.”

• For open cross-sectional shapes (such as W, M, S, C,

HP, WT, ST and angles), the value of U need not be less
than the ratio of the connected element gross area to the
total gross area.
 33

CALCULATION OF NET AREA

Reduction in Area for One Fastener

• In fabricating structural steel, which is to be connected with rivets or bolts, the

holes are usually punched larger than the diameter of the rivet or bolt.

• The usual practice is to drill or punch standard holes (i.e., not oversized) with a
diameter 1/16 inch larger than the fastener diameter.

• To account for possible roughness around the edges of the hole, AISC B4.3)
requires the addition of 1/16 inch to the actual hole diameter.

• This amounts to using an effective hole diameter 1/8 inch larger than the
fastener diameter.

An = Ag - Aholes
 38

FASTENER SPACING
a) Pitch of fasteners

The centre-to-centre distance of the fasteners along

the longitudinal axis of the member is called pitch
and is denoted by p, as shown in Figure.
A

B
T g T
C

p s
Failure Plane = A – B – C – D
D

Fastener Spacing In Various Directions.

b) Gage distance of fasteners
The centre-to-centre distance between the fasteners along
the transverse direction is called gage denoted by g; refer to
Figure.

B
T g T
C

p s
Failure Plane = A – B – C – D
D

Fastener Spacing In Various Directions.

 40

g2

g1

g
 41

c) Stagger of fasteners
The longitudinal distance between two nearest rivets lying
in two adjacent layers of rivets is called stagger denoted by
s and shown in Figure.

B
T g T
C

p s
Failure Plane = A – B – C – D
D

Fastener Spacing In Various Directions.

 Stagger of fasteners
• In bolted tension members, the net area will be maximized if the fasteners
are placed in a single line.
• Due to space limitations (Figure a), more than one line of bolts is provided.
The reduction in cross-sectional area is minimized if the fasteners are
arranged in a staggered pattern.
• Also, staggered fasteners are required by the geometry of a connection,
Figure b.
• In either case, any cross section passing through holes will pass through
fewer holes than if the fasteners are not staggered.
If the amount of stagger is small enough, the influence of an offset hole may
be felt by a nearby cross section, and fracture along an inclined path such as
abcd in Figure c is possible.

The relationship f = P/A does not apply, and stresses on the inclined portion
b–c are a combination of tensile and shearing stresses.
AISC Section B4.3b:

• If the net area is the product of a thickness times a net

width, and the diameter d is used, the net width in a
failure line consisting of both staggered and unstaggered
holes is:

for rivets

where wn is the net width, wg is the gross width

d is the hole diameter, s is the stagger, g is the gage (transverse spacing)

The second term is the sum of all hole diameters, and the third term is the sum
of for all inclined lines in the failure pattern.
Area approach Width approach

Reduced diameter

When more than one failure pattern is conceivable, all possibilities should be
investigated, and the one corresponding to the smallest load capacity should be used.
 46

2LS

Preferably not Gusset

less than 23 Plate

Rivet
Centre-Lines

Gusset
Plate
2LS

Gusset plate
transversely placed
within the member

Front View End View

Figure 2.7. A Typical Riveted Truss Connection.

As each fastener resists an equal share of the load (an assumption used in the
design of simple connections), different potential failure lines may be
subjected to different loads.

For example, line abcde must resist the full load, whereas ijfh will be subjected to 8/11
of the applied load. 3/11 of the load will have been transferred from the member before
ijfh receives any load.
 50

.
Gage distance of Structural Shapes AISC B4.3b specifies that any gage line crossing
the heel of the angle be reduced by an amount that equals the angle thickness.

ga  ga + gb  t/2  t/2

= ga+gb  t Turning Line

gb

Leg is turned through 90

Tensile strength

A tension member can fail by reaching one of two limit

states:
• excessive deformation
• fracture
• To prevent excessive deformation, initiated by yielding,
the load on the gross section must be small enough that
the stress on the gross section is less than the yield
stress Fy

• To prevent fracture, the stress on the net section must

be less than the tensile strength Fu.

• The stress P/A must be less than a limiting stress F.

where Ae is the effective net area, which may be equal to either the net
area or, in some cases, a smaller area (shear lag factor).
Why yielding in gross area is considered?

Yielding will first occur on the net cross section, the deformation within the
length of the connection will generally be smaller than the deformation in the
remainder of the tension member.

The net section exists over a relatively small length of the member, and the
total elongation is a product of the length and the strain (a function of the
stress).

Most of the member will have an unreduced cross section, so attainment of

the yield stress on the gross area will result in larger total elongation.

It is this larger deformation, not the first yield, that is the limit state.
The design equation:
Block Shear
• For certain connection configurations, a segment or “block” of material
at the end of the member can tear out.
• Part of the failure plane is transverse subjected to tension while the
other part is longitudinal subjected to shear.
• In the figure, ab part is subjected to shear and bc part is having tension.
• For certain arrangements of bolts, block shear can also occur in gusset plates.

• Figure shows a plate tension member connected to a gusset plate.

• Block shear could occur in both the gusset plate and the tension member.

• For the gusset plate, tension failure would be along the transverse section df, and
shear failure would occur on two longitudinal surfaces, de and fg.

• Block shear failure in the plate tension member would be tension on ik and shear
on both hi and jk.
 68
T
a Tension

b c

c Gusset plate
b
T
Shear
a area
Shaded
Tension
may tear out a d
T b c
a)Failure by tearing out b)Large tension, small shear
T
c TGusset plate
b
Tension

b c Shaded
c area
b a d
may tear out
T
a)Failure by tearing out b)Large tension, small shear
a Shear T
d
a c)Large
d shear, small tension

b)Large tension, small shear b c

T

c a d
c)Large shear, small tension
 69

Block Shear Tear-out

The AISC failure model for block shear (tearing out)
failure

• The tearing out failure is either a fracture failure on both the tension resisting
and shear resisting sections (tension fracture + shear fracture) together or
tension fracture combined with shear yielding (tension fracture + Shear
yielding) failure.

• Both surfaces contribute to the total strength, and the resistance to block
shear will be the sum of the strengths of the two surfaces.
• The shear rupture (fracture) stress is taken as 60% of the tensile
ultimate stress, so
• 0.6Fy  yield shear strength = y
0.6Fu  ultimate shear strength = u

• the nominal strength in shear is 0.6FuAnv and

• the nominal strength in tension is FuAnt where,

Anv = net area along the shear surface or surfaces

Ant = net area along the tension surface
 72

Rn = lesser of
1) 0.6 Fu Anv + Ubs Fu Ant
2) 0.6 Fy Agv + Ubs Fu Ant
Nominal tension rupture strength = Ubs Fu Ant
Nominal shear rupture strength = 0.6 Fu Anv
Shear yielding strength = 0.6 Fy Agv

0.6Fy  yield shear strength = y

0.6Fu  ultimate shear strength = u
 73

 = 0.75 (LRFD)
Agv = gross area subjected to shear
Anv = net area in shear
Ant = net area in tension
Ubs = tensile rupture strength reduction factor
(subscript ‘bs’ stands for block shear)
= 1.0 when tensile stress is uniform, such as
in all tensile members and gusset plates
and single row beam end connections
= 0.5 when tensile stress is not uniform
such as for multiple row beam end
connections
 DESIGN OF TENSION MEMBERS
Finding a member with adequate gross and net areas and has the
required strength that does not exceed the available strength.

If the member has a bolted connection, the selection of a suitable cross

section requires an accounting for the area lost because of holes.

For a member with a rectangular cross section, the calculations are

relatively straightforward.

If a rolled shape is to be used, the area to be deducted cannot be

predicted in advance because the member’s thickness at the location of
the holes is unknown.
• Slenderness: L/r, where L is the member length and r is the
minimum* radius of gyration of the cross-sectional area.

• If a structural member has a small cross section in relation to its

length, it is said to be slender.

• inconsequential for a tension member but can be limited. AISC

D1 suggests a maximum slenderness ratio of 300.

*corresponding to the minor principal axis of the cross section

 78

DESIGN PROCEDURE/DESIGN FLOW CHART

Known Data:

Service or working loads, PD, PL etc and length

of member, L

1) Find factored tension (Pu) in LRFD method using

load combinations.
For example, Pu = 1.2 PD + 1.6 PL for gravity
loads alone
 79

2) Areq is calculated to avoid two failures

LRFD
Areq. for riveted members

Pu
a) yielding in the gross section Ag 
0.9 Fy
Pu
b) fracture in the net section Ae 
0.75 Fu
 80

3) Minimum radius of gyration

rmin =  / 300

4) Minimum width of connected leg based on number

of lines of bolts/rivets and gage lengths.
 81

Selection of Trial Section: It depends on the

following four criteria:

A. Areq.
B. Section should be of minimum weight and
smaller size.
C. Connected leg width
D. Radius of gyration.
 82

Check Tensile Capacity: Find actual values of U

and An if rivet pattern and diameter of rivets are
known from connection design.

LRFD
a) Yielding of gross section
t Pn = 0.90 Fy Ag  Pu (OK)

b) Fracture in net section

t Pn = 0.75 Fu Ae  Pu (OK)
 83

Check Fatigue Strength:

If loading cycles > 20,000  increase the
section accordingly
 84

Design the Connections

Check Block Shear Failure: The block shear strength must be

checked at the connection, if the connection details are available.

Write the final selection very clearly

• If an angle shape is used as a tension member and the
connection is made by bolting, there must be enough
room for the bolts.

• Space will be a problem only when there are two lines of

bolts in a leg. The usual fabrication practice is to punch or
drill holes in standard locations in angle legs [Table 1-7A
in Part 1of the Manual].
• Gage distance g applies when there is one line of bolts, and g1 and
g2 apply when there are two lines.

• Figure shows that an angle leg must be at least 5 inches long to

accommodate two lines of bolts.
Tables for the Design of Tension
Members

• Part 5 of the Manual contains tables to assist in the design of

tension members of various cross-sectional shapes, including
Table 5-2 for angles
• Note that if the effective net area must be computed, the
tables do not save much effort.

• In addition, you must still refer to the dimensions and

properties tables to find the radius of gyration.

• The tables for design do, however, provide all other

information in a compact form, and the search may go
more quickly.
Built-up members
• When structural shapes or plates are connected to form a built-up
shape, they must be connected not only at the ends of the member
but also at intervals along its length.

• A continuous connection is not required.

• This type of connection is called stitching, and the fasteners used are
termed stitch bolts.

• The usual practice is to locate the points of stitching so that L/r for any
component part does not exceed L/r for the built-up member.
Spacing of fasteners
• The spacing of fasteners or welds should not exceed 24 times the
thickness of the thinner plate, or 12 inches.

• Spacing must not exceed 1) 24 x thickness of thinner plate

2) or 12 inches

• If the member is of “weathering” steel subject to atmospheric

corrosion, the maximum spacing is 14 times the thickness of the
thinner part, or 7 inches.
 TENSION MEMBERS IN ROOF TRUSSES

• Many of the tension members that structural engineers design are

components of trusses.
• In buildings > the main supporting elements of roof systems where
long spans are required. Industrial or mill buildings.
• Used when the cost and weight of a beam would be prohibitive.

supporting walls
can be reinforced
concrete, concrete
block, brick
Roof trusses normally are spaced uniformly along the length of the building
and are tied together by longitudinal beams called purlins and by x-bracing.
• The primary function of the purlins is to transfer loads to the top chord
of the truss, but they can also act as part of the bracing system.

• Bracing is usually provided in the planes of both the top and bottom
chords.
• Sag rods are tension members - Provide lateral support for the purlins.
• Loads applied to the purlins are vertical with a component parallel to a
sloping roof, which will cause the purlin to bend (sag) in that direction.

Sag rods can be located

at the midpoint, the third
points, or at more
frequent intervals along
the purlins
The End