Sei sulla pagina 1di 6

STRENGTH OF MATERIALS(Part 2&3) CN213

Part 2 Torsion
Introduction
In this part we will first discuss the effects of applying a torsional loading to a thin-walled tube and a thick-
walled tube. Then a member under transverse shearing will be discussed. Finally, a open thin-walled
section under combined bending and torsion will be introduced, including shear centre calculation.
2.1 Torsion of a thick tubular section
The formula to determine shear stress and twist angle of a circular shaft under a torque should already
introduced in your first year study. Here we just review some theory and show an example.
The following formula is used to calculate the shear stress in a circular member
J
T
(2.1)
where T is the resultant internal torque acting at the cross section,
is the distance from the centre of the shaft,
J is the polar moment of inertia of the cross-sectional area.
The maximum shear stress in the shaft, which occurs at the outer surface, is
J
TR
max
(2.2)
where R is the outer radius of the shaft.
Please note that the above equations can be used if the shaft is circular and the material is homogeneous
and behaves in a linear elastic manner. The reason is that the derivation is based on the fact that the shear
stress is proportional to the shear strain, and therefore both vary linearly along every radial line on the cross
section.
If a shaft has a constant cross sectional area, also the material is homogeneous (G=constant), the twist
angle of one end of the shaft with respect to the other end (measured in radians) under a constant torque T
can be determined using following equation
JG
TL
(2.3)
where L is the length of the shaft and G is shear modulus.
2.2 Thin-walled closed sections
Thin-walled non-circular tubes are often used in aircraft. They may be, in some applications, subjected to a
torsional loading. Since the walls are thin, we can obtain an approximate solution for the shear stress by
assuming that this stress is uniformly distributed across the thickness of the tube.
Shear flow. Fig. 2-1 shows a small element of the tube cut from a whole tube under a torque T, which has
a finite length s and differential width dx. Considering force equilibrium within the element, we have
) dx t ( dF ) dx t ( dF
B B B A A A
, which yields
B B A A
t t
This result indicates that the product of the average longitudinal shear stress times the thickness of the tube
is the same at each point along the tubes cross-sectional area. This product is known as shear flow,
expressed using q. Therefore
t q
avg

(2.4)
2.1
STRENGTH OF MATERIALS(Part 2&3) CN213
If wall thickness is variable, owing to constant q through the cross section, the largest average shear stress
occurs where the thickness of a tube is the smallest. Remember that the shear flow and the average shear
stress are always directed tangent to the wall of the tube.
Average shear stress. The average shear stress acting on the an area dA=tds shown in Fig. 2-2 can be
linked to the torque T by taking the torque created by the shear stress about a selected point O with in the
tubes boundary. In the above figure, we can see that the shear stress forms a force
) tds ( dA dF
avg avg

on the element. This force acts tangent to the centreline of the tubes cross-
sectional area. Assuming that the moment arm is h (see Fig. 2-3), we have the torque as
) tds ( h ) dF ( h dT
avg

For the whole section, we can integrate the above equation to obtain the whole torque

s
avg
tds h T

s
means the line integral that is performed around the entire boundary of the area. Owing to constant q
(=
t
avg

), the above equation can be rewrite as

s
avg
hds t T
By introducing the mean area, h d s d A
m
2
1
, we have
m avg m avg
tA dA t T

2 2
Finally, the average shear stress is
m
avg
tA
T
2

(2.5)
Noting
t q
avg

, the equation used to determine the shear flow throughout the cross section is
m
A
T
q
2

(2.6)
Angle of twist. If the material behaves in a linear-elastic manner, the angle of twist of a thin-walled tube of
length L can be determined using following equation

s
m
t
ds
G A
TL
2
4

(2.7)
where G is the shear modulus. The integration above must be performed around the entire boundary of the
tubes cross sectional area.
2.3 Shear flow in thin-walled members
Before we introduce shear flow in open thin-walled members we need to review the shear formula for a
beam under transverse loading. Fig. 2-4 shows a beam under such loading. By considering the force
equilibrium in longitudinal direction we can obtain the formula to determine shear stress in a beam as
follows:
It
VQ
(2.8)
where
V the internal resultant shear force,
I the moment of inertia of the entire cross-sectional area computed about the neutral axis,
2.2
STRENGTH OF MATERIALS(Part 2&3) CN213
t the width of the members cross-sectional area, measured at the point where is to be
calculated,
Q

' A
' A ' y ' ydA
, where A is the top (or bottom) portion of the members cross-
sectional
area, defined from the section where t is measured, and
' y
is the distance to the centroid
of
A, measured from the neutral axis.
Now we can derive the formula to calculate the magnitude of the shear flow along any longitudinal section
of a beam. Cutting a small element from a beam under pure bending, see Fig. 2-5, we can calculate the
incremental force dF as follows

' A
' ydA
I
dM
dF
Clearly, the above integral represents Q. Since the element has a length dx, the shear flow, or force per unit
length along the beam, is q=dF/dx. Thus dividing both sides of the above equation by dx and noting that
V=dM/dx, we obtain
I
VQ
q (2.9)
In a thin-walled member under transverse shearing, there is a shear-flow distribution throughout a
members cross-sectional area. We assume that the member has thin wall, i.e. the wall thickness is small
compared with the height or width of the member. This assumption means that we can take the shear flow
as constant through the wall thickness.
We consider the slice dx of the I-section beam shown in Fig. 2-6, a free body element cut from the beam is
also shown in same diagram. In order to balance the normal forces F and F+dF generated by the moments
M and M+dM the force dF is developed along the side longitudinal section, see Fig. 2-6b. The shear flow
along the section is q=dF/dx. Assuming constant shear stress through the thin wall of the flange, there is
dF= dA= (tdx). Noting that dF=qdx, we have
q= t (2.10)
The above equation can also be derived from the shear-flow equation, q=VQ/I and the shear formula,
=VQ/It.
Using similar analysis we can determine the direction of the shear flow at any point on a beams cross
section. Now we are going to use the shear flow formula, q=VQ/I, to determine the distribution of the shear
flow throughout the cross section of the I-section beam shown in Fig. 2-7a. For doing this, consider the
shear flow q acting on the shadow part, located an arbitrary distance x from the centreline of the cross
section, see Fig. 2-7b. Using Eq. (2.9) and noting
t ) x / b )( / d ( ' A ' y Q 2 2
, there is

,
_

x
b
I
Vtd
I
t ) x / b )( / d ( V
I
VQ
q
2 2
2 2
(2.11)
Clearly, the above equation indicates that the shear-flow distribution in linear, varying from q=0 (x=b/2) to
I / Vtdb ) q (
f max
4
(x=0). Owing to symmetry, the linear distributions can be obtained from the similar
analysis, so as their directions, which are shown in Fig. 2-7d.
The total force allocated in each part of a flange can be obtained using following equation (see Fig. 2-7b)
I
Vtdb
dx x
b
I
Vtd
qdx F
/ b
f
16 2 2
2
2
0



,
_


Alternatively, it can be determined as follows due to its linear distribution from zero
I
Vtdb
)
b
( ) q ( F
f max f
16 2 2
1
2

2.3
STRENGTH OF MATERIALS(Part 2&3) CN213
There are four of these flange forces (see Fig. 2-7e), and clearly they are balanced each other in horizontal
direction.
We can use similar analysis to determine the shear flow and the distribution in the web. Here, Q is
comprised of two parts, one from the top flange and one from the top half of the web, i.e.
) y
d
)(
t
(
btd
) y
d
( t )] y
d
( y [ ) bt )(
d
( ' A ' y Q
2
2
4 2 2 2 2 2
1
2
+ + +
Thus,
1
1
]
1

+ ) y
d
)( (
bd
I
Vt
I
VQ
q
2
2
4 2
1
2
(2.12)
From the above equation we know that the shear flow in the web, like shear stress, varies in a parabolic
manner, from
I / Vtdb ) q ( q
f max
2 2
(y=d/2) to ) / d / b )( I / Vtd ( ) q (
w max
8 2 + (y=0), see
Fig. 2-7d.
The force in the web,
w
F , can be determined by integrating Eq. (2.12) as follows,

) d b (
I
Vtd
dy ) y
d
)( (
bd
I
Vt
qdy F
/ d
/ d
w
3
1
2
4 4 2
1
2
2
2
2
2
2
+

1
1
]
1

The moment of inertia for the cross-section is


3 2 3
12
1
2 12
1
2 td )
d
( bt bt I +
1
]
1

+
Since the thickness of each flange is small and their cubic is even smaller, the first term in the above
equation can be omitted. Thus
) d b (
td
I
3
1
2
4
2
+
Substituting the simplified I into the equation of
w
F , we obtain
w
F =V, which is expected (see Fig. 2-
7e).
More examples of how shear flow q is directed along the segments of thin-walled members are shown in
Fig. 2-8.
2.4 Shear centre
As mentioned before for transverse bending, if the internal shear V is applied along a principal centroidal
axis of inertia that also represents an axis of symmetry of the cross section, the beam is only under bending
without any possible twist. But if this force is applied along the vertical, unsymmetrical axis of a channel
beam that passes through the centroid C of the cross-sectional area shown in Fig. 2- 9a, the channel will not
only bend downward, it will also twist clockwise. Shear flow distribution within the channel, which helps
understand this, is displayed in Fig. 2-9b.
Similar to the last section, when integrate the shear flow distribution over the flange and web areas, we can
obtain resultant forces of f
F
in each flange and a force of V=P in the web, shown in Fig. 2-9c. Clearly,
if the moments of these forces are taken at point A, the couple or torque created by the flange forces causes
the channel to twist. In order to prevent this twisting force P needs to be applied at a point O located a
distance e from the web of the channel, see Fig. 2-9d. The distance e can be determined by requiring
Pe d F M
f A

, which generates
P
d F
e
f

(2.13)
2.4
STRENGTH OF MATERIALS(Part 2&3) CN213
The point O is called the shear centre or flexural centre. When force P is applied at the shear centre, the
beam will bend without twisting, as shown in Fig. 2-9e. It should be noted that the shear centre will always
lie on an axis of symmetry of a members cross-sectional area.
Part 3 Unsymmetrical bending
In this section we are going to develop the flexure formula applied to either to a beam having a cross-
sectional area of any shape or to a beam having a resultant internal moment that acts in any direction.
However, the formula learnt in the first year is still help you to understand the new development.
3.1 Principal axis
Fig. 3-1 shows a straight and prismatic beam with a unsymmetrical shape. The right-handed x, y, z co-
ordinate system is located at the centroid C of the cross-section, and the resultant internal bending moment
M acts along the z axis. The equilibrium conditions can be expressed by considering the force (
dA dF ) acting on the small element dA located at (0, y, z), see Fig. 3-1a, as follows
; F
x
0


A
dA 0
(3.1)
; M
y
0

A
dA z 0
(3.2)
; M M
z



A
M dA y
(3.3)
Since the z axis passes through the centroid of the cross-sectional area Eq. (3.1) is satisfied, i.e. the
resultant force on the area above the z axis is equal to that on the area below the z axis. Moreover, as the z
axis is the neutral axis for the cross section, the normal strain will vary linearly from the zero at the neutral
axis to a maximum at extreme fibre, y=c, from the neutral axis, see Fig. 3-1b. Assuming that the material
behaves linear elastically, the normal stress distribution through the cross-sectional area is also linear.
Therefore, we have
max
) c / y ( , see Fig. 3-1c. Substituting it into Eq. (3.3) and integrating the
equation, we have the flexure formula I / Mc
max
, which is similar to the formula you learnt in the
first year. Now when substitute it into Eq. (3.2), there is


A
max
yzdA
c
0

which leads


A
yzdA 0
(3.4)
The above integral is called the product of inertia for the area. If the y and z axes are principal axes of
inertia for the area, this integral is indeed zero. For an arbitrary area, as shown in Fig. 3-1a, the orientation
of the principal axes can always be determined using either the inertia transformation equations or Mohrs
circle of inertia, which are similar to those to determine principal stresses and strains. The equation are
listed as follows:
2
2
/ ) I I (
I
tan
y x
xy
p


(3.5)
and principal moments of inertia
2
2 2
xy
y x y x
min max,
I
I I I I
I +

,
_


t
+
(3.6)
where

A
xy
A
y
A
x
xydA I , dA x I , dA y I and
2 2
.
Eq. (3.5) has two roots,
1
p

and
2
p

, which are
0
90 apart and so specify the inclination of each
principal axis.
2.5
STRENGTH OF MATERIALS(Part 2&3) CN213
Examples.
3.2 Symmetrical beam under skew load
In some circumstances, a beam may be loaded by a skew moment that does not act about one of the
principal axes of the cross section. When this happens, the moment should first be projected to the
principal axes. Then flexure formula can be used to determine the normal stress caused by each moment
component. Finally, using the principle of superposition, the resultant normal stress at the point can be
determined.
Fig. 3-2a shows that a rectangular beam is subjected to a moment M. According to the right-hand rule,
vector M has an angle with respect to the principal axis z. By projecting M into components along the z
and y axes, we have cos M M
z
and
sin M M
y

, respectively. They are displayed in Fig. 3-2b
and 3-3c. Two components of normal stresses are produced by
z
M and y
M
, which are
y
y
'
x
z
z
x
I
z M
,
I
y M

The negative sign above is coincided with the sign convention of normal stress, i.e. tensile stress is positive
and compressive stress is negative. Using the superposition principle and assuming
max
'
x max x
) ( ) ( > ,
the resultant normal stress is shown in Fig. 3-2d. Finally, we have
y
y
z
z
I
z M
I
y M
+
(3.7)
3.3 Orientation of the neutral axis
Since there is no strain, and therefore, no stress on the neutral axis, we can determine the neutral axis by
letting Eq. (3.7) equal to zero, i.e
0 +
y
y
z
z
I
z M
I
y M

z
I M
I M
y
y z
z y

Substituting cos M M
z
and
sin M M
y

into the above equation, there is
z tan
I
I
y
y
z

,
_


(3.8)
Clearly, this is the equation of the line that defines the neutral axis for the cross section. Since the slope of
the line is
z / y tan
, therefore we have
tan
I
I
tan
y
z

(3.9)
is shown in Fig. 3-2d. It should be noted that for unsymmetrical bending the angle , defining the
direction of the moment M, is not equal to , the angle defining the inclination of the neutral axis, unless
y z
I I
. If y z
I I >
, i.e. the y axis is the principal axis for the minimum moment of inertia, and the z
axis is the principal axis for the maximum moment of inertia, we can conclude that the angle , which is
measured positive from the +z axis toward the +y axis, will lie between the line of action of M and the y
axis, see Fig. 3-2d.
2.6

Potrebbero piacerti anche