Experiment-5: Southwell Method To Study Flexural Instability of Column
Experiment-5: Southwell Method To Study Flexural Instability of Column
Experiment-5: Southwell Method To Study Flexural Instability of Column
APPARATUS:
Imperfect Aluminium Pin ended Column
2 Strain Gauges
Strain Gauge meter
Scale
Vernier Calliper
Digital Gauge
Load cell
THEORY:
Buckling of Ideal Pin-ended column:
Consider a pin-ended column AB. Let us assume it is in the displaced state of
neutral equilibrium associated with buckling so that the compressive load P has
attained critical load 𝑃𝑐𝑟 .
Simple theory of bending can be used to describe the equation of the elastic curve
as:
𝑑2𝑦
𝐸𝐼 = 𝑀 = −𝑃𝑦
𝑑𝑥 2
𝑑2𝑦 𝑃𝑦 2
𝑃
= − ; µ =
𝑑𝑥 2 𝐸𝐼 𝐸𝐼
The solution for the differential equation is,
𝑦 = 𝐴𝑐𝑜𝑠(µ𝑥 ) + 𝐵𝑠𝑖𝑛(µ𝑥)
Applying the boundary conditions, i.e. at x=0,l implies y=0
𝐴 = 0, 𝐵𝑠𝑖𝑛(µ𝑙 ) = 0
For a non-trivial solution,
𝑛2 𝜋2 𝐸𝐼
𝑠𝑖𝑛(µ𝑙 ) = 0 ⇒ µ𝑙 = 𝑛𝜋 ⇒ 𝑃𝑐𝑟 =
𝑙2
So, The smallest value of buckling load is
𝜋2 𝐸𝐼
𝑃𝑐𝑟 = 2
𝑙
When both ends are fixed,
4𝜋2 𝐸𝐼
𝑃𝑐𝑟 =
𝑙2
When one end is fixed and other is hinged,
2.04𝜋2 𝐸𝐼
𝑃𝑐𝑟 =
𝑙2
The critical stress 𝜎𝑐𝑟 for a given 𝑃𝑐𝑟 of a pin-ended column is given by,
𝜋2 𝐸
𝜎𝑐𝑟 =
𝑙 2
(𝑟 )
Where,
l is the length of the column,
r is the radius of gyration of C.S of the column,
E is the Young’s Modulus of the material of the column.
Buckling of Imperfect Column:
Let us assume that an imperfect column is initially bent and is subjected to an
increasingly axial load P. In this case, The B.M is always proportional to the change
in curvature of the column from its initial position. Thus,
𝑑2𝑦 𝑑 2 𝑦0
𝐸𝐼 − 𝐸𝐼 = −𝑃𝑦
𝑑𝑥 2 𝑑𝑥 2
𝑑2𝑦 2
𝑑 2 𝑦0 𝑃
2
+ 𝜆 𝑦 = 2
; 𝜆2 =
𝑑𝑥 𝑑𝑥 𝐸𝐼
The final deflected shape y is dependent on the unloaded shape y 0. Since any
arbitrary curve can be done using Fourier sine series where the initial shape of the
centre line can be assumed as,
𝛼
𝑛𝜋𝑥
𝑦0 = ∑ 𝐴𝑛 𝑠𝑖𝑛
𝑙
𝑛=𝑙
So,
𝛼
𝑑2𝑦 2
𝜋2 2
𝑛𝜋𝑥
+ 𝜆 𝑦 = − ∑ 𝑛 𝐴 𝑛 𝑠𝑖𝑛
𝑑𝑥 2 𝑙2 𝑙
𝑛=𝑙
We can write,
𝜆2 𝑙 2 𝑃𝑙 2 𝑃
𝛼= 2
= 2
=
𝜋 𝜋 𝐸𝐼 𝑃𝑐𝑟
Thus, 𝛼 is always less than 1 and approaches to 1 as P tends to 𝑃𝑐𝑟 . So we can say
that the first term of the series is dominating and therefore, a good approximation
results in neglecting of higher order terms which will change the equation to,
𝐴1 𝜋𝑥
𝑦= 𝑠𝑖𝑛
1−𝛼 𝑙
At the centre of the column, x = l/2,
𝐴1 𝐴1
𝑦= =
1−𝛼 1− 𝑃
𝑃𝑐𝑟
Where, 𝐴1 is the initial central deflection. If central deflection is measured from
initially bowed position,
𝑃𝑐𝑟
𝛿 = 𝑦 − 𝐴1 = 𝛿 − 𝐴1
𝑃
𝛿
A graph of 𝛿 vs 𝑃 has a slope in the region of critical load and slope is 𝑃𝑐𝑟 and
intercept is equal to initial central deflection. This graph is well known as “South-Well
Plot” for experimental determination of buckling load of an imperfect column.
PROCEDURE:
Fix the column in the testing set-up.
Mount two strain gauges back to back in the direction of loading on a chosen
location of the column.
Connect the strain gauges in half-bridge circuit with the two active gauges in
the adjacent arms. The sensitivity of the measured output is doubled in this
configuration and the temperature is also compensated.
Take bending strain readings for different values of the axial load P. Now plot
a graph of 𝛿 versus 𝛿 /P and draw a best fit straight line through all the points.
The slope of the straight line gives the buckling load 𝑃𝑐𝑟 .
Calculate theoretically the buckling loads of the column and compare the
results with the experimental values.
OBSERVATIONS:
Length of the column, l = 51.3 cm
Width of the column, w =
25.21+25.16+25.2+25.2
= 25.1925 𝑚𝑚
4
Thickness of the column, t = 1.68+1.66+1.69
3
= 1.67667 𝑚𝑚
Observation Table:
Load(N) Deflection in y-direction Strain in x-direction
(mm) (*10^-6)
0.1 0 0
17 0 0
27.1 0.03 1
41.4 0.11 3
50.2 0.19 5
55.4 0.26 7
58.8 0.31 9
62.7 0.35 10
62.4 0.46 13
76.6 0.66 19
79.3 0.78 24
81.9 0.84 26
84.8 1.07 33
87.1 1.17 36
90.9 1.27 39
89.8 1.47 46
92.3 1.6 50
90.9 1.7 54
The slope of this curve gives the experimental critical load, So,
𝑃𝑐𝑟,𝑒𝑥𝑝𝑡 = 95.7367 𝑁
RESULTS:
The experimental critical load is 𝑃𝑐𝑟,𝑒𝑥𝑝𝑡 = 95.7367 𝑁
The theoretical critical load is 𝑃𝑐𝑟,𝑡ℎ𝑒𝑜 = 99.4562 𝑁
The error between the experimental and the theoretical critical loads is,
|𝑃𝑐𝑟,𝑒𝑥𝑝𝑡 − 𝑃𝑐𝑟,𝑡ℎ𝑒𝑜 | |95.7367 − 99.4562|
𝐸𝑟𝑟𝑜𝑟 = ∗ 100 = ∗ 100 = 3.739%
𝑃𝑐𝑟,𝑡ℎ𝑒𝑜 99.4562
DISCUSSION:
There is an error that have crept in between theoretical and experimental
values. This error is much low which is very good under experimental
conditions.
There may be an error while measuring the distances and thickness using
scale and vernier callipers.
There can be error in reading of load cell.
There can be error in reading of digital gauge.
There may be an error in the readings using strain gauges.