Advanced Structural Analysis

Prof. Devdas Menon

Department of Civil Engineering
Indian Institute of Technology, Madras

Module - 1.3
Lecture - 03
Review of Basic Structural Analysis -1

We are now doing the third lecture on the first module - Review of Basic Structural
Analysis.

(Refer Slide Time: 00: 21)

1
If you recall, in the last session we covered a bit about statically determinate structures.
(Refer Slide Time: 00:31)

I think we were looking at beams and we stopped with the conjugate beam method. In
this session, I hope to finish that topic and so that we can do work and energy methods in
lecture four.

(Refer Slide Time: 00:42)

So, just to remind you, the book we are following is Structural Analysis and you need to
refer to the book to look at problems and the solved examples.

2
(Refer Slide Time: 00:59)

We were looking at the conjugate beam method. Here, you will see how it can be applied
to solving fixed beams. This is a fixed beam AB with arbitrary loading, gravity loading.
You can release the fixity and so you get a statically determinate structure which is a
simply supported beam. You can divide into two parts as you can see there, the one -
where you have the loading, and the other - where you apply the fixed end moments
which are hogging moments. MFAB and MFBA are at the two ends and you can now look
at the conjugate beam, which will be now free free because, the original beam is fixed
fixed and you will find that on the top, you will be putting the curvature diagram caused
by the negative moments. So, that is a straight line. At the bottom you will have the free
moment diagram M0 x by EI. These two must self-equilibrate. So, you will find that the
resultant force must be equal and the lines of action must also be equal. What is
interesting here is that, if we equate the two areas, you will find that the bottom areas
integral 0 to L M0 x dx by EI and there are some advantages you get when the loading is
symmetric.

3
(Refer Slide Time: 02:39)

So, let us look at those special cases, where you have symmetric loading and I will show
you here, two pictures of symmetric loading- the first is a standard concentrated load W
acting at mid span and the second is a uniformly distributed load of total magnitude W
acting at the mid span. The support reactions are W by 2, W by 2 and the two fixed
moments at the ends will be actually equal and opposite to each other. So, we can call it
just MF instead of MFAB and MFBA. So, if you were to draw the conjugate beams, you
will find that finding the fixed end moment is quite easy.

The area on the top is the area of that diagram MF L divided by EI and that must be
equal to the area of the free moment diagram divided by EI. So, if you equate the two,
you have this interesting equation, which suggests that the fixed end moment is nothing
but the average value of the free moment diagram because 1 by L integral M0 dx
actually means that it is an average value and so, it is easy to calculate.

If you take the first case, it is a triangle; so, the area is easy to calculate. The second case,
it is a parabola; the area is two-thirds of equivalent rectangle. You can easily find the
answers. The answers are: In the first case, WL by 8, MF is equal to WL by 8 and the
second case, it is WL by 12. So, you will find that, for symmetric loading it is extremely
easy to do. It will take you just about a minute or so to crack the problem. So, the
conjugate beam method is a very useful method to find fixed end moments when the
loading is symmetric.

4
(Refer Slide Time: 04:44)

When it is not symmetric, then you need two equations. This is another example that
needs a little more work but you will still get the results. Here, you see that if you have
load P acting not in the middle, then you will have two moments, which will not be equal
to each other. The larger moment will always be nearer to the load. So, in this case, if A
is less than B, MFAB will be numerically more than MFBA. You have to equate the area of
a trapezium with that of a triangle and you have to find their lines of action; this can be
done. We will not go through this, but these are the formulas that you need to recall;
especially, when you do displacement method. The fixed end moment formula here turns
out to be P into a into b square by L square for the moment MFAB, and P into a square
into b by L square for the moment MFBA. The previous case where you had a
concentrated load in the mid span is obviously a special case of this span, a is equal to b
is equal to L by 2.

5
(Refer Slide Time: 05:56)

Conjugate beam method can also be used to find fixed end moments in propped
cantilever beams. Here, you have only one unknown fixed end moment and here is a
convention that we will consistently follow. You will find that we are not using MFAB,
here, but we are putting small naught there in MF0AB to remind ourselves that the far end
B is a hinge and the small round that you see there is a reminder, and this also can be
done in exactly the same way.

Here, you will find that the conjugate beam is not free free; it is free-hinged. So, the
support reaction in the conjugate beam will take some reaction and that reaction will be
equal to the slope of the beam at B. Again, you equate the forces and you can get the
solutions. In this case, it works out to be q0 L square by 8 for a uniformly distributed
loading. So, you can draw the shear force diagram, bending moment diagram once you
get the fixed end moment because, once you get the fixed end moment, the beam is now
statically determinate. You know how to draw shear force and bending moment diagrams
for a statically determinate beam. The maximum bending moment will be where the
shear force changes sign. You are quite familiar with all this.

6
(Refer Slide Time: 07:23)

We are done with the topic of beams. We looked at how to draw shear force and bending
moment diagrams. We also did some interesting cases, where the bending moment
diagram was given to you and you have to figure out, what loading caused that bending
moment. We also looked at how to calculate deflections and slopes using direct
equilibrium method.

We now quickly move on to the topic of trusses. The basic assumption made in trusses is
that the axial forces are the only forces that act in the various members. Bending
moments and shear forces are assumed to be negligible and this is valid, if we assume
ideal pin joints at the junctions of various bars and if the loads are applied exactly at the
joints. Now, we know that the first assumption is strictly not true because you need to
have just need to (Remove 'need to') one rivet or one pin for free relative movement
between the members, but in practice you have more than one bolt in trusses.

In bridge trusses, you may have about thirty to forty bolts in a junction. If the elements
are welded, then it is actually rigid jointed connection, but that makes truss highly
statically indeterminate. Even if you were to do a very rigorous analysis, you will find
that the departure from the assumption of ideal pin joints is not very significant. You get
moments, but they are called secondary moments. For a first order analysis, we can
normally ignore them. In design practice, if it is a heavy bridge truss, we try to account
for it indirectly by assuming a 10 percent increase in the stresses. So, please note the

7
whole idea of doing first order structural analysis is to get quick results, which are
reasonably accurate. You are familiar with this conventional method of joints and the
tension coefficient method which is a generalized method of joints that can handle any
type of statically determinate truss with any loading. Then, you have the method of
section which should be applied intelligently to save time.

Here, is an example of two bar truss and you can see that the calculations are very easy
to work out. We need to take advantage of the fact, that there is an internal hinge at B
and there is no bending moment transferred from one member to the other. You only
have axial forces in the two members and it is very easy to compute. So, let us move
ahead, once you have the axial force in the two bars, then you also have the axial
deformations because, we know that there is a relationship between axial force and axial
deformation. We use a word elongation to refer to the axial deformation, where
elongation is strictly appropriate, if the axial force is tensile and this is what we assume
when the force is positive.

If the force is negative, then it is compressive and the elongation turns out to be a
contraction. What is the relationship between the axial force and the elongation?
Elongation is equal to what? [Noise] (Refer Slide Time: 10:56).

(Refer Slide Time: 11:02)

8
The elongation in the i th bar would be equal to the axial force in that bar divided by
length and the axial rigidity of that part. Therefore, this follows directly from Hookes
law. So, we can do that. Once, you got the bar elongation, you can locate the displaced
position of the joint B and you can do it geometrically.

I have shown a picture here, where AB has reduced in length and BC has reduced in
length, but if they were to be the way shown in this picture, then they do not meet but we
need to make them join. So, the only way to do that is to rotate AB about A and CB
about B till they meet. You can draw arcs or tangents to do that and that is how you
locate geometrically. That is how you construct and you can actually get the deflection.
We will find there are better ways of finding deflections.

(Refer Slide Time: 12:08)

Trusses are commonly used in roofs and they are also used commonly in bridges. (Refer
Slide Time: 12:26) Bridge trusses are referred to either through type trusses when the
deck is below and the traffic goes where you can see the two trusses on either side of
your carriageway, or they are called deck trusses when the truss is below the deck and
there are other intermediate types called semi through and so on.

9
(Refer Slide Time: 12:45)

Now, we first studied beams and then we looked at trusses. So, it is good to see the
equivalence between the truss and the beam. Here, I have shown two pictures of what are
called as N trusses because they look like the letter N. One of them does and the other is
a mirror image. Let me ask you a question - for the loading shown here, which of the two
is better? Which is more structurally efficient; the top one or the lower one (Remove
this) lower one? [Noise not audible] (Refer Slide Time: 13:20)Can one of you raise
your hands and answer? [Noise] Lower one. Top one. Answer is top one.(Remove this)
Why did you say the top? [Noise] The members take tension (Refer Slide Time:
13:34).Which member?

Let us talk about the top chord in both the members. Dont they both take roughly the
same order of axial compression? [Noise Bottom part] (Refer Slide Time: 13:34). What
is so special about the bottom? When you cut the beam at any place the reaction at the
left side support is upward and so, the shear acting when you cut will be downward. So,
the member there is aligned downward and it will be having a tension. So, in the second
truss, the diagonal members will be in tension and in the first truss diagonal members
will be in compression. That you get from the equivalence with the beam, when you do
this method of section. So, we should prefer the bars going into tension. Why? [Noise
not audible] (Refer Slide Time: 13:34) Because you have a problem associated with
buckling, the size of the members has to be larger and so on. So, these are the important
points to note. So, the only difference between the two is the diagonals; the alignment of

10
the diagonals. You see, overall, the truss deflects like a beam and if your right brain is
very good, you can actually see the contraction that is taking place in the top diagonals
and the elongation that is taking place in the bottom diagonals.

This is an equivalent beam (Refer Slide Time: 15:03) and you will find that the method
of section essentially establishes a relationship between the axial forces in the various
truss members with bending moment and shear force values in the equivalent beam over
the same span. So, if you cut a section, here for example, (Refer Slide Time: 15:03) if
you take moments about the point B, you can find out the bend.(Remove bend) the axial
force in the top chord in the second truss and the bottom chord in the first truss is given
by that bending moment at that location three, and where B (Remove 'where B') is
divided by the height of the truss. So, this is a quick way and those are the maximum
values of forces in the truss.

If you had to design a truss quickly, you can do it very fast; find out the equilibrium
bending moment in a beam and divide it by the height. So, this is a quick way of doing it
and you will find that, this is how it can work through the different axial forces in the top
chord and bottom chord. The axial forces will be the maximum in the middle just like the
bending moment is in middle region and slowly dies down and sometimes your end
members carry zero force. You see that picture (Refer Slide Time: 16:22) where the
rectangle becomes a rhombus. One of the diagonals increases in length and the other
diagonals decrease in length. So, that gives you a visible idea of why one diagonal goes
into compression and the other goes into tension and you can calculate those shear
forces.

11
(Refer Slide Time: 16:38)

The other simple thing that you can take advantage of is a fact that, whenever you have
two sets of lines, they need not be orthogonal; they could be inclined and the aligned in
the way which is shown here. Then the forces must be equal and opposite in any one
line. So, for example, I have shown you four pictures there and each of those lines could
be a bar. So, if it is a bar, then you are talking about an axial force in that bar. So, you
can take advantage of these pictures. If you look at the picture (d), you will find that if
say, one of those bars has zero force. Sorry, if we do not have the opposing force at that
joint, then you can take that force to be zero. So, you must develop the skill of
identifying which are the zero forces in a truss by inspection and also the ability to
identify whether the force that you get is tension or compression. So, there are many
problems and we will quickly see some examples.

12
(Refer Slide Time: 17:48)

Here is a truss which looks little complicated because you have larger number of
members and you can do a quick check and you find that m plus r is equal to 2j. So, it is
a simple truss, but you do not know where to start because of so many forces, but
perhaps some of them are zero and you need some skill to figure out. We have done this
problem earlier and so, I will go through it quickly. The first thing to recognize is if your
load is P acting downwards, those are your reactions and the reactions are statically
determinate. They all turned out to be P, then if you cut a section say, where I have
shown and you blow up that part you will find...

What can you make out from this picture of the section? [Noise not audible] (Refer
Slide Time: 18:42) N3 is equal to P - that follows from the fact that if you take moments
about G, then you can prove N3 is equal to P. Then, once N3 is equal to P, you go to the
joint D. You will find that necessarily there are many forces which are zero and you will
find that actually you can remove those bars once they are zero and your truss reduces to
this problem. (Refer Slide Time: 19:12). So, it is a much simpler truss to analyse and this
can be done easily. You can analyse by method of joints, triangle of forces.

13
(Refer Slide Time: 19:25)

This is an easy problem and is an extremely interesting problem. This actually was posed
by Timoshenko in his book on Structural Analysis and the normal way of cutting
sections does not work here. Wherever you cut a section you have far too many
unknowns and you know that when you deal with the plane truss, you can at the most
have three unknowns, cut section. If you are doing method of joints, you can have at
most two unknowns.

You will find that traditional techniques do not work here. Here is a clever way of doing
it and this again is a compound truss, but statically determinate. What you can do is - you
know (Remove you know) to view that segment as a triangle and separate the triangle,
which means using a pair of mental scissors and snipping through those connecting bars.
You expose those bars and here you have one look at this picture (Refer Slide Time:
20:36) and you know that there are only three unknown forces and you have three
equations equilibrium available. You can solve it or you can take the other triangle and
solve it. Once you get the idea, it is easy to calculate because you can use any three
equations and solve it.

14
(Refer Slide Time: 20:41)

Finally, the robust method of cracking any problem is the tension coefficient method,
which means you can take any joint and this could be useful, especially when you do
space trusses. You write the coordinates of those joints and you define the tension
coefficient as the axial force in any bar divided by the length of the bar. You can write a
generalized equation of equilibrium along the x direction, along the y direction and along
the z direction. So, you can systematically write this equation moving from joint to joint
and solve them. This is the generalized way of doing it. This is a way, you would
probably do if you had to do using computer to solve the problem because, you do not
take any shortcuts, but you just go through it and solve series of simultaneous equations.
You can make use of the direction cosines l, m, and n for each bar and this is a
systematic way of doing it. We are done with trusses for the time being. (Refer Slide
Time: 20:43).

15
(Refer Slide Time: 21:51)

Now, we move on to the next topic and this is a good interface. Look at this problem,
which we looked at earlier, you have a pin joint at B and what happens if you make that
joint rigid? Strictly, what all can happen? You have got axial forces in the first case.
Now, what is the extra that you can get? You can get bending moment because now the
joint B is rigid, which means you have a moment resisting joint. Now, if you get a
moment and the question I wish to ask you is - well, your vertical reactions do not
change; your horizontal reaction is your unknown; if you get a moment at B, is that
moment going to be sagging or hogging? I have shown here, sagging. Will this work and
why won't it work? No, that is not the reason. Why won't sagging moment work?

Now, to really understand this, try to visualize it. Visualize this by looking at deflected
shape, the possible deflected shape. Now please note, we normally assume, when you
deal with frames, this is a frame and not truss any more. We make the assumption and
their axial deformations are negligible. So, B does not really move down. B will move
down only if the two bars will reduce in length. So, you will have difficulty in drawing
the deflected shape because if you draw a deflected shape like this (Refer Slide Time:
23:32) and you have to preserve the angle at B between the two bars because it is a rigid
joint. If it were a hinged joint, you do not have to worry about that compatibility
requirement. So, you will always find that if one of the elements is shown sagging, the
other will be hogging but your arrows will not satisfy that and either you draw like this
or you have to draw like this, which means you will never get both sagging and hogging.

16
The deflected shape will not look like that. (Refer Slide Time: 23:57) So, you should
know intuitively that something is wrong here.

What is the conclusion that you draw from this? Though the rigid joint has the potential
to transmit a bending moment across the joint, in this particular instance, that bending
moment cannot be transferred and it will not work out. The bending moment is actually
going to be zero. So, it is like saying you have a savings bank account and the bank
account can take in any amount of cash value, but being a student very often it turns out
to be a zero. It just shows there is a capacity, but you need not have any value to that. So
in this case, you do not have that. There is another reason for this.

Can you tell me why there is no bending moment? [Noise not audible] perfect. This
shape is funicular for that particular loading. How do you recognize it is a funicular
shape? [Noise not audible] (Refer Slide Time: 25:18) If you were to assume that there
is a beam spanning across the support say, a simply supported beam and if you were to
draw a bending moment diagram and if the shape of the frame that you erect between
those supports resembles to some scale, the bending moment diagram, then you will get
only axial forces in that structure. That is called the funicular configuration and often we
miss this out because, once you recognize a funicular configuration, it is ideal.

(Refer Slide Time: 25:59)

The word funicular comes from the word funicle, which means a string and it is the
ideal shape. So, you will find that this shape is funicular. In this instance, AB and BC

17
will only have axial compression. If you flip it over, then it is like a cable and it will only
have axial tension. What is the main difference between arch and the cable? [Noise not
audible] (Refer Slide Time: 26:23) Yeah, who is answering? Yeah, arch takes
compression and cable has tension. Right, but there are some more differences between
an arch and a cable in addition to compression. The moment is always zero. [Noise not
audible] (Refer Slide Time: 26:45)

You see a cable and by definition is very flexible. An arch is rigid, is it not? It does not
move around and by virtue of its rigidity, it is capable of resisting; in addition to axial
compression, bending moment and shear force. A cable just does not have that ability.
So, a cable will always assume a funicular shape for a given loading, but an arch is
something that you can make at the most funicular for one loading, but when some other
loads act on that arch, it will be non-funicular and there will be bending moments and
shear forces in addition to the axial forces. This is important to understand.

(Refer Slide Time: 27:34)

To fully understand that bending moment relationship, look at that boulder on top; it is
simply supported and there is no horizontal reaction. Look at the frame, which is also
simply supported. In both incidents, the distribution of bending moment along the x- axis
will resemble the triangle, if you have a concentrative load.

If you are able to generate a horizontal reaction at the two supports, then that has a
tremendously beneficial effect because at any section you get a hogging moment, which

18
is that horizontal reaction times a height y and your net moment at any section is now
reduced to this free moment or the sagging moment M0 x minus hogging moment H into
y. A Funicular shape is one, where M0 is equal to H into y or rather you can say y is
equal to M0 by H and if H is a constant. So, it clearly follows the shape of y, the ordinate
of your frame at any location resembles your bending moment diagram, then you get a
funicular shape. So, it is very easy to appreciate this point.

(Refer Slide Time: 29:01)

So, here is something on cables. Cables are very interesting to analyse. Again, you can
compare with a simply supported beam. You get shear forces in the beam, you get
bending moments in the cable. The cable will be made up of a series of straight lines, if
the loads are acting on a series of concentrative loads, then the vertical reactions are very
easy to calculate as you would in a simply supported beam. Let us say the total load is W
and it is located at the distance 'a' from the left end and 'b' from the right end. Then VA
and VB are very easy to calculate.

The horizontal reaction is actually obtained by the bending moment, the free moment in
the equivalent beam divided by the height at that location. So, you can work out the
heights at different locations, once you know the bending moment. All you need is a
height at any one location and usually you know the height at maximum dip of the cable
and that height is called sag. Here, I have shown hc and so you can you can get the
horizontal reactions. If a cable is inclined as shown, then the axial force in that cable is

19
related to the shear force in the beam, as well as the horizontal force at the two ends.
Now, if you cut a section anywhere, you find that the horizontal component is constant
and sometimes, it is referred to as horizontal tension in a cable. In an arch, it is referred
to as horizontal thrust. So, you will find that H is constant for all cables and for all
segments. V is changing depending on the shear force and the axial force is the resultant
of H and V as shown there. (Refer Slide Time: 31:04)

If the support is at two different levels, then you can relate it to the first case, you will
find that if the arrangement of loads is same, the horizontal force does not change. We
can prove that it is a vertical force that changes, you now get a couple, the two H is
separated by a level difference and is self-equilibrated by reduction and an increase in
the vertical forces. So, this can be proved. We will move ahead. The difference between
the first case, where the two supports are at same elevation, and the second case, where
the two supports are at different elevation is only in the vertical reactions. The horizontal
force is the same. VA increases to some extent and VB reduces.

(Refer Slide Time: 32:01)

If the cable is subjected to distributed loads, then it is not going to be made up of straight
lines, it will be curved and you can actually generate an equation which relates geometry
of the cable to the forces in the cable. If you take vertical equilibrium of a segment of the
cable, you get this beautiful differential equation, which is similar to the equation that we
derived for a beam. Is it not? So, q here is a load intensity and it is related to the shape of

20
the cable and if you take the slope of the cable as y dash, then you have a direct relation
between theta and the cable forces; tan theta is V by H. V can change depending on the
location of x, but H is a constant if you have only gravity loads, and that is equal to dy by
dx. If you substitute this value in the first equation, you get another equation where y of
x comes into play and this is a basic differential equation of a cable.

The axial force can be easily generated from the triangle of forces and you can use this
equation to work out what is the profile of a cable for any given distribution and loading?
The most common type of distributed loading is uniformly distributed load and this can
be done into two ways: one is uniformly distributed along the horizontal span, the other
is uniformly distributed along the curved span. When do you encounter these two cases?
Well, self-weight is uniformly distributed along the curved span and the shape of that
cable is called catenary from the word catena, which means chain, but this is not that
common in practice.

In practice, we use cables to support structures, to supports roofs, to support bridge.

Cable Suspended bridges are of two types: you have the cable stayed bridges, where the
cables are straight and they have force attention. Each cable will have a different force
and you have the classic suspension bridges. What is the shape of a cable in suspension
bridges? [Noise] It is parabolic (Refer Slide Time: 34:31). Why is it parabolic? Uniform
loading, it is parabolic with the idea that the parabola is the ideal shape, the funicular
shape, when you have a uniformly distributed load. So, you need to ensure that it will
remain a parabola even when vehicles move on the bridge. How do you do that?

You have to make the deck stiff and these are the equations (Refer Slide Time: 34:59)
this relationship is very simple. If the sag in the cable is H in the middle, then the
horizontal tension in the cable is nothing but the bending moment, which is q0 l square by
8 divided by H. So, it is a very simple formula and you can work out the rest of the
equation very simply.

21
(Refer Slide Time: 35:30)

Here, is an example of one of the most famous suspension bridges way back in 1937.
This bridge was designed, analyzed by using simple manual calculations, the kind of
formulas that we saw in the previous slide, but todays engineers and todays students
would rather use a computer to solve this problem, but at least for a first order analysis, I
think it is important to know the simple calculations for a parabola.

(Refer Slide Time: 35:57)

Here, is an example of cable suspended bridge. The girder that you saw in the previous
slide at the bottom is called the stiffening girder. It has to be fairly deep and rigid

22
because, the idea is it should not allow the cable on top to change its shape. All the load
that goes on the deck is actually transmitted to the parabolic cable through those vertical
elements called suspenders. If the cable is parabolic, then you can safely assume
suspenders carry a uniformly distributed load. This can be facilitated by providing a
hinge at the center as shown here (Refer Slide Time: 36:38). Because, if you provide a
hinge at the center and you make this assumption of uniformly distributed load, this
problem becomes statically determinate; otherwise, it is indeterminate. If it is
indeterminate, then actually the force in every cable can also be different and that is a
real problem, but this is simplification.

If you isolate the cable, you will find that there is only one unknown - uniformly
distribute load q0. If you know q0, you will know the rest of the problem. How do you
solve for q0? Well, if you take a free body of that stiffening girder and you analyze it,
you take moments about one of the ends; you can find the relationship between the
vertical reaction and q0. Next, you take that beam apart and you invoke the equation of
the condition that, at the hinge there is no moment transfer and MC equal to 0. Then, you
will find that you can solve this equation and get another relationship from which you
can prove and establish the value of q0. Once you know q0, both systems can be
separated out and the cable is easy to analyze and design.

In olden days, the real worry was about (Remove about), what is the size of the cable to
be used - the wire rope. So, you need to have an idea of what is a maximum tension in
the cable. So, it is very simple. Find the total load that can come on the bridge and half
the load will go to one end - on the cable - that is you need is value of q0 and q0 into L by
2 is a vertical reaction and H is q0 L square by 8H and the resultant of that is the axial
force. If you know the allowable stress in that cable, you can work out backwards on
what is the diameter you need. You apply some factor of safety and that is how it needs
to be done. It is a simple straightforward calculation.

23
(Refer Slide Time: 38:52)

Now, in this problem you can solve and work this out, you can get the results. Once you
separate out the stiffening girder, then you can work out, statically determinate. If you
cut any section, you can work out shear force and bending moment diagrams.

Here is the picture of something we did (Refer Slide Time: 38:58). I think in 2004 or so,
your senior students built this bridge, made entirely of newspaper and rope. So, it is a
cable stayed structure, where they use these simple analytical tools to do it. You will find
that the two vertical pylons are made up of newspaper. Lot of testing was done in our
laboratory, those ropes are manila, and they actually want to put this up in the Guinness
book of world records. I do not think they succeeded, but it did come out as a journal
paper in the ASCE journal for Professional Issues in Engineering Education and Practice.
Anita Rao was a civil engineering student, who led that team of about 100 other students
and you can look at this article to see the history of how this was done.(Refer Slide
Time: 39:10)

24
(Refer Slide Time: 39:53)

Funicular arches are ideal arches. If you have concentrated load, they have to be made up
of straight lines. Sometimes, they are called linear arches and they take this shape. They
are very easy to derive. Now, you can find the horizontal thrust by simply taking the
bending moment at any point and dividing it by the height of that point. So, these are
called linear arches.

(Refer Slide Time: 40:23)

25
This is a parabolic arch. Where do you encounter these arches? [Noise] Bridges. Yes, in
bowstring girder bridges. You can also encounter them in buildings and I will show you
picture, where we did this. (Refer Slide Time: 40:36).

Now, this is a two-hinged arch, which is funicular. The boundary conditions are not
important. If you can have uniformly distributed load along the horizontal span, parabola
is the ideal shape. If you have a curved, if you have a self-weight kind of loading, then
the catenary arch is the ideal shape. You can build walls with this shape and traditionally
this was done. Here, some work (Refer Slide Time: 41:10) that we did many years ago,
where we built these catenary shapes vaulted structures made up of coconut shells, which
can take lot of compressive load. The entire structure is made up of coconut shells with
plaster and with a catenary shaped formwork inside which can be dismantled and reused.
(Replace 'when we use' to 'and reused') .

This is a favorite quiz question. (Refer Slide Time: 41:35) that you see in this picture,
three bridges; iron arch bridges built during 1850-1900 over the Harlem river in New
York. What is special about these three bridges? All of them refer to three stages, three
historical advents in the history of structural analysis. What is so special about them?
Here, the first is a three-hinged arch and you can barely see that hinge in the middle
(Refer Slide Time: 42:09). The three-hinged arch is a statically determinate structure and
as we discussed earlier, people were afraid of building indeterminate structures. They got
bolder and they went for the two-hinged arch, and later, the fixed arch. We will study
this in the next module.

26
(Refer Slide Time: 43:15)

How to analyze indeterminate arches? Here are some useful principles. I think I have
asked you this question earlier, but here goes again. Take a simply supported beam and
let us say, half the beam is loaded and the deflection at the mid span is delta1. Let us say
the full beam is loaded, the deflection is delta in that beam. Can you say that you can
get a formula for delta, if you wish. Can we say that the mid span deflection delta1 is half
of delta? Yes or No? Yes. [Noise] but if you remember, last time I asked you this
question and all of you said no. so, why do we say yes? because of parity.

Now to understand parity, you take another beam with the other half-loaded and you
apply the Principle of Superposition. You will find that the sum delta1 plus delta2 must
be equal to delta, but you can argue that delta1 and delta2 are the same. Because, let us
say, we do an experiment and we load this half of the beam and we measure the
deflection at the mid span. What you would draw is one, in which the right half is
loaded. The picture that I will draw is the one with left half loaded, but we are actually
dealing with the same beam and this is not symmetry, this is called parity and this is
called the Principle of Parity. It all depends on how you look at the beam and the
deflection is exactly the same.

If you invoke the principles of parity and superposition, you can prove that delta1 plus
delta2 equal to delta and delta1 is equal to delta2, and so, it follows that delta1 is delta by
2. This combination of parity and superposition, we can invoke in many cases.

27
(Refer Slide Time: 44:18)

For example, in this two-hinged arch subjected to a distributed load on one-half, is this is
a funicular arch? No, because the bending moment diagram for this is not parabola, but
still, can you say it is statically determinate? What is the horizontal reaction? Vertical
reactions are known. What is the horizontal reaction? [Noise] (Refer Slide time: 44:54)
Tell me the basis of how to find. What should we do? Do not cut the arch. You load the
other half. We do not know H and you load the other half and when you add these two
you end up with funicular arch, whose horizontal reaction is known. What is the value? It
is q0 L square by 8H. How do you get the horizontal thrust in any one of these arches? It
will be equal to half that value. That is it.

So, when you put them all together, you get 2H [Remove 'two edge'] and you can easily
work this out. If you go one step further, once you know H it [Remove 'edge'] is
statically determinate and you can draw the bending moment diagram and shear force
diagram for this structure and you will find that the bending moment at the crown is zero.
Why will the crown moment be zero? [Noise] Why will the crown moment be zero? Yes,
it is like a three-hinged arch because if the moment there is zero, it is like there is
(Remove 'you call') a hinge at the crown. Why will it be zero? [Noise] (Refer Slide time:
46:31) One side it is hogging and one side it is sagging.

If we look from the other side... If it is not zero, it should be either sagging or hogging, in
which case, in the total arch should be two times that moment because, you are adding

28
and you are doing superposition but you will not be able to satisfy the relationships. The
only answer is it has to be zero and you can actually draw the bending moment diagram.

Can you explain why the bending moments in the left half of the arch will be sagging
and those in the right half will be hogging? Any sound argument? Yes.

[Noise] (Refer Slide Time: 47:29) Balance the moments. If (Remove 'when') we take it
individually and suppose we consider the loads in left side and it would be sagging, at
the same time we consider the right side, it would be sagging and this is not possible at
the same time.

No, that is the reason why the moment at the crown is zero. Why is left half is (Remove
'is') sagging and the right half hogging?

[Noise] (Refer Slide Time: 47:53) Because, it is the left half where the loads are applied.

No, intuitively you are right, but that is not good enough. Do you remember Eddies
theorem? So, you have to invoke Eddies theorem; we will look at that soon.

(Refer Slide Time: 48:12)

Let us say, you have a load, concentrated load acting on a three-hinged arch like this.
This is certainly not a funicular arch. Which would be a funicular arch? This would be a
funicular arch. So, you superimpose both of them with the same height at the crown and

29
you will find that the deviation from the funicular arch is actually a measure of your
bending moment. So, that is what you have to do.

So, if you go back to the previous case, (Refer Slide Time: 48:52) you will find that the
bending moment diagram would be as indicated there in the funicular arch. You gave
more height than what is required for the funicular arch in the left half. If you take the
equation, M0 minus Hy and minus Hy is more than M0. so, you get hogging. On the other
side, you give less 'y' than the required one, and so, it is sagging. It is a simple argument.
Let us move ahead. The other thing you need to know is - you will also get shear forces
and axial forces.

(Refer Slide Time: 49:24)

You can actually resolve the vertical force and the horizontal thrust into axial and shear
forces using geometry of the arch, you have studied all this. These are simple examples
that can be solved and these type of arches are often either segmental, which means part
of a circle for which you can derive the equations or they are parabolic and we do not
have time right now to solve a full example, but you studied this and it is pretty easy to
solve and you can also invoke Eddies theorem to draw shape of the diagram, you can
write the expressions for bending moments and shear force and they will look like this
(Refer Slide Time: 50:11).

You can also calculate the axial force and shear force. Once you have axial force and...
the axial force based on the shear force, based on V and H, that is, the vertical force and

30
the horizontal thrust. So, you will find that there is not too much difference between the
segmental arch and the parabolic arch because the shapes are more or less same. For this
reason, you can even idealize the catenary arch as the parabolic arch which is easier to
handle.

(Refer Slide Time: 50:50)

We come to the end of this topic, this is a frame suspended by its own weight. So, it is an
interesting question: can you draw the bending moment diagram for this frame? So, what
you need to do is to see how it will look, if you hang that frame? You will find that the
vertical line from A must pass through the center of gravity. You need to locate the
center of gravity and then you can draw the free bodies of these ends. So, typically
frames are like this (Refer Slide Time: 51:18). In a cantilever frame, it is wise to start
from the free end of the cantilever and your bending moment diagram will look like this
(Refer Slide Time: 51:24) and a rough shape of the deflected shape look like this. So,
intuitively this must appeal to you. You must also note that this diagram is drawn on the
tension side. So, the curvature must give you the feeling of tension on the right side.

31
(Refer Slide Time: 51:45)

Some concluding points: this is the favourite question, which I can assure you, one in
thousand engineers answers correctly, but since you studied this, you will probably give
me the right answer. Can you draw the bending moment diagram? Most people will draw
this. Is this right or wrong? [Noise] It is wrong. Why is it wrong? Because there is a
vertical component and there is no bending moment. Yeah, there cannot be bending
moments in the vertical components, that is, the columns. Why not? [Noise] Sir, only
axial loads are there. Why only axial loads? Because, only vertical reactions are there
and no horizontal reaction. You cannot have a horizontal reaction at the supports in the
simply supported beams, subjected to only vertical loading. So, if the horizontal reaction
is not there, you cannot get a bending moment in that vertical element. So, this is wrong.
So, that structure will be a just like a simply supported beam and the bending moment
diagram is just like a simply supported beam. Those two vertical elements are like some
legs and props holding it together.

What about the deflected shape? That is a real challenge to draw. You will find that
many people draw the diagram like this, (Refer Slide Time: 53:01) which is not correct.
Why it is not correct? [Noise not audible] (Refer Slide Time: 53:10) Because the
vertical elements are seen to be bending here, but there is no bending moment. So, what
should you do? Well, you first reproduce the simply supported beam drawing there. Why
should you reproduce it? Because the curvatures are the same and the bending moment
diagram is the same. What should you do for the vertical element? [Noise] 90 degree,

32
you preserve that. This is wrong and this would be violating the 90 degrees. So, you have
to draw like this (Refer Slide Time: 53:42) and this is still feasible because, the roller
support allows the movement. So, the original shape was this (Refer Slide Time: 53:49).
So, a student who can answer this completely is a rare student.

(Refer Slide Time: 54:02)

Now, we have drawn this picture earlier, you will find this bending moment diagram is
correct, if there is no roller support and if both are hinged, then the picture is correct.
Then how to find the reaction? It can be done from superposition. So, take the first case,
where it is roller and that is a primary structure, you can move it to the right and so you
need to apply that much force horizontally to eliminate that. That is how this can be
proved.

33
(Refer Slide Time: 54:28)

One last picture - this is the last topic. Is this structure statically determinate or
indeterminate? - This is a frame. [Noise] It is indeterminate because you have four
unknown reactions. So, how do you say it is determinate? [Noise] (Refer Slide Time:
54:48) By parity; we separate it into two diagrams. Which diagram? Sir, in one diagram,
it is P by 2 and P by 2. Very good. So, I will help you there. First of all, the vertical
reactions are statically determinate, but the horizontal reactions are not. So, as you
rightly say, you divide. This frame is symmetric; that is the reason, why it is going to be
determinate. So, you divide two loading diagrams. One is the symmetric loading where
you have P by 2, P by 2 and you have an asymmetric loading or anti-symmetric loading,
where the P by 2 act in the opposite direction.

In the first diagram, if we assume the axial deformation is negligible, what is the
deflected shape? There is no deflected shape. It will remain as it is because there is no
change in shape. You have only an axial force. Whereas in the second one, if you invoke
parity and superposition, you will find that you will have to have P by 2, P by 2 and so
the deflected shape will look like that (Refer Slide Time: 55:56) and the deflection
comes from the anti-symmetric loading part. So, once you say P is shared equally P by 2
and P by 2, the indeterminacy is gone. You can analyze this and you can draw the
bending moment of the diagram. Thank you.

34
Keywords : conjugate beam, truss, arch, cable, funicular, axial, thrust

35