Lecture 41 : Lagrange's equation of first kind

IIT Kanpur July 201
20 followers
Sep 26, 2024
Hello, and welcome to another lecture of classical motion of a single particle. In this lecture,
we simply continue our previous discussion on motion of a particle under different type
of constraints. So, already I have introduced several type of constraints, if you remember
that before even classifying the constraints, I defined constraints in the framework of
classical mechanics. Constraints are nothing but the conditions or some limitations imposed
on the kinematics of the particle ok. So, I am just trying to emphasize on this point
that if for example, if the force nature is specifically following some symmetric property
for example, like the central force that force is a function of the modulus of r ok and force
is directed along the line joining the particle and some source or the origin of the force
ok. So, this is a constraint on the force if you
want to say ok. So, this is a limitation or it is a specific type of force. So, can we
call this is a constraint? No, in this context we will not call that as a constraint ok.
So, we will simply try to do the constraints I mean try to limit ourselves to conditions
and limitations which are imposed on the trajectory on the velocity of the particle ok. So, you
can also have something a little bit complicated than that. So, for example, you can say that
some combination of velocity is related to some combination of the position coordinates
ok that is also possible. So, for example, in the last lecture I give
an example of this if you have a constraint like where x dot is equal to lambda times
x y ok lambda is a constant, then your x component of velocity is actually proportional to the
product of the position coordinates x and y ok. So, in this case you see that this is
not I mean just merely a constraint only on the trajectory or only on the velocity, but
it is a mixture of 2. Also we can say that the constraints can consist of time explicitly
and so, for example, just for a simple pendulum whose length is changing with time or an inclined
plane whose angle is changing with time they can be thought to be the examples of constraints
with explicit time dependence and those constraints we called that rheonomic constraints whereas,
the constraints without explicit time dependence we called scleronomic constraints.
So, once again so, rheonomic and scleronomic that is depending on their time dependence
of course, dependence means explicit dependence ok. Then we talked about unilateral and bilateral
constraints we said that if a constraint is a form of equation then this is a bilateral
constraint if it is consisting of inequalities or a combination of inequality and equality
then it is called a unilateral constraint. Then the most important one the holonomic
constraint and non-holonomic constraint ok. So, holonomic we said that it has several
features, first it should be an integrable form of equation which we sometimes loosely
say algebraic, but it actually can also consist of trigonometric or exponential type of functions
as well which we know how to integrate ok. So, I can actually write algebraic or trigonometric
or exponential or a mix mixture of them some functions like that ok.
Then it should be always so, it is an equation it should be integrable then it can consist
only position coordinates and at best time explicitly ok. Finally, we say that holonomic
constraint is such a constraint which is actually able to reduce the number of useful
coordinates in the problem. So, just remember the case for a simple pendulum where if we
say that the length L of the pendulum is fixed in time then this is x direction this is z
direction ok, then x square plus z square is equal to L square makes actually the situation
simpler because then you can actually write every x to be equal to root over L square
minus z square ok. And of course, you can take plus minus according
to your convention ok. For example, here in this problem in this problem z is directing
vertically upward. So, we know all these coordinates vertically downward I mean from
this point which can be considered as origin ok. From this point every displacement is
having a coordinate z coordinate which is negative. So, z should be negative. So, when
we calculate z that will be actually minus root over L square minus x square ok. These
things we should just keep in mind ok while solving a problem.
Then we also said that it is not always compulsory that the a priori form of a holonomic constraint
is containing only position coordinate and time. Sometimes it can be possible that the
a priori form that the apparent form of the holonomic constraint is actually containing
velocities, but in such a way that they can be integrated to reduce the equation in a
form where you can actually have x, y, z and t. So, the position coordinates and the
time. So, for example, I said that if you simply
say x dot equal to lambda z dot for a 2D problem like this ok, then actually this is nothing
but equal to x is equal to lambda z plus c. This c you can get rid of ok by just choosing
suitable initial conditions ok. So, you can simply say that ok I choose my origin in such
a way that when z is equal to 0 my x is equal to 0, then actually c is 0 and your x is equal
to simply lambda z ok. So, you are in any case you see this is nothing
but a holonomic constraint ok. And this type of forms which are actually holonomic, but
does not look like a holonomic at first sight ok they are known as pseudo holonomic constraints
ok. Then we said that in nature whenever a particle is undergoing a holonomic constraint
most of the cases they are associated with some force of constraints and for example,
in the case of a simple pendulum. So, I can draw here in the case of the simple pendulum
ok. So, this is at any arbitrary position ok, if I say that the string is of fixed length
then it always has some t to maintain the length. Also if I say that length is not constant,
but is a function of time ok. So, let us say L 0 1 plus sin omega t ok even then it should
have a tension ok, but that tension can actually vary now ok.
But then we actually also showed that one good thing is that. So, if the length is just
independent of time is constant in time, then in this case the tension force is actually
perpendicular to the direction of the motion of the particle ok that we all know. Now if
we are considering another case where a particle is actually sliding ok, a particle is actually
moving I should not use the word slide because it somehow assumes a surface ok. So, just
the particle is moving downwards along this inclined plane with this angle theta and we
are simply saying that it is always necessary for the particle we impose the condition that
the particle should be always at the contact of this surface, then we know that there should
be due to this contact there should be a normal reaction force R ok.
And which is also if you see that which is also perpendicular to the direction of the
motion of the particle. However, in this case also I said that if your theta changes with
time for example, let us say at a time instant t 1 the theta is theta 1, but at time instant
t 2 the theta is actually let us say I mean I would actually do something like this a
different surface like this theta 2 and then what this particle would have actually the
particle will at this point ok. So, you see now in order to maintain the constraint
that the particle will always be on the surface of this inclined plane the particle should
go in the direction which is like this, but again the reaction force is directed perpendicular
to the inclined plane where the particle is instantaneously located. And that is why you
will see that whenever this theta is also changing as a function of time then this reaction
force is no longer perpendicular to this case. And this is also very similar to the case
of this pendulum where the length of the string is actually changing explicitly with
time ok. In both cases we said that we have one smart
trick to get rid of this thing first of all if this reaction forces and this tension forces
can be somehow classified under a universal umbrella, we can actually think that they
are actually coming because we are posing some constraint and due to that constraint
there is a somehow a contact or some commitment of that medium. For example, when whenever
this mass is attached to the string the string is actually trying to respond to that attachment
of mass. Also when the particle is moving on this surface the particle touches the surface
and then the surface also wants to respond to the particle ok. And that is why they are
called somehow some forces of constraints and which are actually coming from a very
intrinsic property like elasticity, rigidity all this type of properties of the material
ok on which or to which the particle is attached. So, these forces of constraints in general
in general for our usual everyday problem associating one particle motion we see that
they are actually perpendicular to the infinite decimal displacement of the particle ok.
Just for example, here the reaction force here the tension force, but if this constraint
force is an explicit function of time then this is not necessarily perpendicular. So,
if you just consider a real displacement then the work done by the forces of constraint
is 0 when you are talking about time independent constraints, but if your constraints are explicit
functions of time then basically you have to consider a new thing which I introduced
that is the virtual displacement. We can actually show if we instead of considering
real displacement if we constitute a virtual displacement this is a displacement which
you cannot see in during the motion of the particle, but what you can actually imagine
and you actually can visualize virtually in your brain that is in your brain due to I
mean between the two possible configurations ok both are compatible with your constraints,
but at a given time instant. What is the meaning? The meaning is that you can just prepare your
system ok and you say ok my simple pendulum system let us say I have prepared this in
my lab and at some t is equal to 0 I am just letting this particle to move from this point.
In another experimental setup you just make an initial displacement which is
here ok. In both these experiments you see that there will be actually
two I mean there will be a discrepancy in this ok, but this is not possible in one single
experiment or you can say I am just setting up one experiment and this is just a thought
experiment I am thinking that in my brain for the second one ok and then you say at
t is equal to 0 in my real system the particles initial position is let us say x 0 1 and in
my brain I have another configuration where the particle starts from x 0 2. So, x 0 1
minus x 0 2 is the virtual displacement ok for a for the coordinate x ok.
We now for virtual displacement concept we always use the concept of infinitesimal
displacement ok otherwise there are some other problems ok. So, just for this course this is
very important to understand that virtual displacements are useful in infinitesimal sense
ok. So, that means, we always write delta r ok. So, d r means that your time is evolving
and that is why your particle is starting from this point it may come to this point ok.
Virtual displacement says that no you are just making two setups maybe two in reality or you
are making one setup in reality and another one you are thinking in your brain and then you fix
your time point. So, although this is a function of time now since you are fixing a time point
let us say t is equal to 0 or t is equal to 2 your L is now fixed and that is why your imaginary
system ok will also have a position which actually remains on the arc of the same circle as that of
your real system ok for example, for a given instant of time given instant ok.
So, that is very important if they are all satisfied then we say that F c. So,
the constraint force dot delta r is equal to 0 ok that is the principle of virtual work we said
and from that also we said that ok if for a particle of mass m if r double dot is the
acceleration then the total force acting on this is the force due to some external field
ok on some external effect and the force. So, which I called F i and some force of constraint
F c then if I do a contraction with delta r in both sides you will actually have this term to be
equal to 0 and that is why you will have m r double dot minus F i ok dot delta r is equal to 0
ok. Here I did not say that this is 0 this is simply perpendicular to this
ok both for rheonomic and scleronomic holonomic constraints ok and that is why this is an equation
an equation of dynamics as well ok where you can solve the systems of the motion of the particle
even without introducing the constraint forces ok because you see the constraint
forces are going away. So, this is known as the D'Alembert's principle.
So, today we will start from there and after this I presented a simple example of a simple pendulum
ok where the length of the string is fixed and I told you to try to use this principle of D'Alembert
ok for solving the motion of a simple pendulum whose length of the string is changing with time
ok. Then today I will present another way a subtly different way from this one this principle,
but where we will also use the constraints explicitly with their mathematical forms.
So, in today's lecture we will discuss actually the method of Lagrange equation of first kind.
So, you can easily see that I can actually say that from D'Alembert's principle
from D'Alembert's principle I can say that M r double dot minus F i dot delta r is equal to 0
that we know ok and we actually used that thing ok along with the constraint itself.
So, we then we said that ok if there is a constraints let us say ok. So, from this
you will have some conditions and then you separately use the constraint and then you
like actually eliminate those delta x delta y delta z and finally, you find the solution of
motion ok finally, you find the equation of dynamics ok. Now from D'Alembert's principle
we can get this here we will do something much more interesting we actually involve
the constraints together with this equation. So, now, we consider a constraint which is
holonomic ok. So, holonomic means that they should have a function phi which can involve
maximum all this 3 coordinates and t is equal to 0. Let us say I am just talking of a very simple
system one single particle moving in 3D space and I have one constraint one holonomic constraint
ok. So, which is given by this ok that can be actually if you see. So, this function ok.
So, for a simple pendulum moving in x z plane this phi is given by x square plus z square minus
L square ok. For a particle moving on the surface of a sphere of radius capital R is given by x
square plus y square plus z square minus R square ok. Of course, here if L is a function of t then
instead of just writing L as constant you can actually write also L 0 square 1 plus for example,
sin omega t whole square ok. Here you can also write some R if R is equal to some alpha t
then you can write alpha t whole square like this ok. So, this is somehow the general form of a
holonomic constraint. If this is a holonomic constraint what is the
differential form of this holonomic constraint d phi. So, this is actually a scalar function here
ok and so, d phi will simply be equal to del phi del x dx plus del phi del y dy plus del phi del z
dz plus del phi del t dt ok. So, this is when we are just talking of change in phi ok in a real
evolution where you have real displacements and real interval of time. But if we now do
for virtual displacements delta R ok where where time is frozen ok we have
d phi will be equal to. So, then I will actually do not use d phi I will use this thing as delta
phi ok and that will be equal to del phi del x delta x plus del phi del y delta y plus del phi
del z delta z you see all these 3 things they are actually virtual ok.
What about this term there will be no term here ok which is a counterpart of this term
ok. So, because delta t is 0 as delta t by definition is 0 ok. So, I will have simply this.
If you have this thing then you actually have this is equal to because phi is 0
ok this is the definition this will also be equal to 0 ok. Now you can also write this thing
in Cartesian coordinate system ok. So, here we have used Cartesian coordinate system
ok you can also write here m x double dot minus f x ok delta x plus m y double dot minus f y
delta y plus m z double dot minus f z delta z that is equal to 0.
So, I have now 2 equations one is this where delta x delta y delta z are related by some
other functions and finally, gives a result which is 0 ok and I have this where also
some combinations of delta x delta y and delta z they are adding up to be 0 ok.
So, of course, you know that if I add 2 0s I will also have 0, but I can do something
more general I can actually write I can take one of this equation intact and actually I can
multiply an arbitrary scalar quantity let us say number let us say lambda ok. If I just multiply
with this this will remain 0 and if I actually add this now to this one it will still remain 0.
So, mathematically if you say that ok you have 2 equations one is like this
and another is like this ok then since both of them are 0 instead of just saying that
this plus this is 0 another more general possibility is that you write some constant
time this one plus some constant time this one is also 0 instead of saying that individually
they are 0 and their exact sum is 0 another more general mathematical conclusion is that
if you multiply some arbitrary function here. So, let us say some M 1 here and some M 2 here
ok and then you add then also you will get 0. So, instead of M 1 and M 2 I just say that
call this lambda as M 2 by M 1 ok and I just get rid of 2 functions I just say that ok I just keep
this equation intact and multiply this equation by a function lambda ok, but it should be multiplied
uniformly by this function lambda and then if you add you will still have 0.
So, finally, adding the above equations we get M x double dot minus F x plus lambda del phi del x
delta x plus M y double dot minus F y plus lambda del phi del y delta y ok and finally,
you can say plus M z double dot minus F z plus lambda del phi del z delta z is equal to 0.
Once again you see in this equation there is no force of constraint ok, but now we have written
these two explicitly that is the d'Alembert's principle and the part of the constraint
explicitly together ok. So, if we write these things then we actually can say ok what is the
vector form of this. So, vector form is simply A M r double dot minus A F ok or as you wrote F I
for example, the part of force which is not related to the constraints ok and then you will
have plus lambda. So, del phi del x you can take like just component wise you can now think that
this is coming from. So, these things actually come from this vector lambda del phi del x i cap
plus del phi del y j cap plus del phi del z k cap and which is nothing, but equal to lambda
gradient of phi ok. So, I can write lambda gradient of phi is equal to 0.
This is known as the Lagrange's equation of first kind. So, this is actually not
primarily very different from d'Alembert's principle. d'Alembert's principle said that
ok I have this part this first two terms dot delta r is equal to 0 and then I again take
separately the constraints and I actually from that constraints I actually try to find a form
where something dot delta r is also 0 and then I do the same thing here ok.
On the other hand here what I am doing doing right from the beginning I take the constraint
ok taking this as for simplicity as a holonomic constraint and for which you have this actually
specific form this form and you finally, can show that this is given by this one ok.
So, this is a little bit at this point I have done a cheating in general.
It is not a cheating I am I am coming to that point. So, I have said ok if this is 0 is this
plus this plus this is equal to 0 this is equal to 0. This is possible only when we assume or
we can show that delta x delta y and delta z they are arbitrarily chosen or they can be
arbitrarily chosen that means, they are independent. But remember delta x delta y delta z they are
related by constraints. So, they are not independent quantities ok. See if they are not
independence quantities because they should always be related by the constraint equation.
So, for example, if you just remember that x square plus z square is equal to L square
simply gives you that 2 x delta x plus 2 z delta z equal to 0. So, delta x and delta z
if they are independent then actually x and z should be 0 and that is not possible because
everything is then 0 identically always ok. So, always this means that the particle is
sitting at origin which has no meaning here ok. So, delta x delta y delta z they are not in general
independent. So, right from this equation I cannot come to this what I have to do I can say.
So, ok since they are not independent in general for any arbitrary value of lambda I cannot claim
this thing this thing and this thing they are 0. But I can now ask this question is there any
possible value of lambda for which even if delta x delta y delta z they are for which even if delta
x delta y delta z they are not independent finally, I can write m r double dot minus f i
plus lambda gradient of phi is equal to 0 this is a question. Actually you can show that yes
there exist a specific value or a specific functional form of lambda ok. So, lambda is
actually a function of x y z for the time being I just write x y z it can be a function of t as
well ok. So, if lambda is a function for which although they are not independent this can be
true this can be true for a specific value specific functional form of lambda and this
lambda is known as undetermined multiplier of Lagrange. If you really want to know a little
bit more about undetermined multipliers of Lagrange which is mathematically a very sound
technique you can go through a good material on internet I do not have any specific suggestion,
but just read it and if you have any question you can always come back to me through forum.
So, now you just admit that ok I admit there exist a specific form of function for lambda
for which this is still 0 even if these things are not arbitrarily chosen and we will now just
with this claim we can say that this equation is a valid equation and this is the Lagrange's equation
of first kind. So, with the assumption that there exist a specific functional form for lambda for
which this thing in the green box ok is true I can talk of Lagrange's equation of first kind ok.
So, before wrapping this lecture up I want to mention somehow two very interesting point about
this Lagrange's equation of first kind. The first thing is that here we have done a very simple case
with only one constraint ok. Now it can have actually more than one constraints
ok. So, for a motion in a in 3D space you can actually have ok more than one constraints. So,
let us say two constraints one can be actually say that x is equal to some a cos omega t and
then y square plus z square is equal to some a square ok if you have such things.
So, first of all you have to make that this constraints are holonomic
ok. Now for every single holonomic constraints actually one undetermined Lagrange's multiplier
comes into play. So, if you have two holonomic constraints ok then actually you have to introduce
two undetermined multipliers for example, lambda 1 and lambda 2 ok. So, because then actually what
you will have you will have two equations like phi 1 equal to 0 and phi 2 equal to 0 and then
actually you will have some equation like delta phi which is equal to 0 then you will have delta
phi 1 which is 0 and delta phi 2 which is also 0 and that is why you will have two different
multipliers of Lagrange ok. I will give you an interesting problem on that in the assignment ok.
The final thing is that although for the sake of this course which is a very introductory level
course I have just introduced the Lagrange's equation of first kind exclusively with the
holonomic constraints it is not necessary that the constraint should be strictly holonomic.
If a constraint has a specific form called a Fafian form ok, but this constraint is not
holonomic that mean it contains actually an equation involving even velocities
ok in addition to position coordinates and time then also for certain form of
non-holonomic constraints you can also use this method of Lagrange's undetermined multiplier
and you can use Lagrange's equation of first kind to solve a problem ok.
This type of problem is now beyond the scope of our current interest of this course ok,
but if you are interested you can always go through very good books for example,
Goldstein books you can see some examples discussed on this matter ok. Thank you very much.
