noc19-ph12-lec41

IIT Kanpur July 201
20 followers
1 Video View·Dec 2, 2023
So, I start the new topic, energy turbulence, but this on the framework of flow with a vector.
Now, magnetic field is a like a vector embedded in velocity, in the flow, in the flow. So,
the two vectors basically u and w, but velocity field I formulate as if it is advecting the
magnetic field. And now, we will use magnetic field as a b variable, not w, that is standard
notation in literature. So, the topics will be formalism, I will set the equations.
The energy transfer in fluxes, then I will just describe few turbulence models. There
are many models, but I will describe what I consider important ones and then dynamo.
So, we will quickly go through these topics, each of it is like one full course, but we
will try to do in one hour and we will see. So, formalism, so let us first set equation
for u. So, we already have set equation for u, we need to only worry about f u. So, that
is general formalism. In fact, I do like the way it has been written and it is very useful
to think in a global way. So, let us rewrite that equation for u, plus f u, vector u plus
nu grad square u. So, what is f u? So, I have assumed constant density. So, we will assume
incompressible, which is good approximation for many fluxes, not so good for some astrophysical
applications, but it is quite good for sun or this, they work quite well. Sun, sun is
nearly incompressible roughly, quite good, I will not say best, but it is quite good.
For galaxy also, but I will not discuss it, privately we can discuss. So, there are ways
to say that incompressible is reasonably ok, may be I will make a remark. So, I think let
us skip this, because this will go into various approximations on. So, f u, just think of
liquid metal. So, inside the earth there is a lot of metal and there is a magnetic field.
So, and it is like molten metal, it is like water. So, it is incompressible. So, that
is what you should think in your head. So, what is the force density? So, this is
the force density, all of this is rho equal to, this is force density, force per unit
volume. So, what is force density formula for in a electromagnetic fluid or it is called
magneto fluid, you will find this word, fluid which responds to magnetic field. Water does
not respond very well, but salt water will respond weakly, but metals will respond heavily.
So, you need free electrons. So, f u is j cross B, current density is like, it is a,
it is a low range, low range force. Now, if you use S i or if you use C g s, then there
are different constant in front, normalization factors. So, I will use S i, no, I will use
C g s, I will use C g s, unlike your other courses, but I will use C g s. So, plasma
I think lot of you will use C g s. So, you will use C, 1 by C, C is speed of light.
Now, under M H D approximation, so there is another Maxwell equation, which is d by
d t, no, d B by d t is j minus, so 4 pi by C j minus, no, no, curl of E plus or minus,
I should not, yeah, curl of E, curl of B and these are d E by d t, right, plus, but in
M H D approximation, we drop this term. So, we assume that electric field is much smaller
than B. Now, I will not describe this approximation,
justification and so on, with this what is assumed. So, electric field is much smaller,
is highly conducting fluid. So, conductivity is tending to infinity and for that, this
is a good approximation. So, we will focus on energy transfers, fluxes, models, we will
not discuss too much into these details. So, if you put that in, you will get C c.
So, 4 pi by C, no, C by, I made too many mistakes, C by 4 pi. So, C c cancels. So, you get j
is curl of B, curl B by 4 pi and the density, I am assuming 1 right now, but the density
also comes. So, in fact, force will be, there is a density, should be sitting here, force
density. So, you divide by rho. So, this is formula for F u. So, force on u is depends
on B. So, that means, B must be passive scalar or active scalar, active scalar. So, B affects
u. So, B is a active scalar, an active scalar, you know, active vector, sorry, not that,
B is an active vector. So, our formula, which I derived for passive vector will not work.
So, and moreover, the force is quite non-linear, is B B. So, this is a non-linear force, unlike
buoyancy force, which is linear, buoyancy was theta, you know, or density. So, this
is non-linear force. So, that makes life reasonably complicated for M H T. So, this is what we
have to keep in mind, that our force is curl of B cross B. In fact, we can simplify this
force. Let me write it here. So, we can write curl of B cross B as minus curl of B square
by 2, say 1 by 4 pi rho is outside, plus B dot grad B. So, this is an interesting form.
Why it is interesting? Because this part can be absorbed with pressure. So, we can take
this part. So, I say, minus grad P plus B square by 8 pi rho and the other part is
B dot grad B by 4 pi rho. Interesting, no? So, this is called total pressure. So, it
is a hydro pressure P plus B square by 4 pi rho, which is 8 pi rho, which is coming from
magnetic pressure. So, P is called thermodynamic pressure plus magnetic pressure. So, this
is the total pressure and this is the, so I am missing out some vector signs, but these
are all vectors. So, this is our force, which is non-linear, well this is also non-linear,
but I have, you remember this pressure does not participate in energy transfer. So, this
term will be silent on energy transfer and we make another change of variable. So, I
will make a remark here. In CGS, you make this variable B by 4 pi
rho, this has dimension of velocity. Now, this B CGS, this has dimension of velocity.
So, we make a change of variable B CGS to this. So, this has very convenient, no? Now,
B has dimension of velocity. So, we can write this as, this part is B dot B dot grad B and
4 pi rho has disappeared. So, now let us go to velocity, magnetic field. So, magnetic
field equation, we will again borrow the equation for active vector or for the vector. So, instead
of w, I am just writing as B. So, d B by d t plus u dot grad B. Now, the right hand side
is F w plus eta grad square B. So, I need to derive this from some Maxwell equation.
So, which Maxwell equation I should use? We have one equation is curl of E is minus d
B by d t and in CGS, E and B have same dimension. So, there must be 1 by C here, no? Now, these
are nit picking on dimensions, where I will not, I will, I will basically ignore units,
rather I will ignore this constant, but this must, there must be C. Now, E in a, I mean
in an image approximation is minus v. So, in the moving frame of the fluid, if you go
with the fluid, the electric field inside is 0 for ideal flow, no? Electric field inside
a conductor is 0. So, E plus, so by Maxwell, not by Maxwell, by the relativistic transformation,
electric field inside the fluid is this, under u by C much less than 1. This is the transformation
of E and B under relativity. So, you will find this is my E prime. So, this will assume
to be 0, but if the conductivity is large, then this is j by sigma. Now, if you want
to get this proper unit, no, I do not derive it here. So, there is a j by sigma, but there
is a some appropriate factor. So, I can find E, no? E is minus v cross B
by C plus j by sigma, which is curl of B. In fact, eta comes here, magnetic diffusivity.
So, plug this in, you plug this in for electric field here, you plug this in, then I will
get a term for F w. So, this you define, this is how we derive equation for B. So, I write
down the equation now. So, F w under my new units, my units where B is in the terms of
velocity, F w is B dot grad u. So, this is simple, this is nothing complicated. So, you
have to use the formula curl of A cross B or curl of u cross B and you will get this.
In fact, you can also write in other way, d B by d t is curl of u cross B. Now, these
derivations I am not emphasizing because of lack of time, but these are interesting derivations.
Now, this equation looks very similar to another equation, which you have seen in the course.
So, equation for the vorticity, no? So, B and w are similar. In fact, they are not identical,
but equation is identical, but they are not same because vorticity is curl of u, but B
is an independent vector. That is the difference. Otherwise, equation is exactly the same.
There is also some of the dynamics. So, vorticity is stretched by the velocity field and this
is called stretching of the vorticity, no? Same thing happens for the magnetic field.
Magnetic field is stretched by the velocity field and that is how the magnetic field increases
in dynamo. So, magnetic field is generated by this process of stretching of magnetic
field. So, may be briefly I will describe it in the end. This is the equation for B.
Now, so this is the advection term and this is u is acting on B of course. So, this is
like B dot grad u, but it has very similar form as u dot grad B, no? Except B has come
before and u has gone to later. So, these are similar non-linear terms. So, what is
the conservation law? Now, because of F u and F B, u square is not conserved. F u is
not 0, right? I mean if you put in magnetic field, I cannot set the force to be 0 and
F B is also not 0. So, they make the conservation law quite complicated. So, u square is not
conserved, B square is also not conserved, but together they are conserved. This is called
total energy. So, integral, integral of this quantity over whole space is conserved.
Another quantity is called F c. It is called cross helicity. It is defined as half u dot
B. So, u dot B, right now I am keeping B as a total field. It can have also mean component.
So, B field could be B naught plus B. When I put B naught, u is fluctuating. So, B naught
will not contribute, no? So, that will get cancelled. u dot B naught will be 0. So, this
is same as half u dot B. Another quantity, which is constant is kinetic helicity, a
magnetic helicity, magnetic helicity, which is half a dot B. A is a vector potential.
So, curl of A is, so proof I will not discuss, but these are three conserved quantities for
energy. u square and B square are not conserved individually, totally. This is for 3D. 2D
has certain modification, but we will skip 2D discussion.
Now, Fourier space, how the equation look like? So, it is again, I mean it is straight
forward if you just have to look at your notes and pick up the thread. So, e q d t plus n
u. So, my notation in the paper, in this book is, so I write like n u of k and minus i k
vector p of k plus f u of k is 4. So, it is function of k and minus nu k square u k. Something
is coming here. I am going to write n u and f u soon and for B, n B of k, non-linear term
coming from euro grade B and there is a f B of k minus eta k square B of k. Now, n u
of k you know already, right. So, let us, for completeness I write it here, k dot u
of q, q of p, q is k minus p. What about n B of k? All vectors, no minus, this plus i
k dot u q B of p, f u, it is B dot grad B. B is also divergence free. You can write them
as del j B j B i. So, you know what is del j u i u j. So, it is very similar. So, i k
dot B of q B of p, right. It is straight forward and f B of p, f B of k. So, this comes from
B dot grad u. Remember, I just derived it in the last slide. So, this will be del j
B j u i. So, which is the advector? B j, B j is one with constant thing. So, it is going
to i k dot B q u of, so this is straight forward and these are the terms which will exchange
energy among modes. They transfer energy from one to other. So, you see that very nice way
to interpret these transfers and derive flux formulas.
So, I will make one more discussion is Alfven waves. So, these are the waves. So, in hydrodynamic
flows, there is no really wave. Though some of you have got ODs, which are not like wave
solutions, but there are oscillations, right. Like we found that three more model has oscillations,
but there is no wave really, but MHD has wave. If you turn off the non-linear term,
we get wave solution and what is the wave solution? So, let me just make it linear equation.
So, B gives you one term. So, the pressure part is, let us say k, p of k. So, this p
is total pressure, but this term comes, linear term is B dot grad B. So, B is a total B.
So, I can write as B naught dot grad B naught dot grad B. So, I can write this as, in Fourier
space, I B naught dot k B of k. So, this is the term coming from B naught. The equation
for B also has a term, which is coming from B naught. In the right hand side, remember
you have B naught dot grad U, right. So, these are linear terms because B naught is a constant.
So, the term which comes out is I B naught dot U of k. So, these are two linear equations.
I turn off viscosity and diffusivity. So, we can solve for them. Now, it is there in
the notes. You solve and you get two solutions. So, you apply magnetic field along z direction.
Then if the wave vector is along any direction, wave vector can be along any direction. So,
this is the direction k. Then you get a solution, where U and B are in opposite direction. Can
you see that? U is here and B is there and they will oscillate. Also, the wave will move.
Wave moves with C plus, alpha in speed and the C plus is B naught dot k. So, the alpha
in speed is B naught dot k cap. So, thanks to Naren. So, this is because it has dimension
of velocity only if it is k k cap. So, this angle is zeta. So, B naught B naught cos zeta
is the magnitude of the alpha in speed, which is different for different zeta. So, that
is why it becomes 0 for zeta equal to 90. This derivation I am skipping, lack of time.
It is nice derivation. You can do it yourself, but so C plus moves with B naught k and solution
is U and B are opposite and they are equal. So, this is the beauty of this unit that U
and B have same dimension and such are equal. So, it is equi-partition. So, if you look
at kinetic energy and magnetic energy on the average, they are equal.
So, there is another solution. If B naught is in that direction, same thing, but k is
going rightward. The wave moves in the minus k direction. Wave is moving in C minus in
that direction with the same magnitude B naught dot k, but it goes in the opposite direction.
So, there is a wave, which is going in that direction, the wave going in that direction
for given k and for this U and B are in the same direction. So, if U is here and B will
be same. So, these are called alpha in waves and these are fluid mode. These are wave modes
in MHD and they have very important role to play in the dynamics. Now, this I am very
briefly describing it. Now, we also have variable for z plus and z minus. Any question? Small?
Yeah, that is 0 because of incompressibility. So, small b is perpendicular to k. So, fluctuations
are perpendicular to k always. There is one subtle point. What happens when your wave
is along x axis or k is along x axis? What happens to B naught dot k? It is 0. Cos 90
degree is 0. So, there is no wave on the plane or the equatorial plane. So, theta equal to
90 degree has no alpha in waves and so their non-linear term plays a big role on the wave
on that on the plane. Now, we have two variables z plus and z minus. We define it for convenience
U plus B and U minus B. So, given B, you can write always z plus z minus. Some in difference
and the equations are much nicer in z plus and z minus. So, let me write that equation.
In fact, this is just for, in fact, we do lot of work with that. So, let me write the
equation for z plus z plus dot vector. So, this is a mean magnetic field. It is a minus
I B naught dot k z plus z minus dot z plus is minus grad pressure plus viscous term,
but I am not. So, let me write this. I will not really discuss this term. U plus minus
r nu plus minus eta by 2. Now, equations are nice because now we do not have four non-linear
terms. We have one z minus dot z plus. This is for z plus. What happens to z minus? You
just flip plus to minus. It is symmetric in plus minus and except this sign, it becomes
plus because wave moves in opposite direction. So, these are convenient non-linear terms,
which I will discuss in next set of discussion. So, what will be the non-linear term by the
way in Fourier space n plus? In fact, I write this n z plus to differentiate between helical
plus and helical minus. Too many pluses if you write for everything. So, there is helical
plus and helical minus. So, I write n z plus, which is going to be i k dot z minus of q
and z plus of p. So, this is my equation sum over. So, I think this is the last slide.
Thank you.
