Lecture 14

IIT Kanpur July 201
20 followers
Aug 2, 2024
[Music]
hello everyone welcome back in the last
few classes we derived the governing
equations uh for fluid flows and we saw
the incompressible flow limit and the
comp flow limit for the compos flows we
learned that we require all three
equations Mass conservation momentum
conservation and energy
conservation and also to close the
equations we needed the ideal gas
equation for the ideal gas flows okay or
any relation between uh the pressure
density and the
temperature in today's class what we are
going to do is we are going to
non-dimensionalize the
equations so why it is important to
non-dimensionalize
equations it is very essential you will
see that in some of the cfd
applications uh the parameters that will
be given to you are basically the
non-nominal numbers for example the r
number Mark number maybe the plal number
okay and so on and so forth and you'll
be asked to basically find the forces
okay maybe the lift and drag
coefficients okay the pressure
coefficient the skin fiction coefficient
drag Etc
okay so in those cases how do you if you
look at the equations all the equations
have your the velocity pressure density
temperature Etc right but the
non-dimensional numbers do not appear in
those equations like Ros number and Mark
number uh Etc right so how do this flow
vary according to this uh
non-dimensional parameters right so it's
very important to cost these equations
in terms of these nonnal numbers
and then uh you will see that how the
variations of the in these particular
numbers uh your equations or your forces
are basically changing right you might
have learned about the laminar flows
turbulent flows transitional flows and
if Theos is low enough it will be
laminar and there will be certain amount
of drag if the flow is laminar
especially the skin fixtion drag will be
lower if you the flow becomes turbulent
the drag becomes higher right so these
findings that we have from the
experiment should also be separated by
the governing equations and they are
indeed supported and you can actually
directly see it also some of you might
have taken courses in bondar theory and
you might have basically derived the
bonday equations and you might have made
some approximations that the variation
in the y direction is much much higher
than the VAR in the X direction and you
might have derived the bondar equations
so by using uh similar kind of uh uh
basically uh analysis uh we can find and
derive the non-d equations this is also
important uh from the point of view that
uh if you look at let's say uh a real
flight uh experiment and uh let's say
what is happening in a wind Tel right so
uh in the flight experiment let's say
you are flying in an A320 and you have
certain amount of drag and certain
amount of left lift and certain amount
of moment forces right moment
coefficients
right but uh before it was designed it
must have been basically experimented or
it must I mean the parameters must have
been evaluated in the Wind Tunnel right
now fitting that kind of aircraft in a
winel is very very difficult right the
size is too big and in order to
basically uh exclude the wall effects
and uh all other things you need to have
a Vel which is like at least four times
the size of the aircraft right and
building that winnal itself is very very
expensive right so how do you basically
uh experiment that kind of situation
right so what you can do you can use a
scale down model right so basically the
same aircraft uh the same uh Wings Sam
impan Sam control surfaces Sam flap
slats and everything that you have in
at20 will be there in your scale down
model right but how do you basically uh
cross Bond uh so you'll get some result
from from the windel but how do you make
a relation with what is happening in
your real flight situation right there
must be some number that is governing
the uh the Dynamics right so that you
can say that okay what is happening in
this model must also be happening in
your real prototype okay so we look into
that so before we go into the
nalization I want you to first talk
about the similarity or the simil okay
so what is similitude basically it is
about uh uh making a relation between
the Prototype and a model OKAY under
certain constraints
okay so there are three kind of
similarities okay as you can see here in
the slide okay one is your geometric
similarity
okay this is geometric
similarity uh the second is the chematic
similarity and third is the dynamic
similarity okay we will look into all
these aspects one by one okay so what
geometric similarity says that the shape
of the uh prototype and the model must
be the same okay not the size the shape
must be the same and the model can be
basically scaled down or maybe scaled up
also it's it's up to you uh by same
factor
everywhere okay so for example if you
see this example here here so this is a
wing here okay
and it has like uh if you can see
there's a cord length is around 8 m okay
so it is scaled down by 0.1 so this is
now become 8 m right what was 8 m here
this is in the Prototype now become
point8 M so this is scaled down by 10th
Vector right so the same should be
followed everywhere let's say if you
have a maximum thickness somewhere which
is 1 M here should become .1 M okay
and uh if it is 40 m in the span it is
becomes 40 by 10 which is 4 m in Span
okay that okay not only that it should
maintain the same angle of attack right
so if it is at let's say 10° angle of
attack here okay it must be the 10°
angle here okay that means it should be
of the same shape okay so let's say uh
if you have an N foil like this okay
which is let's say cord length of uh 1 M
okay at an angle of some degre let say
30° and if you're trying to have the
geometric similarity of let's say 1 by
100 right so then you will have an arrow
file which is small okay which is uh uh
1 cm right and then the angle ofex
should be the same is the
30° okay and then if you have the
geometric similarity okay and then we
can talk about the chematic similarity
and the dynamic similarity but the
geometric similarity is the first
prerequisite or the precondition for the
chematic similarity and the uh Dynamic
similarity
right so now that we have the geometric
similarity that means the model and
prototype have the same shape and the
model has been scaled by the same factor
everywhere in all the dimensions right
we can talk about the ktic similarity so
what does the dtic similarity says is
that at two corresponding points okay
between the model and the Prototype
okay uh the velocities are in the same
direction and must be scaled by the same
factor okay at any two location so let's
say this is on this particular Wing we
are talking about a height okay let's
say we are talking about a height uh
let's say somewhere here okay
okay this is uh in the if I can say this
is in the J Direction let's say okay
this is the y direction okay let's let's
fix our coordinate system so that it
will be easy to
explain okay so let's say this is my X
in the quad wise
Direction y in the
uh the vertical Direction okay and z in
the span direction right so at some
height in across the bond layer let's
say maybe the the 05 mm
okay you're measuring some Force let's
say you're measuring the velocity right
so some point let's say your velocity
uh is in this direction okay somewhere
okay so if you are doing it same on the
Prototype okay so because it is SC by
0.1 in the Prototype you have to go to
the height okay which will
be let's
05 okay because it was 05 here now it is
0.5 M okay because we have to the
velocities at the corresponding point so
whatever distance you have here you have
to take the one10 of the distance here
in this direction and then you can take
talk about the forces
here okay
and the chematic similarity says that
the velocity should be in the same
direction okay and uh it must be scaled
by the same factor so here you have
taken the factor is for the geometric
similarity you have taken the factor of
1x 10 so the factor is 1x 10 here
between the uh model and the Prototype
right so model is prototype by 10 right
but need that need not be the factor for
the velocity also it can be 1/3 or one
one 1/2 so for this is for your
geometric similarity for the dynamic
similarity your velocity may be let's
say if I say VM by VP it may be 1/2
right but that will be followed
everywhere right so at every
corresponding point in the domain your
uh velocities will be scaled by the same
factor okay so that is your chematic
similarity and last one is your
dynamical similarity okay so before you
go to the dynamical similarity that is
one thing I want to mention which is not
often emphasized is that for the
chematic similarity it is very important
that the flow should be in the same
regime
okay so whether the flow is in lamin
regime or the transitional regime or the
turbulent regime that matters a lot okay
so you should not be comparing let's say
you have a critical anos number so the
critical number is Theos number at which
the flow basically Transitions okay from
a laminar flow to a transitional flow
and then eventually to the turent flow
right so let's say you are talking about
a number which is in the transitional
regime then it will be very difficult to
basically uh correspond a prototype to a
model uh unless the Prototype is also in
the same regime right so if we talking
about lamina the both should be in the
lamina regime if we're talking about
turbulent both should in the turbulant
regime right and transitional flow
particularly uh where the uh the
transition happens very very fast you
have to very very careful that whether
you're talking about in the lamina
regime or in the D regime you might have
heard in your fluid mechanic course
about the drag crisis that uh the drag
increases and suddenly the drag drops
down as you the regime changes right so
those things we need to be very very
careful about okay now this is about
your cantic similarity now last one is
the dynamic similarity so the dynamic
similarity says is that at two sponding
points let's talk about the the same
corresponding Point here or we can
basically talk take an example of let's
say two circles or two AR files let's
take the example of two file here okay
that I have drawn here already and we
talking about two forces here okay okay
let's say forces here and here okay so
the dynamic similarity say that uh that
corresponding points the forces on the
body should be in the same Direction and
this must be scaled by the same factor
right and again this Factor can be
different from what factors that you
have used in the geometric similarity or
the ktic similarity but the the the same
factor should be applied everywhere in
the domain okay so that is what is the
dynamic
similarity so the forces should be
identical okay and that they are
parallel in the sense that in they're in
the same direction and scaled by the
same factor everywhere
so now for the dynamic similarity okay
you must have the geometric similarity
and the ktic similarity so if do these
two similarity are satisfied okay then
you can talk about the dynamic
similarity okay so how does it help is
that like the example that I we gave
that okay let's say that we cannot uh
this is your your real prototype and
this is the model that you can use in
the winnal test so you have done the
testing on this model and you got some
forces but you know that these forces
are a factor of what is happening
actually uh in your prototype so you can
just multiply that by that factor and
you can find out the actual force that
is experienced by the wing right or you
can talk about the the velocity field
and you if you know the factor you can
find that out
okay and uh uh uh so if you look at many
curves okay in the flu Dynamics
literature okay you will find out for
example there is a moodi chart okay
which says that how with the renos
number okay your drag is changing right
this is in this side you have Renos
number this is your side your drag okay
and then let's say this is how your cve
looks like okay so this is the famous
modood
chart now you might ask question that
Ros number is a function of many
different parameters
okay row u l by mu you must upd in your
fluid dynamics
[Music]
your density velocity length scale and
uh the viscosity okay and here is the CD
is basically d by let's say half r u
sare
okay and maybe the L Square also if you
divide by area okay or some area okay so
you might uh ask the question that uh
let's say in this model I have the
length is small and here the length is
higher then how do we talk about the
correspondence because the
uh the density is different in the model
the the velocity different in the model
uh the viscosity is different in the
model so what the the dynamic similarity
says is that as long as your length
runos number which is the governing
parameter here governing non diens here
which basically dictates the Dynamics
right so as long as you not number is
same between the model and the Prototype
okay regardless of what is the length
scale you're using what is the velocity
scale you're using what is the density
scale you're using you're going to same
experience get the same drag which is
given as the CD here right so that's how
you can uh basically make correspondence
between your simulations your experiment
on your theory right because all the
flows are basically uh governed by some
or other the non-al parameters okay so
let's look at some of these non-al
parameters
okay so the first and the most important
parameter for uh the people who are
interested in cfd is the runs number
right and I'm sure you know uh the not
some by now okay uh through your flu
mechanics courses because if you're
taking this cfd course you must have
taken a basic fluid mechanics course and
in any basic flu mechan course the first
thing that you learn about is the Run
number so what is run number so run
number is given as R is equal to here
it's row v l by mu right so you can
write it as row
v² / mu V by
L
right and then you can multiply this
entire thing by l s which is some kind
of area Okay so this is r v Square L
square and mu V by L so what are these
two things okay let's look at the bottom
one first okay so what is this so if you
know if you have know the bound layer
okay and if you want to find out what is
the drag experience here okay the skin
friction drag or the viscous drag so you
say that you you can find a SE stress
which is to W which is basically mu d u
by d y at yal to0 right
so this is your viscous Force right so
mu multip by some velocity gradient so
if you write in the in terms of
dimensional numbers this is like mu V by
L right so it represents your viscous
stress so mu V is a viscous stress
multip by the area the stress multiply
area is the force so this is the viscous
force and what is this row row v²
inertial Force we have just derived the
momentum equation and if you see the
second term that has the velocity coming
twice okay so that is your inertial term
so this is your inertial force that
means how much is the uh the
contribution of the uh of the Velocity
that is coming in divided by how much by
how much Factor the viscosity is
opposing that right so that the ratio
between the two Force this is your
inertial
Force divided by the viscous Force okay
and the ratio of the two is called the
runos number right so if the runos
number is very very low that means your
viscous force is dominating right uh
basically uh the inertial forces are not
that strong to basically find out the uh
basically uh to have the flow overcoming
that viscous nature right and on the
other hand if you go to the higher l
number so the viscous forces will be
small small so the bond layer that is
forming on the body will be very very
thin okay but at the same time you will
have the the flow gradient changing the
very very thin
region okay so this is a very important
number okay given by rods okay this also
tells whether the flow will be in the
lamina region or in the dant region
right similarly we have for the
compressible flow we have seen uh the
importance of Mark number the mark
number tells whether the flow is going
to be incompressible or compressible
right in terms of number you can write
uh Mark number is ratio of inertial
Forces versus the elastic forces and
elastic forces can be related to the
compressibility okay we have earlier
derived
okay how the speed of sound can be
related to the compressibility right
basically entopic compressibility so you
can basically derive that and you'll see
that Mark related to the initial forces
to the elastic
forces so these two are the these two
numbers we'll be talking again and again
in our course okay but then there are
some other situations where other
numbers may become important okay for
example is a fruit number okay which is
given by v² by GL or you can write it as
row
v² by G and you multiply by
area Okay so that is basically this is
the in cell force and when the gravity
is important this number becomes very
very important and so the bottom is the
Gravity Force right so all these numbers
tell you that uh it's an basically
interaction between two forces and which
force is dominating uh at a given
situation right so let's say if the
fluid number is low uh then your gravity
force is dominating over the initial
Force if the fru number is high the
initial force is dominating over the
gravity force right so depending on the
situation or what kind of fluid flow
that you're are trying to simulate you
have to look for the forces which are
dominant in the uh in the flow field and
then you basically balance them out okay
similarly you have this Ila number which
is basically the pressure forces to the
inal forces right so you know that there
are pressure terms also and the initial
term also in your new Stoke equations
and you take the ratio between the two
okay this is the press is Delta P so
there's some pressure gradient there so
your IL number will be Delta P by row v²
and pressure into area that gives you
the pressure force and row v² into area
okay that gives you the initial
force and uh you have to see that how
they balance out each other right so if
High number is higher then pressure
force is dominating and vice versa and
simp similarly if you have surface T
present then you talk about the W number
okay and here the it's a ratio between
the initial forces divided by the
surface tension forces right so
depending on the situation these numbers
are very useful particularly for web
number if you are in chemical
engineering and trying to do simulations
and if you have a free surface okay in
then those cases you worry about this we
number okay similarly if you have a flow
which is dominated by the gravitational
force right then you talk about the
fruit number
okay uh let's now talk about a case uh
where we uh talk about the dynamic
similarity okay so we'll give an example
so that the concept of dynamic
similarity gets clear in your mind okay
so the problem is the consider the flow
of two geometrically similar error files
okay the geometrically similar means
that the geometric similarity is Main
that means uh that two AR files are
basically scaled by some Factor here it
is given one having four times the cord
of the other file okay so the model is
basically 1/4 time of the the Prototype
okay so remember we are talking about
the scaling down in the model because
that is what is generally practiced you
want to have a smaller uh section in
your V Tel then what is there in the
flight but it is not necessary you can
have a scale up also so as long as you
can maintain the Ros number and the mark
number for example then uh it doesn't
matter whether you scale down or scale
up
okay and the free stream parameters are
very different such that your V2 = to V1
row2 = to Row 1 by 4 and t2al to T1 4 T1
so are those flows dynamically similar
so let's Let Me 2 V1 so uh so on here
that uh the one having the four times
the Cod of the other error file right
so let's draw draw the AR files so this
is your small one let's say this is your
biger file so this is the cord of
C this is your cord of 4
C right this is one and this is two
right and then what it says V2 = to 4
V1 so this V2 = to 4
V1 and then root 2 = to Row 1 by
4 and T2 = 4
T1 okay so basically if the temperature
is involved it's a compressible
flow so let's take the ratio of re1
versus re2 so what we need to match is
that re1 should be equal to
re2 and M1 should be equal to M2 right
for the dynamical similarity so let's
test take this example R1 by re2 right
so R1 is row1 V1 C1 by
mu1 and R2 would
be 1 by2 is mu2 by R2 V2
C2
right so express everything in terms of
the uh the one okay this is Row one V1
C1 by
mu1 okay and then mu2 would be let's
write mu2 here only uh row2 = to row 1X
4 in V2 is 4
V1 and C2 is equal to 4 C right or if I
can call it C1 this is 4
C1
right so the row one row one cancels out
B1 B1 cancels out okay this four four
cancels out
okay and uh then you will
have
mu2
by 4
mu1
right so this is the ratio between re1
and re2
right just V2 = 2 V1
right V2 is 2 V1
2 V1 if you do this this is 2 V1 this 44
cancel out so this is mu2 by 2
mu1 right now you know that uh the the
variation of mu uh you know the souland
law so it goes as Mu equal to some mu
reference multip by t^ 3x 2 by t + S
okay this is the s l law that you have
discuss so roughly because this is t^
3x2 and this is T so you can say roughly
new Val 2^ half right sare root of 2
right
so mu2 can be written
as t to the power half okay and some
other factors which is constant for both
thees and T1 to^ half right so this is T
into T2 by T1 to the^ half and T2 by T1
is 4 so 2
into
uh this is 1x
2 1 by 2 T2 half which is 2 2 into 1X 2
right because 22 by 2 1 is 4 right so
this is equal to 1 that means your re2
equal to
re1 right so uh in terms of the
comparison of Ros numbers both are
similar right now let's look at the Mark
number okay so let's look at the Mark
number I'll just do it uh quickly
here so r re2 equal to
re1 now we want M1 by
M2 right so M1 is basically V1 by square
root of gamma T1 K that is sare root of
gamma
T2 by V2 right this is V1 by V2 into
square root of T2 by T1 so B1 by B2 so
B21 by2 so B1 B2 will be 1 by2 and T2 by
T1 is 4 the root of 4 that will become
2x 2 is equal to 1 that means M1 is also
equal to
M2 so that means these flows are
dynamically similar okay I'll just give
you one last example and then we close
the discussion
today okay so the example is flow in a
lead driven cavity right so we have been
discussing about this case of lead
driven cavity
okay so this is domain okay and there's
a gravity which is moving at certain
velocity okay let's say U okay and then
you'll have flow which gets developed
here okay let's say that you want want
to simulate this flow using uh a
dimensional solver and a non-dimensional
solver so what is the dimensional solver
that means you have to assign the value
of pressure uh velocity uh length
temperature all those things okay in the
non-dimensional solver what you need to
assign is your runos number and Mark
number okay and maybe the temperature
right some
scaling okay so you have Ros number
which is it's a very famous example okay
based on this length scale this is your
L okay and your your top moving velocity
so r l = to
row u l by mu is equal to
400 right and Mark is given as U by a =
to 0.1 okay and the temperature is given
if if it is not given you can assume it
okay this is 288 kin
right so now if you want to solve this
problem using a dimensional Sol for
non-dimension solver you have all the
numbers given okay non Dimension solver
what you need to do is that just set up
the problem in the bonding conditions
give the no SL bonding conditions
everywhere okay for the top lead you
just give the velocity which will be in
the non it will be equal to one because
you are non-dimensionalizing with
respect to the the velocity that is
moving okay and they just set up the Ros
and Mark number but let's say you want
to solve it using the dimensional solver
then you need to find out out your
velocity your values of p v l t and also
row right so how do you do that okay is
very easy to do so okay so you can take
any side length L equal to one or you
can take1 .01 it's your thing so
whatever thing that you want to take so
let's say I'm taking Lal to 1 okay now
the first thing that you know is the
temperature right so from the
temperature what are the things that you
going to calculate your mu directly
depends on the temperature the viscosity
and your uh speed of sound depends
directly on the temperature right so use
the otherlands law
okay and you know the values of the C1
and S which are constant you can
calculate the value of mu okay which
will come out to be
1.79 or so into 10us 5 Pascal second
similarly you can calculate AAL to
square root of GMA RT and you'll get the
value of square root of 340 m/s okay
which makes sense because the
temperature is around 288 Kelvin and we
doing it for air so gamma is equal to
1.4 okay T is 1 2 288 K and R will be
around for perfect gas 287 something
okay JW per kg
Kelvin okay so that will give you the
speed of sound as 340 m/s right so
because of Mark number is 0.1 your
velocity will become Mark into a which
is
34.2 okay now now that you know your
speed of sound velocity and the
viscosity and your length scale also you
know you fix it to one you can calculate
your density how because r
l r l is equal
to row u l by mu right so if you want to
calculate row it will be r l into mu
by U into L right
so if you put all those values there
you'll get the value 2105 10us 4 right
so you might be finding it very
surprised that why the density value is
very very low it does it is not
realistic right so it does not matter in
cfd okay right you can set up the value
and as long as the value is reasonable
okay that means it is not going into
some kind of floating Point error or uh
some error which is like below the
Precision of the the computer it will
run right but for the experiment you
have to be very very careful you have to
choose a dimension l so that your
density is matching to the density of
the air or density of the water visable
Med Medium you're selecting for the
simulation right so you have to fix fix
some other things
there and uh after you got the density
you can find the pressure also P equal
to row RT okay which is basically 17.45
99 okay so you can use these values to
basically do the simulation in a using a
dimensional solver you know the value of
row u t p mu and everything right so
this is how you can convert from the
dimensional to the non-dimensional and
vice versa right so in cfd it is pretty
easy because you are not really bothered
about uh uh what is the actual value of
density or actual value of pressure and
how it matches to the realc situation
but it is very difficult in an
experiment to basically get the dynamic
similarity one to one between uh what is
happening in the realistic condition and
what is happening in your internal
condition okay for example you want to
match the both the and Mark number and
often it is uh uh not uh very easy to do
so okay so in that case what we could we
do is we talk about the partial
similarity in the sense that let's say
even uh you're not able to do it a very
low Mark number okay let's say you not
able to do it mark equal to point one
right but if you can do it to let's say
mark to point2 but by maintaining a
similar Ros number you can assume that
flow to be nearly similar right so you
can have this kind of partial similarity
uh in your realistic uh experiments in
the uh in V Tel but in cfd you can
basically find out uh the values of the
Primitive variables and do the
simulation and remember this is not the
only way to do that you can basically
change your lengthy scale you can change
your viscosity okay uh you can fix your
different temperature uh and you can get
different values of primitive variables
okay so with that uh we will conclude
The Talk today uh we'll meet in the next
lecture we will where we will nalize the
governing equations thank you very much
[Music]
[Music]
[Music]
