Lecture 09

IIT Kanpur July 201
20 followers
3 Video Views·Dec 1, 2023
good morning everybody
today we will discuss about
2d heat conduction
in a
finite cylinder
here we can see
a finite cylinder and
we will consider
the following boundary conditions
at
r equal to r naught
temperature is t one
throughout
and
then at z equal to 0
again
ah
throughout
the diameter or
you know
in the r direction
it is
uh at z z equal to 0 t 1
at
z equal to l
it is
temperature is t2
it may be a
finite temperature
ah constant all over
or it
is basically
excess temperature over t1
which is a function of radius
so basically
this side temperature
is higher than
t1
and at r equal to 0
that means along the axis at r equal to
0 theta is finite
so
t is finite so temperature is finite
ah and the axis are equal to 0
at r equal to r naught
temperature is t1
at z equal to 0 temperature is
t1 and at z equal to l temperature is t1
plus fr
now governing equation
del 2 t del r2 plus 1 by r del t del r
plus 1 by r square del 2 t
del phi 2 plus del 2 t del z 2 equal to
1 by alpha del temperature del time and
here
we have a rotational symmetry
along phi
so
we can neglect delta t del phi 2
and therefore
uh
under steady state first of all
uh before neglecting that under steady
state
del temperature del time
1 by alpha del temperature del time this
term is dropped under steady state
and
of course
we have said
already mentioned about it
phi symmetry
so this term can be dropped then
delta t del phi 2
and finally we put
theta equal to t minus t 1
then we can write
del 2 theta delta 2
plus 1 by r
del theta del r
plus
del 2 t
del z 2 of theta
equal to 0
this is our governing equation
and then boundary conditions will
express in terms of theta
at r equal to r naught
theta equal to 0
at r equal to
0
along the axis theta is finite
at z equal to 0 theta equal to 0
domain is
r
greater than 0 less than r 0
and at z equal to l
theta is f r
which is x s over t one temperature
and
again
domain is
r greater than equal to zero
less than equal to r naught
so
these are the
uh boundary conditions and this is the
governing equation
now
we have to solve it
we now apply
the separation of variables technique
by assuming that
the desired solution
can be expressed
as the
product of two functions
one of which
depends
only on r
and the other depends only on
z
this is
basically
standard technique
of separation of variables
we assume
the solution final solution
is product of two function
one is
function of r only and other is function
of
z only
so theta
r z equal to
r
into z
now if we
find out del 2 theta delta 2
this will be
upper case r
double prime into
z
del theta del r it is uppercase r
prime into z
and delta theta del z 2
is
uppercase r into uppercase
z
double prime
so
our governing equation now becomes
1 by
uppercase r
multiplied by
r
double prime plus 1 by r r prime
all are function of r only
plus
z double prime
divided by z
all are function of z
only this is equal to 0
then we separate out
r dependent part
minus 1 by
uppercase r into uppercase r double
prime plus
1 by r in 2 upper case r prime
equal to
z double prime by z
equal to lambda square
which is separation constant
now
from the z dependent part we get
z double prime minus lambda square z
equal to 0
and obviously the solution is z equal to
a
into e to the power
minus
lambda z
plus b into e to the power
plus lambda
z
and
our dependent part
we can write as
r double prime
plus 1 by r
r prime
plus
lambda square
into r equal to zero
just to equate
our dependent part straight away with
lambda square
we get this
now
if we look at the
equation carefully
we can see
a this form matches with
delta y del x 2
plus 1 by x
d y d x
plus y
equal to 0 d 2 y d x 2 plus 1 by x d y d
x plus y equal to 0
which is
bessel's equation
of
order 0 this is bessel's equation
of order 0
and
solutions are
bessel function
of the
first kind of order 0
and
bessel function of the second kind of
order 0
and we can write r equal to c
j naught lambda r
plus d
y naught
lambda r c and d are constants
so we have mistakenly written here
basal's function this is basil's
equation
so bessel's equation of
order 0
and its solution is bessel's function
and vessel function of
first kind of order 0 and bessel
function of second kind of what is 0 so
final solution is r
c
bessel function of
first kind of order 0
plus d
bessel function of second kind of order
0 y naught
lambda r
so this is the solution
now
z dependent part solution is given by 1
and our dependent part solution is given
by
2
if we combine them
we get the complete solution
but
just before that
you know to remind all of you
uh
we will just
recapitulate
the
uh you know
nuances of
basis function and bessel's equation
bessel's equation of order n
is given by x squared
d 2 y d x 2 plus x d y d x plus x square
minus n square y equal to 0
this is of order n
and
the
its solutions are
basis function of first kind and
business function of second kind
and this is bessel's function of first
kind of order in
and this is basis function of
second kind of order in
so we have discussed this already
while discussing
about
fins
but
let us again recapitulate
so this is basis function
of first kind of order n
given by x by 2 to the power n
and this is this is a series
uh summation m equal to 0 to infinity
minus 1 power
m divided by
factorial m into gamma this is gamma
function
m plus n plus 1
into
x by 2 to the power 2 m
similarly basically we have written
basis function
of
second kind
of order n
now
in this equation if we
put
n equal to 0
this will be
basis function
vessels equation
of
ordered zero
then solutions of course
everywhere we have to substitute n by 0
then we will get
solutions
as
summation of
basis function
of first kind of order 0 plus a constant
into basis function of second kind of
order 0
y equal to a plus
j naught ax plus
b
y naught x and
with the argument
basis function vary this way
this is a variation of j naught
and this is the variation of y naught
with x when x is the argument
so
now keeping this in mind we will again
come back
to our solution
solutions are
we go back by two slides solutions are
one is
j dependent part another is our
dependent part
if we combine these two we'll get the
complete solution
and complete solution is equation 3
which is
theta equal to c
j naught
lambda r
plus d y naught lambda r
into a e to the power minus lambda z
plus b e to the power plus lambda z
now
we will start applying boundary
conditions first boundary condition we
will apply
at r equal to 0
we know theta is finite
now in this
bezel function
of
first kind of order zero and vessels
function of first sky second kind of
order zero
this solution
we know
that when the argument
is zero
j naught is 1
but when the argument is 0 y naught is
infinity
so in order to accommodate
this condition
that
y naught 0 equal to infinity we have to
make
d equal to 0
so
we evaluate d
having evaluated d we can write
theta equal to c so we can
here d is 0 we can straight away right
from here this part into
this part multiplied by this part
so
c c
j naught lambda r into a e to the power
minus lambda j plus b e to the power
plus lambda z
next boundary condition is z equal to 0
theta equal to 0
so we substitute z equal to 0 we get
c
j naught lambda r a plus b
so
equal to zero
c j naught lambda r
multiplied by a plus b equal to zero now
this part our dependent part cannot be
zero then we totally get rid of
ah if c becomes 0 totally get rid of r
dependence so this part a plus b has to
be 0
that means a equal to minus b
if we substitute a equal to minus b in
this equation
then we can write theta equal to 2 c b
j naught lambda r
into e to the power minus minus e to the
power minus lambda z
plus e to the power
lambda z
by 2
which is nothing but
sine hyperbolic lambda
z
so
where this is this becomes sine
hyperbolic
lambda z
and
2 c into b
combine them together we define a new
constant c
into j naught lambda r
sine hyperbolic lambda z
this becomes now
theta
equation 5.
now at r equal to r naught
that means at the periphery
temperature is t one
so theta is zero
so r equal to r naught theta equal to
zero
this gives
if we want to
substitute here
clearly this gives
us
j naught lambda r equal to zero
so c cannot be 0 this part cannot be 0
so that means
that means we will get j naught lambda r
equal to lambda r naught equal to 0
and
this
i mean will be zero as we know
that
the bessel function variation
again
you know we go back by one or two slides
you know when bessel function becomes
zero for those values
uh you know when it is crossing zero
we can clearly see
for those arguments those are the roots
of bessel functions
first order of
first kind of order 0
so
j naught
lambda n are not equal to 0 for specific
values of n
lambda assume assumes discrete values
for which
j naught bessel function
of first kind of order 0 we have seen
its variation
its value becomes
zero for some specific
arguments and that is if it is
represented by
lambda n or not
then
we can say
that those ends are
one two three you know
uh will give
basically the value of bessel function
equal to 0
for specific values of n
now
theta becomes
summation over n when n varies from 1 to
3 up to infinity
c n
j naught lambda n r
sine hyperbolic lambda in z
becomes theta so basically theta
value now
we get for
specific
values of n
lambda n
so j naught lambda in r and sine
hyperbolic lambda in z so
we can get theta equal to summation over
n equal to 1 to 3 up to infinity
c n
j naught lambda n r
sine hyperbolic lambda n z
now only one boundary condition
we are left with that is at the top
surface
z equal to l
theta equal to f r this is
temperature over
t1
we said t2 may be t1 plus fr or t1 plus
some value which will give t2 that is
the temperature on the top surface
so
here
then
at if we substitute
z equal to l
we will get
r
so f r is
then
summation over
c n
j naught lambda n r sine hyperbolic
lambda n
l
and
this sine hyperbolic lambda in l
this value is
you know
can be combined with c n
and we can define a new constant a n
into j naught
lambda in r
summation over this
is f r
and
its
domain is basically r
varies between 0 and r naught
so
here within this domain
we can define basically
orthogonality and
then
we can say
0 to
this is a requirement to express now
this
based on
our
orthogonal functions that if our
jna r j naught
lambda r
lambda n r
d r
equal to 0 to r naught
in
then r j n r square lambda n r d r so
both the sides we are multiplying
basically by
r
j naught
lambda n r
and
integrating between 0 to r
dr
then
a r a n
series n
can be expressed as
integral
0 to r naught
for
r
j naught lambda n r d r
this quantity
divided by
integral
0 to r naught
are
j not square
lambda n r
d r
so if we can evaluate
this part
we are true we will get in
so
basically
in order to integrate this little more
smoothly we are saying for the time
being that afr let it be a constant
value t naught that means t 2 the
temperature at the top surface
basically
you know
in excess of t 1
it is
t0 in excess of
t1 it is higher than t1 by a quantity
which is t naught
and then we can write a n is
integral 0 to r naught t naught
r
j naught lambda inner d r
divided by
0 integral 0 to r naught r
j naught square
lambda in r d r
so just the previous integral
we have substituted for by
t naught
and
now in order to evaluate numerator and
the denominator
separately
we have to use
two recurrence formulae of vessels
functions
now relation one
is this is very well known the fast
recurrence relationship it is called
ddx of x power n
j n x
is x power n
j n minus 1 x
now if we substitute n equal to 1
we can
get
d dx of x
j 1 x
equal to x
j naught x
just substitute
n by 1
in this
expression
and from here
we can write
integral x
j naught x dx
equal to x
j 1 x
just from here we can write
and
we will
make use of
this
relationship
in order to evaluate the numerator
we will see little later we'll do that
so recurrence relationship 1
we
will make use of it
and recurrence relationship 2
is
integration between
0 and a when a is the upper limit
x
jn squared
lambda x dx is a square by 2
into
j n prime lambda a square
plus
1 minus n square by lambda square into a
square
n square divided by lambda square into a
square
into jn square argument is lambda
in this case
in our case
we have to evaluate
this where upper limit is r naught
lower limit is 0
and
integrant is
this
so
0 to r naught
are
j naught square lambda n r d r
basically
zero to r naught are z naught square
lambda n r
d r
that is what
will match with this second recurrence
relationship
here we can just follow and
write are not
squared this should be please again
excuse me for this mistake please read
this r naught square we have also
written it as r naught square in the
next expression so this is r naught
square by 2
into
minus
j 1
lambda n
r naught
whole square
plus here
we can see this term will disappear
for n equal to 0
and this is j naught
lambda n
into r naught upper limit a lambda a
this is upper limit or not lambda n r
naught
and
in our case
j naught lambda n r naught
this was our starting point equal to 0
so this is eventually 0
and then
we can write this equation
will be left with r naught square by 2
into j 1
lambda n are not lambda and r naught is
the argument
and this is square so this becomes j 1
square
lambda in r naught into as i said this
will be r naught square please read this
as r naught square so r naught square by
2 exactly r naught square by 2
j 1 square lambda n are not this has
become the denominator
and the
numerator
we
have seen this is the
term
here
if we substitute lambda in into r
as
psi
we can
write
uh in
psi equal to lambda n into r
so n equal to 0 to r
this will be t naught
and
lambda n into r is i so r is psi by
lambda n
psi by lambda n
z naught
psi
dzi
divided by lambda n
so t naught divided by lambda n square
will come out
0 to
r naught
psi
j 0
psi
dzi
so this
exactly we can see matches with this
recurrence relationship
x j naught x dx is x j 1 x
what we get that
get as our
numerator is
psi
j naught psi dz
so
this will become straight away
psi
j naught
j
ah this is basically
x g 1 x so
this will become psi
j 1 psi
so t naught by lambda square psi
j 1
j
upper limit and we put the limit after
integration is done we have integrated
after integration is done we put the
limit
upper limit and lower limit
upper limit and
lower limit and
the
denominator
is r square by 2 r naught square by 2 g
m1 squared
lambda in
r naught
and then you know it's matter of just
little bit of algebra
twice t naught
lambda n square
r naught square
and here
psi means lambda and r naught
g 1 psi
g 1 lambda n r naught
r actually an upper limit or not or not
so divided by j 1
lambda n r naught
so a n becomes
2 times
t naught
divided by lambda n
are not r naught are not cancels out
lambda n one lambda n cancels out one
are not it can uh cancels out
divided by lambda n r naught
into 1 by
j 1
lambda n r naught
so this is n
and
n was
c n if we recall
into sine hyperbolic lambda
in l
so that means c n will be
this
series uh
summation over 1 2
1 2 3 up to infinity
twice t naught divided by lambda in r
naught
1 by j 1 lambda n r naught into
so
a 1 into this quantity was c n
was right hand side c n into sorry c n
into sine hyperbolic lambda n l is
basically n
so c n will be this quantity divided by
1 by
sine hyperbolic lambda n l
now
since c n has been evaluated
we can now that was the only remaining
part
if we
look into
our
equation
see
this c n was to be evaluated so
now
we can write
theta equal to
twice
since c n is given by this
so
twice t naught by r naught
summation
sine hyperbolic
lambda in
z
divided by
lambda n
sine hyperbolic lambda nz
into
j naught
lambda n r
divided by j 1
lambda n
r naught
this is the
complete solution so this is the final
expression for the temperature
distribution at any r z location
within the domain
and we may recall
that we said
while
imposing the problem while posing the
problem we
mentioned that t2 is t1 plus a4
so t2 minus t1 is fr
and here we have taken
f r as a constant
value
t naught
we can take that as also a function of r
so whatever it is we can say
that part is t2 minus t1
so
is basically
the
excess temperature over t1 is t naught
which makes t2
so
theta rj at any rj is given by t2 minus
t1 divided by r naught
into
a series
this series is
extended
from
for
values of n
1 to 3 up to infinity
this is sine hyperbolic
lambda n z
divided by lambda n sine hyperbolic
lambda in l
into
j naught
lambda in r
divided by j 1
lambda n r naught
so if we substitute r and
z coordinate correctly
and
maybe we take first three n values or
first two n values
we will get
correct distribution of theta
on the
domain of interest
so
that is what was
our task
basically two dimensional heat
conduction
in a finite cylinder
thank you very much
thank you for your interest
thank you
