HOME

HW 14 Mathematics 503, Mathematical Modeling, CSUF , August 6, 2007

Nasser M. Abbasi

June 15, 2014

Contents

1 Problem 9 page 346 section 6.2 (PDE's)
2 Problem 3 page 365 section 6.2 (PDE's)
3 Problem 5 page 365 section 6.2 Conservation laws

1 Problem 9 page 346 section 6.2 (PDE's)

problem:

Find all solutions to the heat equation ut = κuxx  of the form u(x,t) = U (z)  where     x√--
z=   κt

answer:

We have that z(x,t) = √xκt  , hence ∂∂zx = √1κt  and  2
∂∂zx2-= 0  and             3       − 3
∂z∂t = −2x(κt)−2 κ = −2xt√2k-

Now

pict

and

pict

and

pict

Plug in the above expressions into the PDE we obtain

pict

But     √x-
z =  κt  , hence the above becomes

  1   ′      ′′
− 2z U (z) = U (z)

or

|---------------------|
|   ′′                 |
| U  (z)+ 12z U ′(z)= 0 |
-----------------------

Let U ′(z)=  y(z), hence the above becomes

pict

Integrate both sides

       1
lny = − -z2+ C
       4

Hence

        −1z2
y(z) = Ae 4

But since  ′
U (z)= y(z)  , then

pict

I think now I need to write the above in terms of x,t again. Fix time, and change x and so we have      ∂z     1--
dz = ∂xdx=  √κtdx and the above integral becomes

          ∫
                  −(x−4κξt)2√-1-
u (x,t;ξ)=   A (ξ)e       κtd ξ + B (ξ )

for any ξ location along the space dimension x , where A(ξ),B (ξ )  are functions that depend on the value ξ

2 Problem 3 page 365 section 6.2 (PDE's)

problem:

Use the energy method to prove the uniqueness for the problem

pict

Solution

First note that        2    2        2
∇2u ≡ ∂∂ux2-+ ∂∂ux2+ ⋅⋅⋅+ ∂∂x2nu
       1    2   i.e. the Laplacian.

Proof by contradiction. Assume there is no unique solution. Let u1(x,t)  and u2(x,t)  be 2 different solutions to the above PDE. Let w (x,t)  be the difference between these 2 solutions. i.e. w (x,t)= u1 (x,t)− u2(x,t)  , hence w (x,t)  must satisfy the following conditions: it must be zero at the boundaries x∈ ∂ Ω  for all time, and also it must be zero inside Ω  initially. Hence

pict

Now if we can show that w(x,t)= 0  for t > 0  inside Ω  , then this would imply that u1 (x,t)= u2(x,t)  , showing a contradiction, hence completing the proof.

i.e. we need to show that wt(x,t) = ∇2w (x,t)  yields a solution w (x,t)= 0  for x ∈ Ω,t > 0

Using the energy argument, we write

       ∫
E (t)=   w2 (x,t)dx
        Ω

First we note that E (0)= 0  since w (x,0)= 0  from the initial conditions above.

pict

But ∂∂tw (x,t)=  ∇2w(x,t)  from the PDE itself, hence the above becomes

         ∫
E ′(t)= 2   w (x,t) ∇2w(x,t) d x
          Ω
(1)

But from Green first identity which states the following

∫  (             )     ∫
   u ∇2w+ ∇u ⋅∇w  dx =    udw-dA
 Ω                      ∂Ω  dn

Replace u by w in the above, we obtain

pict

Comparing (1) and (2) we see that LHS of (2) is 1E′(t)
2  Hence the above become

        ∫            ∫
1E ′(t)=     wdw-dA −   ∇w ⋅∇w  dx
2        ∂Ω  dn       Ω

But ∇w  ⋅∇w = ∥∇w ∥2   , so the above becomes

        ∫           ∫
1E′(t)=    w dw-dA−    ∥∇w ∥2 dx
2        ∂Ω  dn      Ω

But w(x,t)= 0  on ∂Ω  for t > 0  , since this is the boundary conditions. Hence the above becomes

          ∫
E′(t) = − 2  ∥∇w ∥2 dx
           Ω

Therefore we showed that E′(t)  is ≤ 0  since ∫
 Ω∥∇w ∥2 dx≥  0

So energy inside Ω  is nonincreasing with time. But since E (0 )= 0  then E (t) = 0  (since energy can not be negative, this is the only choice left).

Therefore, from       ∫   2
E(t)=  Ω w (x,t)dx  , we conclude that w(x,t)= 0  everywhere in Ω  for t > 0  since w (x,t)  is continuous in both its arguments.

Hence we conclude since w (x,t) = u1(x,t)− u2(x,t) = 0  then u1(x,t) = u2(x,t)  , then the PDE solution is unique.

3 Problem 5 page 365 section 6.2 Conservation laws

problem:

In absence of sources derive the diffusion equation for radial motion in the plane     D
ut = r (rur)r  from first principles. That is, take an arbitrary domain between circles r = a,r = b and apply conservation law for the density u = u(r,t)  assuming the flux is J(r,t)= − Dur  . Assume no sources.

Answer:

PIC

First note that the density u (r,t)  is measured in quantity per unit volume.

Consider a cross sectional area through circle ra = a . This area is 2πhra  where h is the width of the strip.

Let J(r,t)  be the flux at r at time t , measured in quantity per unit area per unit time.

Hence amount u that passes though cross sectional area at ra  , per unit time, is A (ra)J (ra,t)  where A (ra)=  2πhra

Similarly, amount u that passes though cross sectional area at rb  , per unit time, is A (rb)J (ra,t)  where A (rb)=  2πhrb

Hence the net amount that flows, per unit time, between rb  and ra  is A(ra)J(ra,t)− A(rb)J(rb,t)

Since there is no source nor sink inside this region, then the above equal the rate at which the amount u itself changes between rb  and ra  , which is d-
dt (u (r,t)× volume between ra and rb)  .

Hence we have

pict

Apply fundamental theorem of calculus on the RHS above where                ∫a
J(a,t)− J (b,t)=  b Jrdr hence the above becomes

∫ b               ∫ a
   ut(r,t)A (r)dr =    ∂--[A (r)J (r,t)]dr
  a                b ∂r

But A (r)= 2πrh so the above becomes

∫             ∫
  bu (r,t)rdr =   a∂--[rJ(r,t)]dr
 a  t          b ∂r

Changing the limits on the integral in the RHS above to make it match the LHS, we obtain

∫               ∫
  bu (r,t)rdr = −   b∂--[rJ(r,t)]dr
 a  t            a ∂r

Because the above holds for all intervals of integration and the functions involved are continuous, then we can remove the integrals and just write

           -∂-
ut(r,t)r = − ∂r [rJ(r,t)]

Now assuming diffusion model for the flux, i.e. J (r,t)= − Dur(r,t)  , then the above becomes

           ∂--
ut(r,t)r = D ∂r [rur(r,t)]

Hence

|-------------|
|      D      |
| ut = r [rur]r|
---------------