CRANK-NICOLSON SCHEME TO SOLVE HEAT DFFUSION EQUATIONI

CRANK-NICOLSON SCHEME TO SOLVE HEAT DFFUSION EQUATIONI

QUESTION: Heat diffusion equation is u_t= (D(u)u_x)_x.
Modify this program to investigate the following developments:
Allow for the diffusivity D(u) to change discontinuously, with initial data as u(x,0)= (1+x)(1-x)^2. boundary values u(+-1,t)=0.
Treat in detail the case D(u)=1 when x<1/2 and D(u)=1/2 otherwise.
=> This is my normal code:
program crank_nicolson
implicit none
real, allocatable :: x(:),u(:),a(:),b(:),c(:),d(:)
real:: m,dx,dt,tmax
integer:: j,ni,ji

print*, 'enter the total number of time steps'
read*, ni
print*, 'enter the final time'
read*, tmax
dt=tmax/ni !the size of timestep
print*, 'this gives stepsize of time dt=',dt
dx= 0.1 !delta x =0.1
ji= 20
m = dt/(2*dx**2)

allocate (x(0:ji),u(0:ji+1),a(ji),b(ji),c(ji),d(ji))
open(10,file='crank_nicolsan.m')

!initial condition
do j=1,ji
  x(j)= -1+j*dx
  u(j)=(1+x(j))*(1-x(j))**2 !x=-1+j*dx
end do
x(0)= -1
x(JI)= 1

!boundary condition
u(0)=0
u(ji+1)=0

  do j = 1, ji ! a,b,c are the coefficients of c-n scheme and d is right part
    a(j) = -m
    b(j) = 1+2.0*m
    c(j) = -m
    d(j) = m*u(j+1)+(1-2.0*m)*u(j)+m*u(j-1)
  end do
  call thomas(a,b,c,d)
  do j=1,ji-1
    u(j)=d(j)
  end do

!Print out the Approximate solution in matlab file
write(10,*)  'ApproximateSolution =[',x(0),u(0)
do j =1, JI-1 
write(10,*) x(j),u(j)
end do
write(10,*)x(JI),u(JI),']'
write(10,*) "plot(ApproximateSolution(:,1),ApproximateSolution(:,2),'r')"
write(10,*) "xlabel('x'),ylabel('temperature'),legend('Approximate C-N Scheme')"
close(10)
contains

subroutine thomas (a,b,c,d)

  real, intent(inout) :: a(:),b(:),c(:),d(:)
  integer j

  do j = 2,ji !combined decomposition and forward substitution
! a - sub-diagonal (means it is the diagonal below the main diagonal)
! b - the main diagonal
! c - sup-diagonal (means it is the diagonal above the main diagonal)
! d - right part
   
    a(j) = a(j)/b(j-1)
    b(j) = b(j)-a(j)*c(j-1)
    d(j) = d(j)-a(j)*d(j-1)
  end do

  !back substitution
  d(ji) = d(ji)/b(ji)
  do j = ji-1,1,-1
    d(j) = (d(j)-c(j)*d(j+1))/b(j)
  end do

end subroutine thomas

end program crank_nicolson

(THERE ARE NO ERRORS IN THE CODE. JUST NEED TO MODIFY THE ABOVE CODE Allowing for the diffusivity D(u) to change discontinuously WITH the case D(u)=1 when x<1/2 and D(u)=1/2 otherwise.)

 

 

3 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

Is there a real MIC issue here, or are you simply asking us to solve your homework question? 

If the latter you'll be out of luck, and even if someone did think it worth playing with, how would you know if they had done it right, rather than producing something that would make you look like an idiot when you submitted it?

Taylor Kidd (Intel)'s picture

I notice that you posted this at least twice, probably to different forums.

Unless you have a real issue relevant to the forums, please do not post here. We are a special resource for those very difficult technical issues having to do with Intel tools and products. If you abuse the forums now, you may be ignored later even if you have a very relevant issue we can help you with.

Regards
--
Taylor
 

Login to leave a comment.