[SOLVED] Help Me: 2D FDTD Electromagnetic Wave Propagation code in C program

Nenita · 11-08-2014, 12:24 PM

Hi everyone!! My name is Nenita, I'm looking for help on 2D FDTD Electromagnetic Wave Propagation code in C program. Please anyone with knowledge of this please help me!

psionl0 · 11-08-2014, 12:33 PM

What have you done so far?

Nenita · 11-08-2014, 12:38 PM

This is what I've done so far, but when I run the program, it does not show a wave graph. It's just a flat surface.

Code:

# include <stdio.h>
# include <stdlib.h>
# include <math.h>

# define IE 60
# define JE 60

int main ()
{
  double ez[IE][JE], hx[IE][JE], hy[IE][JE];
  int l, n, i, j, ic, jc, nsteps;
  double dx, dy, dt, T, epsz, pi, sigma, mu;
  double t0, spread, pulse;
  double c  =  2.99792458e8;  // propagation velocity
  double freq  = 1.0e9;
  double alamda  =  c/freq;
  FILE *fp, *fopen();

  ic  =  IE/2;
  jc  =  JE/2;
  dx  =  .01;        /* cell size */
  dy  =  .01;        /* cell size */
  dt  =  1.0/freq/6e8;     /* Time steps */
  epsz  =  8.8e-12;
  pi  = 3.14159;
  mu  = 1.0e-7*4.0*pi;
  sigma  = 1.0e-8 ;

  for( j = 0; j < JE; j++) {
    //printf("%2d", j);
    for(i = 0; i < IE; i++) {
      ez[i][j]  =  0.;
      hx[i][j]  =  0.;
      hy[i][j]  =  0.;
    }
    printf("\n");
  }

  t0 = 20.0;
  spread = 6.0;
  T = 0;
  nsteps = 1;

  while(nsteps > 0){
    printf("nsteps -> ");
    scanf("%d", &nsteps);
    printf("%d \n", nsteps);

    for(n = 1; n <= nsteps; n++) {
      T = T + 1;

        /*  - Start of Main FDTD loop  - */

      /*  -  Put a Gaussian pulse in the middle  -  */
    
      pulse = sin(2*pi*1500*1e6*dt);
      ez[ic][jc] = pulse;

      /*  -  Calculate Ez field  -  */

     for(j = 1; j < JE; j++){
        for(i = 1; i < IE; i++) {
          ez[i][j] = (1.0 - sigma*dt/(2.0*epsz))/(1.0+sigma*dt)/(2.0*epsz)*ez[i][j]+(dt/epsz)
            /(1.0+sigma*dt/(2.0*epsz*(1/dx))*(hy[i][j] - hy[i - 1][j]) - (dt/epsz)/((1.0+sigma*dt)/(2.0*epsz))*(1/dy)
            *(hx[i][j] - hx[i][j - 1]));
        }}

      /*  -  Calculate Hx field  -  */
      for(j = 0; j < JE - 1; j++){
        for(i = 0; i < IE - 1; i++){
          hx[i][j] = hx[i][j] - (dt/mu*1/dy)*(ez[i][j+1] - ez[i][j]);
        }
      }
      
      /* Calculate Hy field */
      for(j = 0; j < JE - 1; j++){
        for(i = 0; i <= IE - 1; i++) {
       hy[i][j] = hy[i][j]+(dt/mu*1/dx)*(ez[i+1][j] - ez[i][j]);
        }
      }
    }

    /*  -  End of main FDTD loop  -  */

    for(j = 1; j < jc; j++) {
      printf("%2d", j);
      for(i = 1; i < ic; i++) {
        printf("%5.2f", ez[2*i][2*j]);
      }
      printf("\n");
    }
    printf("T = %5.0f \n", T);

    /* Write E field out to file "Ez" */

    fp = fopen("Ez", "w");
    for(j = 0; j < JE; j++) {
      for(i = 0; i < IE; i++) {
        fprintf(fp,"%d\t%d\t%6.3f\n",i, j, ez[i][j]);
      }
      fprintf(fp, "\n");
    }
    fclose(fp);
  }
}

norobro · 11-08-2014, 01:36 PM

I know nothing about this, but just eyeballing your code, the ez formula looks suspect to me.

Code:

ez[i][j] = (1.0 - sigma*dt/(2.0*epsz))/(1.0+sigma*dt)/(2.0*epsz)*ez[i][j]+(dt/epsz)
            /(1.0+sigma*dt/(2.0*epsz*(1/dx))*(hy[i][j] - hy[i - 1][j]) - (dt/epsz)/((1.0+sigma*dt)/(2.0*epsz))*(1/dy)
            *(hx[i][j] - hx[i][j - 1]));

All of the elements of the arrays ez, hx, and hy are 0.

Nenita · 11-08-2014, 01:59 PM

Quote:

Originally Posted by norobro

I know nothing about this, but just eyeballing your code, the ez formula looks suspect to me.

Code:

ez[i][j] = (1.0 - sigma*dt/(2.0*epsz))/(1.0+sigma*dt)/(2.0*epsz)*ez[i][j]+(dt/epsz)
            /(1.0+sigma*dt/(2.0*epsz*(1/dx))*(hy[i][j] - hy[i - 1][j]) - (dt/epsz)/((1.0+sigma*dt)/(2.0*epsz))*(1/dy)
            *(hx[i][j] - hx[i][j - 1]));

All of the elements of the arrays ez, hx, and hy are 0.

I'm also very new to the topic

I've tried re-writing ez, hx and hy formulas but I'm stuck so I reached out to other programmers who may be able to help me find what the problem is.

metaschima · 11-08-2014, 02:22 PM

Put some printfs in there to output the values as you go. If you know the correct values you should be able to find out what is wrong.

jlinkels · 11-08-2014, 02:38 PM

The basics of writing such a program:

- Make sure you have the formulas correct in writing, not in your program
- Calculate some values by hand so you know what the matrix element values must be
- Translate the formulas in C
- Put printf statements after each calculation which print the index variables (i and j), the calculation result and further any input values to formulas which are variable as well.

Anything less than that tends to demonstrate lazyness.

jlinkels