Items agreed to during April 28th lunch discussion on NIMROD/M3D nonlinear comparison.  Present:  W. Park, S. Kruger, J. Breslau, S. Jardin

 

  1. Both codes to start with the exact same, accurate equilibrium, with q(0) near 0.955.   Jardin to send email documenting where this equilibrium comes from and where the files are.
  2. Initialize runs to a pure n=1 mode with the ratio of the maximum perturbed magnetic field to the vacuum toroidal magnetic field on axis to be 10**(-4).
  3. Use Sptizer resistivity, with a constant specified ZEFF (consistent with the equilibrium specification).
  4. Use a viscosity equal to 10 times the resistivity.
  5. Start with the equilibrium density, and evolve it in time using the continuity equation.
  6. Specify a constant loop voltage of zero on the boundary.
  7. In the energy equation, include parallel thermal conductivity with a realistic value of parallel conductivity.

 

 

In addition, the M3D run will be modified in the following ways:

  1. Use finer toroidal grid spacing (note NIMROD is including up to n=10)
  2. Use a smaller “linear chopping” so that the linear phase can be followed to see more clearly when the nonlinear effects set in.

 

 

Items to be compared:

  1. Kinetic energy vs time in the different toroidal modes.
  2. Magnetic energy vs time in the different toroidal modes.
  3. Crash time.
  4. Island widths vs time.

Followup to the 7/25/03 conference call involving Schnack, Kruger, Park, Breslau, and Jardin:

I generated 5 new tar equilibrium files of the CDX-U discharge with the same (surface averaged) current profile as before, but with the pressure profile multiplied by a fraction between .01 and 1.0. The files can be found at the PPPL ftp site: /u/ftp/pub/jardin/cdxu The names of the files are:

where number after "f" at the end is the factor that the pressure was multiplied by before calculating the equilibrium. Thus, the first file should be the same as the previous file run05time09 that was distributed and analyzed.

The other items we agreed to are as follows:

  1. Both codes will use chi_|| = chi_perp = 200 m**2/sec in the energy equation to model isotropic thermal conduction with the approximate experimental value.
  2. The resistivity being used is approximately the spitzer value ~ 5.1 x 10**-5 in the core, increasing towards the edge. Josh will determine what the maximum value of the resistivity should be for the "cuttoff" value and send an email to the others, who will use this same cuttoff to impose a maximum resistivity.
  3. There will be another conference call on Friday Aug 15th at the same time (2:00 Eastern Time Zone).
  4. Well before the Aug 15th conference call, M3D and NIMROD will each converge their linear n=1 eigenmodes and exchange graphics files and growth rates so that the linear comparison will not be an issue at that time. Jardin will post any material on the project web site.
  5. Each team may conduct side-studies, for example, of the dependence of the linear mode structure and growth rate on the pressure, and or, on the thermal conductivity. These side-studies will likely be included in the final paper that results from this activity.
  6. Both teams will explore ways of providing a "high-n cuttoff" for the nonlinear studies. This will be discussed at the next conference call.