C1. Method of manufactured solutions for EB
Case description
This is a similar test case problem as performed for PeleC verification using MASA to construct a manufactured solution. The MMS is performed around a sphere at a Re number and Mach number of 1.
\(L_2\) error norm of fields





Note
The first order convergence observed here is expected because the treatment of the EB surface is first order.
Running study
paren=`pwd`
pelec="${paren}/PeleC3d.gnu.MPI.ex"
mpi_ranks=36
res=( 8 16 32 64 )
for i in "${res[@]}"
do
rm -rf "${i}"
mkdir "${i}"
cd "${i}" || exit
cp "${paren}/inputs_3d" .
srun -n ${mpi_ranks} "${pelec}" inputs_3d amr.n_cell="${i} ${i} ${i}" > out
ls -1v *plt*/Header | tee movie.visit
cd "${paren}" || exit
done