User: rhaas Date: 2012/05/16 11:28 AM

Modified: /trunk/ param.ccl /trunk/src/ GRHydro_HLLEM.F90

Log: GRHydro: decouple Bvec(flux_direction) from other two for divclean

since it and psidc form a 2d subsystem of the characteristic system. See eg. Mignonoe et al. arXiv:1001.2832v2 section 3.1.

File Changes:

Directory: /trunk/src/ ======================

File [modified]: GRHydro_HLLEM.F90 Delta lines: +6 -0 =================================================================== --- trunk/src/GRHydro_HLLEM.F90 2012-05-14 18:47:06 UTC (rev 334) +++ trunk/src/GRHydro_HLLEM.F90 2012-05-16 16:28:21 UTC (rev 335) @@ -442,6 +442,12 @@ !!$ Wavespeeds for psidc are +/-c, not Fast Magnetosonic? psidcf = 0.5d0 * (1.d0 * psidcfp - (-1.d0) * psidcfm + & 1.d0 * (-1.d0) * psidcdiff) + if(decouple_normal_Bfield .ne. 0) then ! same expression for HLLE and LLF + !!$ B^i field decouples from the others and has same propagation + !!$ speed as divergence + f1(5+flux_direction) = 0.5d0 * (1.d0 * fplus(5+flux_direction) - (-1.d0) * fminus(5+flux_direction) + & + 1.d0 * (-1.d0) * qdiff(5+flux_direction)) + end if case(2) charmax = setcharmax charmin = setcharmin

Directory: /trunk/ ==================

File [modified]: param.ccl Delta lines: +4 -0 =================================================================== --- trunk/param.ccl 2012-05-14 18:47:06 UTC (rev 334) +++ trunk/param.ccl 2012-05-16 16:28:21 UTC (rev 335) @@ -549,6 +549,10 @@ -1:0 :: "Any value smaller than speed of light - sign should be negative" } -1.0

+BOOLEAN decouple_normal_Bfield "when using divergence cleaning properly decouple Bx,psidc subsystem" +{ +} yes + boolean track_divB "Track the magnetic field constraint violations" { } "no"