Update 6/14

  • Implemented multispecies gas for Roe solver in AstroBEAR, but need a solution for density or sound speed for van der Waals gas to implement it. Relevant code snippet: 
     ! Entropy fix      !!! Implement a different entropy fix to use van der Waals gas - density equations have complicated/no solution
     lambdaLL = WL(ivx) - PrimSoundSpeed(WL)
     lambdaRR = WR(ivx) + PrimSoundSpeed(WR)
     if (lambdaLL < 0 .or. lambdaRR > 0) then
      call StarPU(WL((/ irho, iE, ivx /)), WR((/ irho, iE, ivx /)), PrimSoundSpeed(WL), PrimSoundSpeed(WR), pstar, ustar)
      ! Passively advected (tracers; y,z momentum)
      ExactL = WL
      ExactR = WR
      ! Solution in interaction region
      ExactL(iE) = pstar
      ExactL(ivx) = ustar
      ExactR(iE) = pstar
      ExactR(ivx) = ustar
      ! Densities differ
      if (mpi_id == 0 .and. iEOS == EOS_vanDerWaals) write (*,*) 'This solution for density will not be accurate. The attraction parameter is not yet included in these calculations.'
      ExactL(irho) = PrimExactDensity(WL,pstar)
      ExactR(irho) = PrimExactDensity(WR,pstar)
      
      lambdaLR = ustar - PrimSoundSpeed(ExactL)
      lambdaRL = ustar + PrimSoundSpeed(ExactR)
     end if
     if (lambdaLL < 0 .and. lambdaLR > 0) then    ! Check for transonic rarefaction on left side
      print *, 'Using entropy fix.'
      lambda(1) = lambdaLL*((lambdaLR - lambda(1))/(lambdaLR - lambdaLL))
     end if
     if (lambdaRL < 0 .and. lambdaRR > 0) then    ! check for transonic rarefaction on right side
      print *, 'Using entropy fix.'
      lambda(nDim+2) = lambdaRR*((lambda(nDim+2) - lambdaRL)/(lambdaRR - lambdaRL))
     end if
  • Wrote WAF scheme for own solver. Working, but results are not as good as I expect based on Toro's. Not yet sure where the problem lies.

http://www.pas.rochester.edu/~adebrech/code/debugging/test1_HLLC.gif

  • Related: Transmissive boundary conditions for second set of ghost cells don't quite make sense yet. Why move back inward to get wave to flow out?
  • Still getting explosion (but better than implosion) with adiabatic atmosphere.

Should be the same as Jonathan's result - not sure why it appears to not be working.

Attachments (4)

Comments

No comments.