mod_nonlinear_roe.t

Go to the documentation of this file.

!> Module containing Roe solver for scalar nonlinear equation
module mod_nonlinear_roe
  use mod_physics_roe
  use mod_nonlinear_phys
 
  implicit none
  private
 
  public :: nonlinear_roe_init
 
contains
 
  subroutine nonlinear_roe_init()
    use mod_physics_roe
 
    nworkroe = 1
 
    phys_average         => nonlinear_average
    phys_get_eigenjump   => nonlinear_get_eigenjump
    phys_rtimes          => nonlinear_rtimes
  end subroutine nonlinear_roe_init
 
  subroutine nonlinear_average(wL, wR, x, ix^L, idim, wroe, workroe)
    use mod_global_parameters
    integer, intent(in)             :: ix^L, idim
    double precision, intent(in)    :: wL(ixG^T, nw), wR(ixG^T, nw)
    double precision, intent(inout) :: wroe(ixG^T, nw)
    double precision, intent(inout) :: workroe(ixG^T, nworkroe)
    double precision, intent(in)    :: x(ixG^T, 1:^ND)
 
    wroe(ix^s, rho_)=half*(wl(ix^s, rho_)+wr(ix^s, rho_))
  end subroutine nonlinear_average
 
  subroutine nonlinear_get_eigenjump(wL, wR, wC, x, ix^L, il, idim, smalla, a, jump, workroe)
 
    ! Calculate the characteristic speed a and the jump in the
    ! characteristic variable in the idim direction within ixL.
    ! For a scalar equation the characteristic and conservative variables coincide
    ! The characteristic speed is just the velocity, but it should be averaged
    ! for the cell interfaces
 
    use mod_global_parameters
 
    integer, intent(in)                          :: ix^L, il, idim
    double precision, dimension(ixG^T, nw)       :: wL, wR, wC
    double precision, dimension(ixG^T)           :: smalla, a, jump, v
    double precision, dimension(ixG^T, nworkroe) :: workroe
    double precision, intent(in)                 :: x(ixG^T, 1:^ND)
    integer                                      :: jx^L, ixC^L
 
    jx^l=ix^l+kr(idim,^d);
    ixcmin^d=ixmin^d; ixcmax^d=jxmax^d;
 
    ! No entropy fix
    smalla(ix^s)= -one
    ! The velocity is dependent of w in the nonlinear scalar equation,
    ! and thus depends on the location
    !> TODO: check this, for advection added argument to get velocity at cell edge!!!
    call nonlinear_get_v(wl, x, ixg^ll, ixc^l, idim, v)
 
    a(ix^s)=(v(jx^s)+v(ix^s))/2
 
    jump(ix^s)=wr(ix^s, rho_)-wl(ix^s, rho_)
 
  end subroutine nonlinear_get_eigenjump
 
  subroutine nonlinear_rtimes(q, w, ix^L, iw, il, idim, rq, workroe)
 
    ! Multiply q by R(il, iw), where R is the right eigenvalue matrix at wC.
    ! For a scalar equation the R matrix is unity
 
    use mod_global_parameters
    integer, intent(in)             :: ix^L, iw, il, idim
    double precision, intent(in)    :: w(ixG^T, nw), q(ixG^T)
    double precision, intent(inout) :: rq(ixG^T)
    double precision, intent(inout) :: workroe(ixG^T, nworkroe)
 
    rq(ix^s)=q(ix^s)
 
  end subroutine nonlinear_rtimes
 
end module mod_nonlinear_roe