Update compute_forces_viscoelastic.F90

Correct a bug which broke axisymmetric simulations
SPECFEM · Nov 10, 2020 · 47a8f3f · 47a8f3f
1 parent a213b4b
commit 47a8f3f
Showing 1 changed file with 6 additions and 0 deletions.
diff --git a/src/specfem2D/compute_forces_viscoelastic.F90 b/src/specfem2D/compute_forces_viscoelastic.F90
@@ -304,6 +304,7 @@ subroutine compute_forces_viscoelastic(accel_elastic,veloc_elastic,displ_elastic
                 sigma_zz = lambdaplus2mu_unrelaxed_elastic*duz_dzl(i,j) &
                          + lambdal_unrelaxed_elastic * (dux_dxl(i,j) + sigma_zz/xxi)
                 sigma_xz = mul_unrelaxed_elastic*(duz_dxl(i,j) + dux_dzl(i,j))
+                sigma_zx = sigma_xz
                 sigma_thetatheta(i,j) = lambdal_unrelaxed_elastic * (duz_dzl(i,j) + dux_dxl(i,j)) &
                                    + lambdaplus2mu_unrelaxed_elastic*sigma_thetatheta(i,j)/xxi
               else
@@ -313,6 +314,7 @@ subroutine compute_forces_viscoelastic(accel_elastic,veloc_elastic,displ_elastic
                 sigma_zz = lambdaplus2mu_unrelaxed_elastic*duz_dzl(i,j) &
                            + lambdal_unrelaxed_elastic * (dux_dxl(i,j) + dummy_loc(1,i,j)/coord(1,ibool(i,j,ispec)))
                 sigma_xz = mul_unrelaxed_elastic*(duz_dxl(i,j) + dux_dzl(i,j))
+                sigma_zx = sigma_xz
                 sigma_thetatheta(i,j) = lambdal_unrelaxed_elastic * (duz_dzl(i,j) + dux_dxl(i,j)) &
                                         + lambdaplus2mu_unrelaxed_elastic * dummy_loc(1,i,j)/coord(1,ibool(i,j,ispec))
                 r_xiplus1(i,j) = coord(1,ibool(i,j,ispec))/(xiglj(i)+ONE)
@@ -324,6 +326,7 @@ subroutine compute_forces_viscoelastic(accel_elastic,veloc_elastic,displ_elastic
               sigma_zz = lambdaplus2mu_unrelaxed_elastic*duz_dzl(i,j) &
                          + lambdal_unrelaxed_elastic * (dux_dxl(i,j) + dummy_loc(1,i,j)/coord(1,ibool(i,j,ispec)))
               sigma_xz = mul_unrelaxed_elastic*(duz_dxl(i,j) + dux_dzl(i,j))
+              sigma_zx = sigma_xz
               sigma_thetatheta(i,j) = lambdal_unrelaxed_elastic * (duz_dzl(i,j) + dux_dxl(i,j)) &
                                       + lambdaplus2mu_unrelaxed_elastic * dummy_loc(1,i,j)/coord(1,ibool(i,j,ispec))
             endif
@@ -390,12 +393,14 @@ subroutine compute_forces_viscoelastic(accel_elastic,veloc_elastic,displ_elastic
                 sigma_xx = c11*dux_dxl(i,j) + c13*duz_dzl(i,j) + c12*sigma_xx/xxi
                 sigma_zz = c13*dux_dxl(i,j) + c33*duz_dzl(i,j) + c23*sigma_zz/xxi
                 sigma_xz = c15*dux_dxl(i,j) + c35*duz_dzl(i,j) + c55*(duz_dxl(i,j) + dux_dzl(i,j))
+                sigma_zx = sigma_xz
                 sigma_thetatheta(i,j) = c12*dux_dxl(i,j) + c23*duz_dzl(i,j) + c22*sigma_thetatheta(i,j)/xxi
               else
                 ! not first GLJ point but not on the axis
                 sigma_xx = c11*dux_dxl(i,j) + c13*duz_dzl(i,j) + c12*dummy_loc(1,i,j)/coord(1,ibool(i,j,ispec))
                 sigma_zz = c13*dux_dxl(i,j) + c33*duz_dzl(i,j) + c23*dummy_loc(1,i,j)/coord(1,ibool(i,j,ispec))
                 sigma_xz = c15*dux_dxl(i,j) + c35*duz_dzl(i,j) + c55*(duz_dxl(i,j) + dux_dzl(i,j))
+                sigma_zx = sigma_xz
                 sigma_thetatheta(i,j) = c12*dux_dxl(i,j) + c23*duz_dzl(i,j) + &
                                         c22*dummy_loc(1,i,j)/coord(1,ibool(i,j,ispec))
               endif
@@ -404,6 +409,7 @@ subroutine compute_forces_viscoelastic(accel_elastic,veloc_elastic,displ_elastic
               sigma_xx = c11*dux_dxl(i,j) + c13*duz_dzl(i,j) + c12*dummy_loc(1,i,j)/coord(1,ibool(i,j,ispec))
               sigma_zz = c13*dux_dxl(i,j) + c33*duz_dzl(i,j) + c23*dummy_loc(1,i,j)/coord(1,ibool(i,j,ispec))
               sigma_xz = c15*dux_dxl(i,j) + c35*duz_dzl(i,j) + c55*(duz_dxl(i,j) + dux_dzl(i,j))
+              sigma_zx = sigma_xz
               sigma_thetatheta(i,j) = c12*dux_dxl(i,j) + c23*duz_dzl(i,j) + c22*dummy_loc(1,i,j)/coord(1,ibool(i,j,ispec))
             endif
           else