Unify boundary conditions

NEWS.md

@@ -8,18 +8,27 @@ for human readability.
 ## Changes when updating to v0.15 from v0.14.x
 
 #### Changed
+
+- Boundary conditions are now passed as a `NamedTuple` consistently across all mesh types, i.e., boundary conditions, which
+  were previously passed as a `Dict` for the `P4estMesh`, `T8codeMesh`, and `UnstructuredMesh2D` mesh types have to be converted to `NamedTuple`s. Also passing boundary conditions as a `Tuple` is no longer supported. This change also makes `boundary_condition_default` obsolete, which is why it was removed ([#2761]).
+- The `TreeMesh`, `StructuredMesh` as well as the hypercube constructors of `P4estMesh` and `T8codeMesh` are now non-periodic by default, i.e., you need to
+  explicitly set the `periodicity` keyword argument to `true` to obtain periodic meshes. This change was made to be consistent with other mesh types and to avoid confusion ([#2770]).
+- There are no default boundary conditions anymore in `SemidiscretizationHyperbolic` and `SemidiscretizationHyperbolicParabolic`.
+  Users now have to explicitly provide boundary conditions via the `boundary_conditions` keyword argument, which were periodic by default before ([#2770]).
 - Renamed `energy_internal` for `NonIdealCompressibleEulerEquations1D` to `energy_internal_specific` ([#2762]). This makes `energy_internal` for `NonIdealCompressibleEulerEquations1D` consistent with the rest of Trixi.jl, where `energy_internal` refers to the specific internal energy scaled by density.
-- Changed `cons2prim(u, ::NonIdealCompressibleEulerEquations1D)` so that it returns `rho, v1, p`. Previously, `cons2prim` returned `V, v1, T`; this functionalityis now provided by the non-exported function `cons2thermo` ([#2769]). 
+- Changed `cons2prim(u, ::NonIdealCompressibleEulerEquations1D)` so that it returns `rho, v1, p`. Previously, `cons2prim` returned `V, v1, T`; this functionality is now provided by the non-exported function `cons2thermo` ([#2769]).
 - The variable name (printed to the screen and files) of the total energy in the compressible Euler equations has been changed from `rho_e` to `rho_e_total` to avoid confusion with the internal energy `rho_e_internal` ([#2778]).
 - `convergence_test` now returns the complete convergence orders and the full errors matrix. To obtain the mean convergence rates, use `Trixi.calc_mean_convergence` on the convergence orders ([#2753]).
 - The serial and parallel mesh types have been renamed from `SerialTreeMesh`, `ParallelTreeMesh`, `SerialP4estMesh`, `ParallelP4estMesh`, `SerialT8codeMesh`, and `ParallelT8codeMesh` to `TreeMeshSerial`, `TreeMeshParallel`, `P4estMeshSerial`, `P4estMeshParallel`, `T8codeMeshSerial`, and `T8codeMeshParallel`, respectively ([#2787]).
 
 #### Added
+
 - Added `PengRobinson` equation of state ([#2769]).
 
 ## Changes in the v0.14 lifecycle
 
 #### Added
+
 - Added `NonIdealCompressibleEulerEquations1D`, which allows users to specify a non-ideal equation of state. Currently `IdealGas` and `VanDerWaals` are supported ([#2739]).
 - Added the APEC (approximate pressure equilibrium preserving with conservation) fluxes of `flux_terashima_etal` and `flux_central_terashima_etal` from [Terashima, Ly, Ihme (2025)](https://doi.org/10.1016/j.jcp.2024.113701) ([#2756]).
 - Support for second-order finite volume subcell volume integral (`VolumeIntegralPureLGLFiniteVolumeO2`) and
@@ -41,6 +50,7 @@ for human readability.
 ## Changes in the v0.13 lifecycle
 
 #### Added
+
 - Initial 3D support for subcell limiting with `P4estMesh` was added ([#2582], [#2647], [#2688], [#2722]).
   In the new version, IDP positivity limiting for conservative variables (using
   the keyword `positivity_variables_cons` in `SubcellLimiterIDP()`) and nonlinear

docs/literate/src/files/DGMulti_1.jl

@@ -172,8 +172,8 @@ meshIO = StartUpDG.triangulate_domain(StartUpDG.RectangularDomainWithHole());
 mesh = DGMultiMesh(dg, meshIO, Dict(:outer_boundary => 1, :inner_boundary => 2))
 #-
 boundary_condition_convergence_test = BoundaryConditionDirichlet(initial_condition)
-boundary_conditions = (; :outer_boundary => boundary_condition_convergence_test,
-                       :inner_boundary => boundary_condition_convergence_test)
+boundary_conditions = (; outer_boundary = boundary_condition_convergence_test,
+                       inner_boundary = boundary_condition_convergence_test)
 
 semi = SemidiscretizationHyperbolic(mesh, equations, initial_condition, dg,
                                     source_terms = source_terms,

docs/literate/src/files/adaptive_mesh_refinement.jl

@@ -117,9 +117,11 @@ coordinates_min = (-5.0, -5.0)
 coordinates_max = (5.0, 5.0)
 mesh = TreeMesh(coordinates_min, coordinates_max,
                 initial_refinement_level = 4,
-                n_cells_max = 30_000)
+                n_cells_max = 30_000,
+                periodicity = true)
 
-semi = SemidiscretizationHyperbolic(mesh, equations, initial_condition, solver)
+semi = SemidiscretizationHyperbolic(mesh, equations, initial_condition, solver;
+                                    boundary_conditions = boundary_condition_periodic)
 
 tspan = (0.0, 10.0)
 ode = semidiscretize(semi, tspan);

docs/literate/src/files/adding_new_scalar_equations.jl

@@ -42,11 +42,13 @@ initial_condition_sine(x, t, equation::CubicEquation) = SVector(sinpi(x[1]))
 
 mesh = TreeMesh(-1.0, 1.0, # min/max coordinates
                 initial_refinement_level = 4,
-                n_cells_max = 10^4)
+                n_cells_max = 10^4,
+                periodicity = true)
 
 solver = DGSEM(3, flux_central) # set polynomial degree to 3
 
-semi = SemidiscretizationHyperbolic(mesh, equation, initial_condition_sine, solver)
+semi = SemidiscretizationHyperbolic(mesh, equation, initial_condition_sine, solver;
+                                    boundary_conditions = boundary_condition_periodic)
 
 # We wrap the return value of the `initial_condition_sine` inside an `SVector` since that's the approach
 # used in Trixi.jl also for systems of equations. We need to index the spatial coordinate `x[1]`,
@@ -103,7 +105,8 @@ plot!(sol)
 
 ## A larger final time: Nonclassical shocks develop (you can even increase the refinement to 12)
 semi = remake(semi,
-              mesh = TreeMesh(-1.0, 1.0, initial_refinement_level = 8, n_cells_max = 10^5))
+              mesh = TreeMesh(-1.0, 1.0, initial_refinement_level = 8, n_cells_max = 10^5,
+                              periodicity = true))
 ode = semidiscretize(semi, (0.0, 0.5)) # set tspan to (0.0, 0.5)
 sol = solve(ode, SSPRK43(); ode_default_options()...)
 plot(sol)
@@ -184,11 +187,13 @@ end
 
 mesh = TreeMesh(-1.0, 1.0, # min/max coordinates
                 initial_refinement_level = 4,
-                n_cells_max = 10^4)
+                n_cells_max = 10^4,
+                periodicity = true)
 
 solver = DGSEM(3, flux_central) # set polynomial degree to 3
 
-semi = SemidiscretizationHyperbolic(mesh, equation, initial_condition_sine, solver)
+semi = SemidiscretizationHyperbolic(mesh, equation, initial_condition_sine, solver;
+                                    boundary_conditions = boundary_condition_periodic)
 
 ## Create ODE problem with given time span
 tspan = (0.0, 0.1)
@@ -206,7 +211,8 @@ plot!(sol)
 
 ## A larger final time: Nonclassical shocks develop (you can even increase the refinement to 12)
 semi = remake(semi,
-              mesh = TreeMesh(-1.0, 1.0, initial_refinement_level = 8, n_cells_max = 10^5))
+              mesh = TreeMesh(-1.0, 1.0, initial_refinement_level = 8, n_cells_max = 10^5,
+                              periodicity = true))
 ode = semidiscretize(semi, (0.0, 0.5))
 sol = solve(ode, SSPRK43(); ode_default_options()...)
 plot(sol)

docs/literate/src/files/adding_nonconservative_equation.jl

@@ -106,7 +106,7 @@ end
 
 ## Create a uniform mesh in 1D in the interval [-π, π] with periodic boundaries
 mesh = TreeMesh(-Float64(π), Float64(π), # min/max coordinates
-                initial_refinement_level = 4, n_cells_max = 10^4)
+                initial_refinement_level = 4, n_cells_max = 10^4, periodicity = true)
 
 ## Create a DGSEM solver with polynomials of degree `polydeg`
 ## Remember to pass a tuple of the form `(conservative_flux, nonconservative_flux)`
@@ -117,7 +117,8 @@ solver = DGSEM(polydeg = 3, surface_flux = surface_flux,
                volume_integral = VolumeIntegralFluxDifferencing(volume_flux))
 
 ## Setup the spatial semidiscretization containing all ingredients
-semi = SemidiscretizationHyperbolic(mesh, equation, initial_condition_sine, solver)
+semi = SemidiscretizationHyperbolic(mesh, equation, initial_condition_sine, solver;
+                                    boundary_conditions = boundary_condition_periodic)
 
 ## Create an ODE problem with given time span
 tspan = (0.0, 1.0)
@@ -150,9 +151,10 @@ error_1 = analysis_callback(sol).l2 |> first
 # simulation again.
 
 mesh = TreeMesh(-Float64(π), Float64(π), # min/max coordinates
-                initial_refinement_level = 5, n_cells_max = 10^4)
+                initial_refinement_level = 5, n_cells_max = 10^4, periodicity = true)
 
-semi = SemidiscretizationHyperbolic(mesh, equation, initial_condition_sine, solver)
+semi = SemidiscretizationHyperbolic(mesh, equation, initial_condition_sine, solver;
+                                    boundary_conditions = boundary_condition_periodic)
 
 tspan = (0.0, 1.0)
 ode = semidiscretize(semi, tspan);
@@ -254,7 +256,7 @@ end
 
 ## Create a uniform mesh in 1D in the interval [-π, π] with periodic boundaries
 mesh = TreeMesh(-Float64(π), Float64(π), # min/max coordinates
-                initial_refinement_level = 4, n_cells_max = 10^4)
+                initial_refinement_level = 4, n_cells_max = 10^4, periodicity = true)
 
 ## Create a DGSEM solver with polynomials of degree `polydeg`
 ## Remember to pass a tuple of the form `(conservative_flux, nonconservative_flux)`
@@ -265,7 +267,8 @@ solver = DGSEM(polydeg = 3, surface_flux = surface_flux,
                volume_integral = VolumeIntegralFluxDifferencing(volume_flux))
 
 ## Setup the spatial semidiscretization containing all ingredients
-semi = SemidiscretizationHyperbolic(mesh, equation, initial_condition_sine, solver)
+semi = SemidiscretizationHyperbolic(mesh, equation, initial_condition_sine, solver;
+                                    boundary_conditions = boundary_condition_periodic)
 
 ## Create an ODE problem with given time span
 tspan = (0.0, 1.0)

docs/literate/src/files/behind_the_scenes_simulation_setup.jl

@@ -33,7 +33,8 @@ coarsening_patches = ((type = "box", coordinates_min = [0.0, -2.0],
 
 mesh = TreeMesh(coordinates_min, coordinates_max, initial_refinement_level = 2,
                 n_cells_max = 30_000,
-                coarsening_patches = coarsening_patches)
+                coarsening_patches = coarsening_patches,
+                periodicity = true)
 
 # The created `TreeMesh` looks like the following:
 
@@ -58,7 +59,8 @@ solver = DGSEM(polydeg = 3)
 # comes into play.
 
 semi = SemidiscretizationHyperbolic(mesh, equations, initial_condition_convergence_test,
-                                    solver)
+                                    solver;
+                                    boundary_conditions = boundary_condition_periodic)
 
 # The constructor for the `SemidiscretizationHyperbolic` object calls numerous sub-functions to
 # perform the necessary initialization steps. A brief description of the key sub-functions is

docs/literate/src/files/custom_semidiscretization.jl

@@ -87,6 +87,7 @@ end
 
 semi = SemidiscretizationHyperbolic(mesh, equations, initial_condition,
                                     solver;
+                                    boundary_conditions = boundary_condition_periodic,
                                     source_terms = source_terms_standard)
 
 # Now, we can create the `ODEProblem`, solve the resulting ODE