Merge branch 'v0.6-dev' into HLL_MHD_Breaking

NEWS.md

@@ -15,6 +15,8 @@ for human readability.
   conceptually identical across equations.
   Users, who have been using `flux_hll` for MHD have now to use `flux_hlle` in order to use the
   Einfeldt wave speed estimate.
+- Parabolic diffusion terms are now officially supported and not marked as experimental
+  anymore.
 
 #### Deprecated
 
@@ -37,6 +39,7 @@ for human readability.
 - Implementation of the polytropic Euler equations in 2D
 - Subcell positivity limiting support for conservative variables in 2D for `TreeMesh`
 - AMR for hyperbolic-parabolic equations on 2D/3D `TreeMesh`
+- Added `GradientVariables` type parameter to `AbstractEquationsParabolic`
 
 #### Changed
 

README.md

@@ -18,7 +18,7 @@
   <img width="300px" src="https://trixi-framework.github.io/assets/logo.png">
 </p>
 
-**Trixi.jl** is a numerical simulation framework for hyperbolic conservation
+**Trixi.jl** is a numerical simulation framework for conservation
 laws written in [Julia](https://julialang.org). A key objective for the
 framework is to be useful to both scientists and students. Therefore, next to
 having an extensible design with a fast implementation, Trixi.jl is
@@ -46,6 +46,7 @@ installation and postprocessing procedures. Its features include:
 * Periodic and weakly-enforced boundary conditions
 * Multiple governing equations:
   * Compressible Euler equations
+  * Compressible Navier-Stokes equations
   * Magnetohydrodynamics (MHD) equations
   * Multi-component compressible Euler and MHD equations
   * Linearized Euler and acoustic perturbation equations
@@ -56,6 +57,7 @@ installation and postprocessing procedures. Its features include:
 * Multi-physics simulations
   * [Self-gravitating gas dynamics](https://github.com/trixi-framework/paper-self-gravitating-gas-dynamics)
 * Shared-memory parallelization via multithreading
+* Multi-node parallelization via MPI
 * Visualization and postprocessing of the results
   * In-situ and a posteriori visualization with [Plots.jl](https://github.com/JuliaPlots/Plots.jl)
   * Interactive visualization with [Makie.jl](https://makie.juliaplots.org/)

docs/literate/src/files/DGMulti_1.jl

@@ -168,7 +168,7 @@ meshIO = StartUpDG.triangulate_domain(StartUpDG.RectangularDomainWithHole());
 
 # The pre-defined Triangulate geometry in StartUpDG has integer boundary tags. With [`DGMultiMesh`](@ref)
 # we assign boundary faces based on these integer boundary tags and create a mesh compatible with Trixi.jl.
-mesh = DGMultiMesh(meshIO, dg, Dict(:outer_boundary=>1, :inner_boundary=>2))
+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,

docs/literate/src/files/adding_new_parabolic_terms.jl

@@ -18,8 +18,15 @@ equations_hyperbolic = LinearScalarAdvectionEquation2D(advection_velocity);
 # `ConstantAnisotropicDiffusion2D` has a field for `equations_hyperbolic`. It is useful to have
 # information about the hyperbolic system available to the parabolic part so that we can reuse
 # functions defined for hyperbolic equations (such as `varnames`).
-
-struct ConstantAnisotropicDiffusion2D{E, T} <: Trixi.AbstractEquationsParabolic{2, 1}
+#
+# The abstract type `Trixi.AbstractEquationsParabolic` has three parameters: `NDIMS` (the spatial dimension,
+# e.g., 1D, 2D, or 3D), `NVARS` (the number of variables), and `GradientVariable`, which we set as
+# `GradientVariablesConservative`. This indicates that the gradient should be taken with respect to the
+# conservative variables (e.g., the same variables used in `equations_hyperbolic`). Users can also take
+# the gradient with respect to a different set of variables; see, for example, the implementation of
+# [`CompressibleNavierStokesDiffusion2D`](@ref), which can utilize either "primitive" or "entropy" variables.
+
+struct ConstantAnisotropicDiffusion2D{E, T} <: Trixi.AbstractEquationsParabolic{2, 1, GradientVariablesConservative}
   diffusivity::T
   equations_hyperbolic::E
 end

docs/src/index.md

@@ -12,7 +12,7 @@
 [![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.3996439.svg)](https://doi.org/10.5281/zenodo.3996439)
 
 [**Trixi.jl**](https://github.com/trixi-framework/Trixi.jl)
-is a numerical simulation framework for hyperbolic conservation
+is a numerical simulation framework for conservation
 laws written in [Julia](https://julialang.org). A key objective for the
 framework is to be useful to both scientists and students. Therefore, next to
 having an extensible design with a fast implementation, Trixi.jl is
@@ -40,6 +40,7 @@ installation and postprocessing procedures. Its features include:
 * Periodic and weakly-enforced boundary conditions
 * Multiple governing equations:
   * Compressible Euler equations
+  * Compressible Navier-Stokes equations
   * Magnetohydrodynamics (MHD) equations
   * Multi-component compressible Euler and MHD equations
   * Linearized Euler and acoustic perturbation equations
@@ -50,6 +51,7 @@ installation and postprocessing procedures. Its features include:
 * Multi-physics simulations
   * [Self-gravitating gas dynamics](https://github.com/trixi-framework/paper-self-gravitating-gas-dynamics)
 * Shared-memory parallelization via multithreading
+* Multi-node parallelization via MPI
 * Visualization and postprocessing of the results
   * In-situ and a posteriori visualization with [Plots.jl](https://github.com/JuliaPlots/Plots.jl)
   * Interactive visualization with [Makie.jl](https://makie.juliaplots.org/)

docs/src/overview.md

@@ -60,10 +60,9 @@ different features on different mesh types.
 | Flux differencing                                            |          ✅         |             ✅            |               ✅              |          ✅          |               ✅            | [`VolumeIntegralFluxDifferencing`](@ref)
 | Shock capturing                                              |          ✅         |             ✅            |               ✅              |          ✅          |               ❌            | [`VolumeIntegralShockCapturingHG`](@ref)
 | Nonconservative equations                                    |          ✅         |             ✅            |               ✅              |          ✅          |               ✅            | e.g., GLM MHD or shallow water equations
-| Parabolic termsᵇ                                             |          ✅         |             ✅            |               ❌              |          ✅          |               ✅            | e.g., [`CompressibleNavierStokesDiffusion2D`](@ref)
+| Parabolic terms                                              |          ✅         |             ✅            |               ❌              |          ✅          |               ✅            | e.g., [`CompressibleNavierStokesDiffusion2D`](@ref)
 
 ᵃ: quad = quadrilateral, hex = hexahedron
-ᵇ: Parabolic terms do not currently support adaptivity. 
 
 ## Time integration methods
 

src/Trixi.jl

@@ -157,7 +157,7 @@ export LaplaceDiffusion1D, LaplaceDiffusion2D, LaplaceDiffusion3D,
        CompressibleNavierStokesDiffusion1D, CompressibleNavierStokesDiffusion2D,
        CompressibleNavierStokesDiffusion3D
 
-export GradientVariablesPrimitive, GradientVariablesEntropy
+export GradientVariablesConservative, GradientVariablesPrimitive, GradientVariablesEntropy
 
 export flux, flux_central, flux_lax_friedrichs, flux_hll, flux_hllc, flux_hlle,
        flux_godunov,

src/equations/compressible_navier_stokes.jl

@@ -6,9 +6,6 @@ Creates a wall-type boundary conditions for the compressible Navier-Stokes equat
 The fields `boundary_condition_velocity` and `boundary_condition_heat_flux` are intended
 to be boundary condition types such as the `NoSlip` velocity boundary condition and the
 `Adiabatic` or `Isothermal` heat boundary condition.
-
-!!! warning "Experimental feature"
-    This is an experimental feature and may change in future releases.
 """
 struct BoundaryConditionNavierStokesWall{V, H}
     boundary_condition_velocity::V
@@ -52,9 +49,6 @@ struct Adiabatic{F}
 end
 
 """
-!!! warning "Experimental code"
-    This code is experimental and may be changed or removed in any future release.
-
 `GradientVariablesPrimitive` and `GradientVariablesEntropy` are gradient variable type parameters
 for `CompressibleNavierStokesDiffusion1D`. By default, the gradient variables are set to be
 `GradientVariablesPrimitive`. Specifying `GradientVariablesEntropy` instead uses the entropy variable

src/equations/compressible_navier_stokes_1d.jl

@@ -79,14 +79,11 @@ where
 ```math
 w_2 = \frac{\rho v1}{p},\, w_3 = -\frac{\rho}{p}
 ```
-
-!!! warning "Experimental code"
-    This code is experimental and may be changed or removed in any future release.
 """
 struct CompressibleNavierStokesDiffusion1D{GradientVariables, RealT <: Real,
                                            E <: AbstractCompressibleEulerEquations{1}
                                            } <:
-       AbstractCompressibleNavierStokesDiffusion{1, 3}
+       AbstractCompressibleNavierStokesDiffusion{1, 3, GradientVariables}
     # TODO: parabolic
     # 1) For now save gamma and inv(gamma-1) again, but could potentially reuse them from the Euler equations
     # 2) Add NGRADS as a type parameter here and in AbstractEquationsParabolic, add `ngradients(...)` accessor function

src/equations/compressible_navier_stokes_2d.jl

@@ -79,14 +79,11 @@ where
 ```math
 w_2 = \frac{\rho v_1}{p},\, w_3 = \frac{\rho v_2}{p},\, w_4 = -\frac{\rho}{p}
 ```
-
-!!! warning "Experimental code"
-    This code is experimental and may be changed or removed in any future release.
 """
 struct CompressibleNavierStokesDiffusion2D{GradientVariables, RealT <: Real,
                                            E <: AbstractCompressibleEulerEquations{2}
                                            } <:
-       AbstractCompressibleNavierStokesDiffusion{2, 4}
+       AbstractCompressibleNavierStokesDiffusion{2, 4, GradientVariables}
     # TODO: parabolic
     # 1) For now save gamma and inv(gamma-1) again, but could potentially reuse them from the Euler equations
     # 2) Add NGRADS as a type parameter here and in AbstractEquationsParabolic, add `ngradients(...)` accessor function

src/equations/compressible_navier_stokes_3d.jl

@@ -79,14 +79,11 @@ where
 ```math
 w_2 = \frac{\rho v_1}{p},\, w_3 = \frac{\rho v_2}{p},\, w_4 = \frac{\rho v_3}{p},\, w_5 = -\frac{\rho}{p}
 ```
-
-!!! warning "Experimental code"
-    This code is experimental and may be changed or removed in any future release.
 """
 struct CompressibleNavierStokesDiffusion3D{GradientVariables, RealT <: Real,
                                            E <: AbstractCompressibleEulerEquations{3}
                                            } <:
-       AbstractCompressibleNavierStokesDiffusion{3, 5}
+       AbstractCompressibleNavierStokesDiffusion{3, 5, GradientVariables}
     # TODO: parabolic
     # 1) For now save gamma and inv(gamma-1) again, but could potentially reuse them from the Euler equations
     # 2) Add NGRADS as a type parameter here and in AbstractEquationsParabolic, add `ngradients(...)` accessor function

src/equations/equations.jl

@@ -479,6 +479,6 @@ abstract type AbstractLinearizedEulerEquations{NDIMS, NVARS} <:
               AbstractEquations{NDIMS, NVARS} end
 include("linearized_euler_2d.jl")
 
-abstract type AbstractEquationsParabolic{NDIMS, NVARS} <:
+abstract type AbstractEquationsParabolic{NDIMS, NVARS, GradientVariables} <:
               AbstractEquations{NDIMS, NVARS} end
 end # @muladd

src/equations/equations_parabolic.jl

@@ -2,16 +2,21 @@
 # specialize this function to compute gradients e.g., of primitive variables instead of conservative
 gradient_variable_transformation(::AbstractEquationsParabolic) = cons2cons
 
+# By default, the gradients are taken with respect to the conservative variables.
+# this is reflected by the type parameter `GradientVariablesConservative` in the abstract
+# type `AbstractEquationsParabolic{NDIMS, NVARS, GradientVariablesConservative}`.
+struct GradientVariablesConservative end
+
 # Linear scalar diffusion for use in linear scalar advection-diffusion problems
 abstract type AbstractLaplaceDiffusion{NDIMS, NVARS} <:
-              AbstractEquationsParabolic{NDIMS, NVARS} end
+              AbstractEquationsParabolic{NDIMS, NVARS, GradientVariablesConservative} end
 include("laplace_diffusion_1d.jl")
 include("laplace_diffusion_2d.jl")
 include("laplace_diffusion_3d.jl")
 
 # Compressible Navier-Stokes equations
-abstract type AbstractCompressibleNavierStokesDiffusion{NDIMS, NVARS} <:
-              AbstractEquationsParabolic{NDIMS, NVARS} end
+abstract type AbstractCompressibleNavierStokesDiffusion{NDIMS, NVARS, GradientVariables} <:
+              AbstractEquationsParabolic{NDIMS, NVARS, GradientVariables} end
 include("compressible_navier_stokes.jl")
 include("compressible_navier_stokes_1d.jl")
 include("compressible_navier_stokes_2d.jl")

src/solvers/dgmulti/types.jl

@@ -433,15 +433,3 @@ function LinearAlgebra.mul!(b_in, A_kronecker::SimpleKronecker{3}, x_in)
     return nothing
 end
 end # @muladd
-
-# TODO: deprecations introduced in Trixi.jl v0.6
-@deprecate DGMultiMesh(dg::DGMulti{NDIMS}; cells_per_dimension, kwargs...) where {NDIMS} DGMultiMesh(dg,
-                                                                                                     cells_per_dimension;
-                                                                                                     kwargs...)
-
-# TODO: deprecations introduced in Trixi.jl v0.5
-@deprecate DGMultiMesh(vertex_coordinates, EToV, dg::DGMulti{NDIMS};
-                       kwargs...) where {NDIMS} DGMultiMesh(dg, vertex_coordinates, EToV;
-                                                            kwargs...)
-@deprecate DGMultiMesh(triangulateIO, dg::DGMulti{2, Tri}, boundary_dict::Dict{Symbol, Int};
-                       kwargs...) DGMultiMesh(dg, triangulateIO, boundary_dict; kwargs...)