Update documentation - CHANGELOG, README

CHANGELOG.md

@@ -54,6 +54,10 @@ The format of this changelog is based on
   - Added Apptainer/Singularity container build definition for Palace.
   - Fixed bugs related to thread-safety for OpenMP builds and parallel tetrahedral meshes in
     the upstream MFEM library.
+  - Introduced Adaptive Mesh Refinement, specified in the `config["Model"]["Refinement"]`,
+    for all problem types aside from transient. To enable AMR, a user must specify
+    `"MaxIts"`, while all other options have reasonable defaults. Nonconformal (all mesh
+    types) and conformal (simplex meshes) refinement are supported.
 
 ## [0.11.2] - 2023-07-14
 

docs/src/config/boundaries.md

@@ -99,8 +99,9 @@ ports can be specified on boundaries which are internal to the computational dom
 
 `"WavePort"` :  Array of objects for configuring numeric wave port boundary conditions. Wave
 ports can only be specified on boundaries which are on the true boundary of the
-computational domain. Addiitionally, wave port boundaries are only available for the
-frequency domain driven simulation type.
+computational domain. Additionally, wave port boundaries are only available for the
+frequency domain driven simulation type, and are presently not compatible with nonconformal
+mesh refinement.
 
 `"WavePortPEC"` :  Top-level object for configuring PEC boundary conditions for boundary
 mode analysis performed on the wave port boundaries. Thus, this object is only relevant

docs/src/config/model.md

@@ -73,8 +73,8 @@ with
 mesh exceeds this value no further adaptation will occur. A value less than 1 means that no
 maximum size constraint will be imposed.
 
-`"Nonconformal" [true]` : Chose whether the adaptation should use nonconformal refinement
-for simplex meshes.
+`"Nonconformal" [true]` : Chose whether the adaptation should use nonconformal refinement.
+Nonconformal refinement is required for non-simplex meshes.
 
 `"UpdateFraction" [0.7]` : Dörfler marking fraction used to specify which elements to
 refine. This marking strategy will mark the smallest number of elements that make up

docs/src/guide/model.md

@@ -43,8 +43,11 @@ maintains a conforming mesh but meshes containing hexahedra, prism, or pyramid e
 will be nonconforming after local refinement (this is not supported at this time).
 
 [Adaptive mesh refinement (AMR)](https://en.wikipedia.org/wiki/Adaptive_mesh_refinement)
-according to error estimates in the computed solution is a work in progress for all
-simulation types.
+according to error estimates calculated from the computed solution can also be specified
+using the [`config["Model"]["Refinement"]`](../config/model.md#model%5B%22Refinement%22%5D)
+object. Nonconformal refinement is supported for all mesh types, and additionally conformal
+refinement is supported for simplex meshes. AMR is available for all problem types apart
+from driven problems in the time domain.
 
 ## Material models
 

docs/src/index.md

@@ -38,6 +38,10 @@ the frequency or time domain, using the [MFEM finite element discretization libr
     (config/solver.md#solver["Linear"]), including geometric multigrid (GMG), parallel
     sparse direct solvers, and algebraic multigrid (AMG) preconditioners, for fast
     performance on platforms ranging from laptops to HPC systems.
+  - Solution based [Adaptive Mesh Refinement
+    (AMR)](config/model.md#model%5B%22Refinement%22%5D) for all simulation types aside
+    from transient. Nonconformal refinement is supported for all mesh types, and conformal
+    refinement for simplex meshes.
 
 ## Contents
 

palace/drivers/basesolver.cpp

@@ -187,7 +187,7 @@ void BaseSolver::SolveEstimateMarkRefine(
                     ? ", maximum size = " + std::to_string(refinement.max_size)
                     : ""));
 
-    // Optionally save of the previous solution.
+    // Optionally save off the previous solution.
     if (refinement.save_adapt_iterations)
     {
       SaveIteration(comm, post_dir, it,

palace/fem/errorindicator.cpp

@@ -28,7 +28,7 @@ void ErrorIndicator::AddIndicator(const ErrorIndicator &indicator)
   //                            E = √(1/N ∑ₙ ∑ₖ ηₖₙ²)
   // from which it follows that:
   //                            E² = 1/N ∑ₙ ∑ₖ ηₖₙ²
-  //                               = 1/N ∑ₙ Eₙ
+  //                               = 1/N ∑ₙ Eₙ²
   // Namely the average of the global error indicators included in the reduction.
   // Squaring both sides means the summation can be rearranged, and then the local error
   // indicators become: