Bug: Prevent spurious powder consolidation following mesh refinement (#108)

Author: colemanjs

applications/solvers/additiveFoam/additiveFoam.C

@@ -94,8 +94,6 @@ int main(int argc, char *argv[])
 
     while (runTime.run())
     {
-        #include "updateProperties.H"
-
         #include "readDyMControls.H"
         #include "CourantNo.H"
         #include "setDeltaT.H"
@@ -104,6 +102,22 @@ int main(int argc, char *argv[])
 
         mesh.update();
 
+        if (mesh.topoChanged())
+        {
+            // Temperature is authoritative after topology changes. A
+            // conservative scheme should consider enthalpy as authoritative.
+            forAll(mesh.cells(), cellI)
+            {
+                scalar alpha1_ =
+                    interpolateXY(T[cellI], thermo.x(), thermo.y());
+                alpha1[cellI] = min(max(alpha1_, 0.0), 1.0);
+            }
+
+            alpha1.correctBoundaryConditions();
+
+            #include "updateProperties.H"
+        }
+
         runTime++;
 
         Info<< "Time = " << runTime.name() << nl << endl;
@@ -123,6 +137,8 @@ int main(int argc, char *argv[])
 
         #include "thermo/TEqn.H"
 
+        #include "updateProperties.H"
+
         runTime.write();
 
         Info<< "ExecutionTime = " << runTime.elapsedCpuTime() << " s"

applications/solvers/additiveFoam/updateProperties.H

@@ -1,13 +1,10 @@
+alpha3 *= pos0(solidus - T);
+alpha3.correctBoundaryConditions();
+
 forAll(mesh.cells(), cellI)
 {
-    // update phase fractions
     scalar alpha1_ = alpha1[cellI];
     scalar alpha2_ = 1.0 - alpha1[cellI];
-
-    if (alpha1_ < 1.0)
-    {
-        alpha3[cellI] = 0.0;
-    }
     scalar alpha3_ = alpha3[cellI];
 
     // update temperature ranges for phase properties
@@ -24,6 +21,5 @@ forAll(mesh.cells(), cellI)
       + alpha3_*Cp3.value(T1);
 }
 
-alpha3.correctBoundaryConditions();
 Cp.correctBoundaryConditions();
 kappa.correctBoundaryConditions();