Bug: Prevent spurious powder consolidation following mesh refinement

[colemanjs]

After topology changes, field mapping can lead to small floating point differences. During coarsening, the solid fraction field alpha1 can become non-unary in the solid, which leads to:

1. consolidation according to the previous model:

if (alpha1[celli] < 1)
{
    alpha3[celli] = 1.0;
}

2. Fictitious latent heat release in the mapped solid.

NOTE: A better long term solution for energy conservation might be to make the enthalpy the authoritative quantity for field mapping.

[kincaidkc]

Looks good. I never used the powder model during my testing so never noticed this issue. Thanks for fixing.