Staged-coupling workflow, gravity/PID/smoothingModel gotchas, prefer …

[cgoniva]

…defaults

Learned building a coupled cfdemSolverInter melting-tracer showcase case (AspherixCFDEM feature_meltingTracer_1088). Single source of truth is the official docs; entries here are genuinely undocumented operational knowledge, not restatements of what's already covered on doc.aspherix-dem.com.

RULES.md: new Default Values rule — prefer a couplingProperties/*Props keyword's documented default over restating it explicitly.

STRATEGIES.md, six new entries:

• Build a coupled case in complexity stages (isothermal -> heat -> phase change) instead of all at once.
• Use Autorun.sh/run.config (based_on/change_dicts/depends_on) for a case's own execution by default, matching the tutorials.
• New CLASSY_BLOCKS.md/CHANGE_DICTS.md: classy_blocks for O-grid blockMeshDicts, and how to patch a dict for a run.config variant without copying the whole file.
• A DEM/CFD gravity mismatch (Section_settings.html) only warns, never errors, so a crash-signature-only log monitor misses it.
• Use a smoothingModel once particles are large compared to the local cell size (pointer to smoothingModels.md).
• Track the actual launch's PIDs, not a fresh process search, before trusting a coupled run has ended (cfdemSimulate.html's own note on DEM/CFD process independence).