Solver settings

FluxCore settings balance accuracy, robustness and runtime without exposing implementation-specific solver architecture. Begin with validated defaults and change a setting only when the physics or a sensitivity study justifies it.

Simulation mode

Mode	Use for
Steady	Design points where a stable mean flow or thermal state is the main result.
Time resolved	Varying loads, wakes, rotating motion, start-up or thermal transients.
Conjugate heat transfer	Coupled fluid convection and solid conduction across multiple regions.

Accuracy profile

Profile	Use for
Robust	Initial checks, difficult geometry or a new operating point.
Standard	Most production engineering runs.
High accuracy	Final studies on a suitable mesh with a validation target.

Higher accuracy cannot recover geometry or gradients that the mesh does not resolve. Review mesh quality and sensitivity before escalating numerical settings.

Run controls

Control	Purpose
Maximum iterations	Limits solver work for a steady run.
Residual target	Defines the requested equation-convergence threshold.
Physical steps	Sets the length of a time-resolved study.
Time step	Resolves the physical time scale of interest.
Inner work budget	Controls the convergence effort within each physical step.
Output cadence	Selects how often accepted states are written for time history.
Early stopping	Ends a run after residual and engineering-monitor criteria are met.

Engineering monitors

Select at least one quantity tied directly to the design decision: force, moment, pressure loss, mass flow, temperature, heat rate or another component output. Judge convergence from residuals and engineering monitors together.

Reference values

Force and moment coefficients require explicit reference area, velocity, density, directions, moment centre and reference lengths. Reuse the same definitions across variants. An automatic projected area is useful for exploration but is not a universal reporting standard.

When to revise a setting

Revise settings when residuals and physical monitors disagree, the case is genuinely unsteady, a mesh or time-step study shows sensitivity, or a benchmark requires a defined accuracy level. Do not use a larger run budget to conceal incorrect geometry, units, materials or boundary conditions.