Settings

Global (solver-wide) settings for ClaymoreUW that are not specific to any body, boundary, or sensor are collected here. They are grouped into:

• Simulation Settings (Ambient, Spatial, Temporal)

• Scaling Settings

• Computer Settings (Deprecated)

Simulation Settings

General solver controls that govern stability, accuracy, and runtime.

Ambient

Parameters that ambiently affect the entire simulation (not tied to space/time discretization).

Parameter	Description	Typical / Notes
CFL Number	Courant–Friedrichs–Lewy stability number used by the material models to compute a stable time step.	Keep in 0.3–0.5 for robust runs.
Gravity (vector)	Uniform acceleration applied globally.	Example: [0, -9.80665, 0] (m/s²)

Important

A lower CFL yields more conservative (smaller) time steps → more stable but slower. A higher CFL approaches the stability limit and risks divergence.

Spatial

Parameters defining the computational grid and domain.

Parameter	Description	Typical / Notes
Grid-Cell Size (dx)	MPM background grid spacing. Directly impacts accuracy and cost (smaller dx → more cells/steps).	Choose to adequately resolve features of interest.
Domain (vector / bbox)	Bounding box of the computational domain (full extent).	Ensure all bodies and boundary motions remain inside for full duration.
Mirror (Deprecated)	Legacy symmetry mirroring across coordinate planes for efficiency.	Avoid for new studies.

Note

As you refine dx, you must also reduce the stable time step due to the CFL condition. Expect runtime to increase superlinearly with refinement.

Temporal

Parameters controlling the simulation clock and integration.

Parameter	Description	Typical / Notes
Max Time Step (dt)	Upper bound on the time step. The effective step is computed by the material models from the CFL and material stiffness; the solver uses the minimum of this value and the CFL-based step.	Acts as a safety cap; do not set above the stable CFL step.
Time-Integration	Currently explicit scheme.	Implicit not yet available.
Initial Time (t0)	Time datum (start time).	Usually 0.
Duration (tf - t0)	Total simulated time to run.	Pick long enough to capture the full response.

Tip

A practical rule of thumb is Δt_effective ≤ CFL * dx / c_max where c_max is the problem’s max wave/signal speed (material dependent). Keep CFL ≤ 0.5 for headroom.

Scaling Settings

Controls to enforce similitude when mapping between model-scale and prototype-scale.

Setting	Description	Notes
Dominant Law	Choose Froude or Cauchy similitude as the governing scaling law.	Determines which primary ratios you specify.
Froude inputs	Provide Length Ratio and Time Ratio; other quantities auto-scale for consistency (e.g., velocities, accelerations).	Useful for gravity-dominated free-surface flows.
Cauchy inputs	Provide Elasticity Ratio (and Length Ratio as needed); elastic and inertial effects are scaled to maintain stress similitude.	Useful for elastic/inertial similarity (e.g., solid response).

Warning

Ensure your units are consistent before and after scaling. Check that derived properties (e.g., densities, moduli, gravity) are scaled as intended by the selected law. Mismatched units or incompatible ratios can silently invalidate results.

Note

This feature may expand in future versions. For now, only the listed inputs are user-editable; other parameters are adjusted automatically to maintain the chosen similitude.

Computer Settings (Deprecated)

Legacy controls tied to the host system.

Parameter	Description	Status
GPU Count	Select number of GPUs to target.	Deprecated
Compile-Time Memory	Static memory reservations used by older builds.	Deprecated

Note

These settings are retained for backwards compatibility and may be removed in future releases. Prefer runtime auto-configuration and modern build presets.

Best-Practice Hints

• Start with CFL = 0.4, moderate dx, then refine based on accuracy checks.

• Verify the Domain encloses all motion for the entire Duration.

• Use Scaling Settings to translate model-scale studies to prototype behavior, but validate with at least one unscaled baseline case.