Verification, Interpretation, and Project Completion

A completed run is not necessarily a verified solution. Assess numerical quality, physical plausibility, and sensitivity before using the results for an engineering conclusion.

Model checks before production runs

Run a reduced model first. Check:

• dimensions, gravity direction, and unit consistency;

• material-point locations, volumes, and groups;

• background-node spacing and active support region;

• constitutive models and properties;

• initial stresses and pore pressures;

• boundary constraints and rigid-body modes;

• timestep, damping, and PIC fraction;

• contact and hybrid coupling behavior.

Histories

Use model history for global solution quantities and mpoint history for local material-point response.

model history mechanical ratio-average
model history name 'Time' dynamic time-total

mpoint history name 'Settlement' displacement component z ...
    position [monitor_position]
mpoint history name 'Vertical velocity' velocity component z ...
    position [monitor_position]
mpoint history name 'Vertical stress' stress quantity zz ...
    position [monitor_position]

Choose history locations before large motion makes a purely coordinate-based interpretation ambiguous. When possible, identify and store a material-point ID or track a group.

Resolution studies

Vary the background-node spacing and material-point resolution independently. A complete study may include:

• background-grid refinement at fixed particle density;

• particle refinement at fixed background spacing;

• combined refinement;

• timestep refinement for dynamic models;

• PIC-fraction and damping sensitivity;

• conversion interval and hybrid radius sensitivity.

Interpretation of stress and strain

Confirm the stress sign convention and distinguish total from effective stress in coupled analyses. Examine principal stress, maximum shear stress, plastic state, deformation, velocity, and energy together. A localized contour should be interpreted with the material-point distribution, background resolution, and constitutive regularization in mind.

Project completion

Archive:

• all data files and FISH functions;

• input geometry and property sources;

• saved states needed to reproduce figures;

• assumptions, limitations, and known numerical issues.