M01 volumetric material response
===============================

This is genuine three-dimensional small-strain linear tetrahedral elasticity on
a20mm solid cube. It is not a device animation, beam approximation or fracture
model. Each node has x/y/z displacement DOFs; the volume contains TET4 elements.
The bottom face is fully fixed. The top face carries a uniform finite-area1N peak
normal force in the negative z direction. Other faces are traction-free.

E, density and Poisson ratio come from the rounded primary-source input record.
The three baselines are generic structural steel, nominal6061 aluminum and
HPFS7980 fused silica. These are catalog approximations, not measured specimens.
Fused silica is amorphous, not crystalline quartz. Hardness, yield, toughness,
flaws, grains, anisotropy, piezoelectricity and crack growth are not model inputs.

The mesh sequence is2/4/6/8subdivisions per axis with6tetrahedra per cell. TET4
stiffness uses engineering shear strains, exact constant element gradients and
consistent mass. Forty-eight mass-normalized elastic modes are retained, with a
24-mode comparison. The161frequency samples span20,000–200,000Hz, not20Hz.

Damping is illustrative Rayleigh C=alpha*M+beta*K, matching the chosen ratio at
20kHz and200kHz. Baseline ratio0.01 is compared with0.005/0.02 at the finest mesh.
These loss assumptions are not material-specific measured damping. Results can
depend strongly on this choice and on boundary/interface constraints.

Displacement is a complex PEAK amplitude with exp(+i*omega*t) convention. Physical
replay is Ureal*cos(theta)-Uimag*sin(theta). The largest instantaneous vector norm
at each node is calculated from its2x2 phase quadratic, not the complex norm.
Eigenvectors are normalized phi.T*M*phi=1 and have units kg^(-1/2). Their displayed
amplitude is arbitrary and must be labeled. A selected basis in a degenerate mode
cluster is not a uniquely trackable mode identity.

Stress is recovered as constant element-average stress. Reported harmonic peak
von Mises uses the exact2x2 real/imag stress quadratic. The global metric is RMS
over time and volume, not a simultaneous peak. Clamp-edge maxima are mesh-dependent
and are not fracture limits. No damage or safe specimen load is inferred.

Run in a Python environment with NumPy and SciPy:
  python -m unittest -v test_fem3d.py
  python run_study.py --output fresh-results
  python make_figures.py --results fresh-results --output fresh-figures

Output folders must be new. Frozen source/material hashes are checked before and
after the study. All convergence failures are retained separately from numerical
algebra checks. The source plan preserves its historical review-pending header;
execution follows a separate approval without changing its matrix.

M01_web.json contains a full volume mesh, six surface modal vectors and the actual
selected physical forced complex vectors/stresses. Its display values are rounded
to9significant digits; the NPZ/CSV preserve float64 data. WEB-SCHEMA.txt defines the
mapping and units. Numerical animation can exaggerate deformation or slow phase,
but must show that gain and the actual calculated frequency/time scale.

There is no contacting hardware, acoustic coupling/laser, amplifier calibration,
fracture experiment, purchase or physical actuation in this model. A1N prescribed
traction does not establish any source device can deliver that load at these
frequencies. All plotted results are numerical calculations, not physical trials.
