NonlinearStatic.arc_length {constraint, max_load_factor, displacement_limit}: Riks–Crisfield
arc-length increments (cylindrical or spherical arc, tangent predictor signed by the angle
criterion, the constraint's quadratic root chosen by the angle with the last iterate), arcs
grown, shrunk and cut back as increments are. The converged state's tangent is factored for the
next predictor (LDLᵀ when indefinite); its negative eigenvalues are recorded per increment and
changes in their count are reported as limit or bifurcation points. Ends at the maximum load
factor, a displacement limit or the increment cap. Abaqus *STATIC, RIKS; LS-DYNA refuses it by
name; forms gain the arc-length group. A shallow arch of hex20 snaps through on the closed-form
path (1.9 %), both limit points found, inertia 0 → 1 → 0 between them.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01UhhZMBTviDQxZvwnyP89fo
Procedure::Buckle {modes, shift}: the step's loads as a linear preload, K_σ from its stresses,
and (K + λ K_σ) φ = 0 by block Lanczos in the K inner product: K⁻¹(−K_σ) on Cholesky without a
shift, the buckling transformation (K + σ K_σ)⁻¹ K on LDLᵀ with one, whose inertia counts the
load factors below σ (Grimes–Lewis–Simon 1994). Lanczos gains an operand and an eigenvalue map
(defaults keep frequency steps bit for bit) and reports converged negative eigenvalues. Results
are frequency results with a `buckling` part; modes display with their load factor.
solve.buckle (in the Static drop-down); Abaqus *BUCKLE, Nastran SOL 105 and LS-DYNA refuse it by
name. Euler columns of hex20 within 0.18 % of Engesser's load.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01UhhZMBTviDQxZvwnyP89fo
Segment-to-segment penalty contact between element faces: each secondary
face and each main face facing it are projected onto the secondary face's
plane, split into triangles, clipped (Sutherland–Hodgman) and integrated
with a three-point rule per sub-triangle (Puso–Laursen segmentation, the
penalty at the integration points as Fischer–Wriggers), so a uniform
pressure gives the consistent nodal forces on both meshes and the contact
patch test passes on non-matching meshes. Every contact runs both ways at
half the penalty: pairs with each surface as the secondary, general and
self contact face against face. A bucket grid of face boxes every
increment, faces integrated in parallel and their forces added in face
order: the same numbers on any thread count.
The penalty per area ε = s γ min(m/A)/Δt² comes from the lumped masses,
each node's directional area and the stable increment, bounded by
Gershgorin's row sums so contact never shortens the increment. Coulomb
friction with exponential decay by a regularised return map at the
secondary nodes; rigid walls take Coulomb impulses. Contact work and
friction dissipation join the energy balance; each contact reports its
peak force, first and last contact, deepest penetration, friction and the
nodes adjusted at the start (initial penetrations adjusted strain-free or
refused with the node, face and depth); a cap on contact faces.
Verified (crates/explicit/tests/contact.rs, tests/oracle/explicit-*):
colliding bars (2L/c +1.2%, ρcAv₀/2, KE kept to 0.005%, contact energy
back to 8e-5 of it), blocks sliding to a stop on a rigid wall and on a
plate under constant and decaying friction (0.005–0.2%), the patch test
(σ_zz uniform to 0.06%), two spheres colliding against Hertz (peak +2.3%,
duration −0.4%; the oracle case waits for a graded sphere mesher), a free
body untouched by general contact, determinism across threads. Oracle
measures gain `over` (ratios). Existing fixtures do not move.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
A supernodal LDLᵀ on the Cholesky factor's symbolic factorization (ordering, supernodes, fill,
cap and parallel subtree tasks shared through a Symbolic and a generic driver; the Cholesky
numerics are unchanged bit for bit). Pivots by the bounded Bunch–Kaufman (rook) search of
Ashcraft, Grimes & Lewis among each supernode's columns (Schenk & Gärtner's supernode
Bunch–Kaufman), 1 × 1 and 2 × 2, right-looking panels whose trailing update reuses the tiled
Cholesky kernels with L D formed on the fly. Reports the inertia; a column without a pivot above
round-off is a located Singular error; NaN is refused. Symmetric::new picks Cholesky, LDLᵀ, or
Cholesky with an LDLᵀ fallback; solve_refined adds iterative refinement. 1.1–1.2× the Cholesky
time on a 140 000-equation hex8 pattern.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
Five cases against closed forms: a Mooney-Rivlin block in Treloar's
uniaxial, equibiaxial and pure-shear modes and a Yeoh block in uniaxial
tension (the compressible energy's stresses, exact); a stretched
Mooney-Rivlin tube inflated by a follower pressure (Rivlin's incompressible
solution, 0.6 % at nu = 0.4995); free and fully held thermal expansion at a
thermal stretch of 1.2 (exact); a bar held at its ends heated past yield
(exact). docs/fem.md, the literature, ROADMAP (M15, M17), parity and the
oracle page follow.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
Four cases, each against its closed form: uniaxial tension of a J2 cube to a
logarithmic strain of 0.2 (Kirchhoff stress exact; 0.13 % from the explicit
solver's hypoelastic Cauchy stress, the elastic volume change between the two
laws), the Bauschinger effect of linear kinematic hardening on a load path
back to zero strain (exact), Hill's elastic-perfectly plastic thick cylinder
(bore displacement at a plastic radius of 1.5a within 0.009 %, collapse
pressure within 2e-8; the pending case is now active) and a cantilever's
plastic collapse, converged 3.1 % above beam theory's hinge, the bound a
continuum exceeds. docs/fem.md, the literature, ROADMAP (M15), parity and the
oracle page follow.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
Theory cases for the nonlinear solver, each frozen as a fixture: uniaxial
stretch of a neo-Hookean block to twice its length (closed form), the
nonlinear patch test on a distorted hex8 mesh, a quarter turn of the same
mesh (objectivity: zero stress), the elastica of a hex20 cantilever at
P L²/(E I) = 1, 2, 5, 10 against Bisshopp–Drucker (0.2–0.5 %), and the
inflation of a thick neo-Hookean tube by a follower pressure against
Rivlin's closed form (0.08 % at the bore). A finite-strain Cook's membrane
is pending its transcribed reference. docs/fem.md describes the
formulation, the material point interface, the follower-pressure choice and
what nonlinear steps refuse; the literature, parity and roadmap pages
follow.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
Crates fieldcraft-*, binaries fieldcraft and fieldcraft-cli, the .fieldcraft file (format tag
fieldcraft/1, manifest fieldcraft.json), FIELDCRAFT_* environment variables, the MCP server
name, egui widget ids (fc_*), the app icon document FieldCraft.icon and its generated family,
docs and licences. No compatibility shim: no files of the earlier name exist.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
The isotropic elasticity splits into D_dev + K m mᵀ; an element-wise
discontinuous pressure is condensed out (p̂ = K M⁻¹ Gᵀ u), leaving
K_dev + K G M⁻¹ Gᵀ (Hughes §4.3–4.5). Q2–P1disc and P2–P0 are inf-sup
stable; Q1–P0 does not lock; tet4 and wedge6 still lock and are refused, as
are orthotropic materials (named at the section). ElementDofs counts the
condensed pressures. Verified: element patch, rigid modes and totals for
every hybrid kind, a block in uniaxial strain at ν = 0.4999 (exact), Lamé's
cylinder in plane strain at ν = 0.4999 (hex20 1e-5, tet10 1.3e-3, hex8
2.5e-3 vs the locking displacement hex8 at −80 %). Every non-hybrid result
of the test suites hashes identically before and after.
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
Co-authored-by: jyviko <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Heat transfer and thermal stress: thermal materials, temperature BCs, heat loads, heat transfer and coupled temperature–displacement steps, solve.heat
Model: Material.thermal (conductivity, specific heat, expansion, isotropic or
per axis; reference temperature), BoundaryCondition::Temperature (dof 11),
loads SurfaceFlux, Convection, HeatSource and Temperature (a field in static
steps), procedures HeatTransfer { transient (θ-method) } and
CoupledTemperatureDisplacement; checks for missing conductivity/specific
heat/expansion. fem: heat conduction kernels (conductivity, capacity,
source, face flux/film, heat flux) on the DofLayout (one T per node), thermal
strain load and stress D(ε − ε_th) for every solid formulation incl.
incompatible modes. Solver: steady and transient heat transfer, thermal loads
in static steps, coupled steps (exact one-way coupling). Engine: solve.heat,
temperature/heat-flux result fields, probes. Decks refuse heat transfer by
name. Analytic verification: slab, source+convection/flux, thick cylinder,
erfc semi-infinite solid, free/constrained expansion, bimetal strip, coupled
step.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* thermal: strains in 2D elements and explicit dynamics
Plane::thermal_load and stresses_with for plane stress (εzz free), plane
strain (σzz = −D₃·ε_th) and axisymmetric elements, incompatible modes
included; the solver's 2D kernels take them. Explicit dynamics turns a
step's temperature loads into a per-element thermal strain rate (reached
linearly over the step) subtracted from each strain increment, so it feeds
J2 and Johnson–Cook updates. Verified: free expansion in every 2D theory
and element, a biaxially held plate, an explicit free bar (ΔL = αΔT L,
stress-free) and a clamped bar at −EαΔT, capped at the yield stress with
perfect plasticity.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* clippy
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* thermal: trusses take the axial thermal strain; keep the roadmap's M7 line
Truss::thermal_load and stresses_with project the thermal strain on the
member's tangent (Σ αₖ tₖ² ΔT); the static solver applies them to rod2/rod3
(#38). Explicit trusses take none yet and a warning says so.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
---------
Co-authored-by: jyviko <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Model, fem and decks: point masses, rotary inertia, springs, dashpots and rod3 trusses
Model: PointMass and Spring definitions (spring | dashpot; to ground or
two-point; along a degree of freedom, a vector or the line of the ends) on a
set or reference point, resolved per node; ElementKind::Rod3. fem: rod2/rod3
truss kernel on Basis::Line2/Line3 declaring ElementDofs (translations).
Decks: Abaqus T3D3, MASS, ROTARYI, SPRING1/2/A, DASHPOT1/2/A; Nastran CONM1,
CELAS2, CDAMP2; LS-DYNA ELEMENT_MASS/INERTIA and ELEMENT_DISCRETE; what a
format cannot take is reported. Commands model.mass.* and model.spring.*.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* Implicit solver: trusses, point masses, rotary inertia and springs on the DofLayout
Trusses declare their translations like solids; point masses, inertia and
springs ask the layout for the kinds they need at their nodes (translations,
rotations), and reference points they use get six degrees of freedom. The
matrix pattern takes rows for those points; dashpots are ignored with a
warning. Existing results are bit-identical (hash-checked).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* Explicit solver: trusses, point masses, springs and dashpots on the node-major layout
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* docs: fem.md on trusses and discrete elements in the layout; regenerate features.md
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* docs: trusses in plane models and corotational explicit trusses
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
---------
Co-authored-by: jyviko <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
- model: materials gain solver-neutral plastic (bilinear isotropic/kinematic,
hardening table, Johnson-Cook with adiabatic heating), damage (Johnson-Cook
D1-D5) and erosion (plastic strain); explicit steps gain nonlinear_geometry
(default on) and stress_rate (jaumann | green_naghdi).
- fem: rotation (Hughes-Winget, Flanagan-Taylor), plasticity (radial return,
capped bracketed Newton for Johnson-Cook, damage; the update takes the
mechanical strain increment so thermal strains can be subtracted), closed-form
hex8 gradients for deformed configurations.
- explicit: updated Lagrangian hex8/tet4, per-element material state, erosion
with reported energy, plastic work in the energy account.
- solver: warns that the linear solver uses the elastic constants of plastic
materials (results hash-identical).
- decks: Abaqus *PLASTIC (+ Johnson-Cook keywords), Nastran MATS1/TABLES1,
LS-DYNA *MAT_PLASTIC_KINEMATIC / *MAT_PIECEWISE_LINEAR_PLASTICITY; what a
format cannot express is reported as Unsupported.
- engine: model.material.plasticity, model.material.damage (catalogue ready),
explicit element probe, eroded elements in the solve summary.
Step commits before the rebase onto the DOF framework: branch explicit-x2-steps.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
Model: planar section kinds with Extent, element edges in sets, edge-weighted 2D loads,
centrifugal load, 2D model checks. fem: Plane kernels on Basis::{Tri3,Tri6,Quad4,Quad8}
declaring ElementDofs::new(DofSet::PLANE); QM6 incompatible modes with the
Wilson-Ibrahimbegovic correction. Solver: Kernel enum on the DofLayout; 3D results
bit-identical (hash-checked). Decks: Abaqus CPS/CPE/CAX, Nastran membranes and RFORCE,
LS-DYNA ELFORM 12/13. Mesh: free triangle mesher, planar region meshing, face boundaries.
Engine: mesh.planar with tagged edges. Analytic verification suite in solver/tests/plane.rs.
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
Co-authored-by: jyviko <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Model: Newtonian fluid materials; fluid domains (box, spacing, conditions on
the box faces, obstacle bodies with conditions on their named faces, initial
velocity, body acceleration); fluid steps with caps. Deck writers refuse
fluid steps; a fluid-only model needs no mesh in model.check; one step
selector (Model::select_step) for every solver.
New L2 crate modelcraft-lbm: D3Q19 two-relaxation-time collision with Guo
forcing; halfway bounce-back walls (moving walls after Ladd), slip faces,
periodic pairs; Zou–He velocity and pressure faces with the D3Q19
transverse correction; obstacles voxelised from the face-tagged boundary
surface with exact predicates and symbolic perturbation; momentum-exchange
forces per box face and per B-rep face; unit conversion, automatic time step
and Mach/tau limits; bounded runs with frames and force histories.
Engine: model.fluid_domain.create/edit/delete, model.fluid_domain.boundary,
solve.fluid (obstacles from the open design or imported surfaces, lowerable
caps), solve.probe for fluid fields and wall force histories. Fluid Domain,
Inlet & Outlet and Lattice Boltzmann toolbar entries are ready.
Verification (analytic): plane Poiseuille exact with Λ = 3/16 and second
order (2.000) with Λ = 1/4; Couette exact with the wall shear; Taylor–Green
decay second order (2.01, 2.00); staircase pipe; Stokes drag of a sphere
array within 1.7% of Hasimoto at R = 6; Zou–He channel; mass conserved to
round-off; bit-for-bit reproducibility. Benchmark: Schäfer–Turek 2D-1.
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
Co-authored-by: jyviko <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
* ui.capture: the whole window drawn offscreen and rasterised on the CPU
ui.screenshot falls back to it when no frame is presented (hidden window,
locked screen), so agents can look at panels and the toolbar.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* Draw the open design and imported surfaces in the viewport
DisplayMesh keeps node labels on every element primitive (for results),
gains add_surface (geometry faces and their boundaries, coincident points
welded) and displaced (elements moved by nodal offsets). Bodies are
tessellated at a thousandth of their size; meshed models show the geometry
as boundaries only.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* results.* commands: deformed shape, contours, mode shapes, animation, reaction sums
The results display is session state (ResultsView), set by commands that are
not edits. Nodal fields (displacement, von Mises, principal and component
stresses; principal stresses in closed form after Smith 1961) and reaction
force and moment sums per set. Results kept from an earlier revision stay
viewable, marked stale; solve.probe says when a result is stale. The Results
tab's display and reaction entries are ready.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* Results display: deformed shape, contours with legend, mode shapes and animation, probe, reaction sums
The viewport draws the session's results view: elements moved by their nodes'
displacements or a mode shape (automatic or given scale), faces split exactly
along contour band boundaries into flat-coloured pieces (same on GPU and CPU),
a legend with levels and extremes, a results panel whose controls run the
results.* commands, a click probing the nearest node (solve.probe), and stale
results marked. Solving from the form shows the result on the Results tab.
sample::cantilever is a solvable model for tests and demos.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* Simplify the results display (review)
One Contour in the plot; the results scene reuses scene::build; a solve shows
its result through the engine (ResultsView::for_result) and the UI raises the
Results tab when a view appears; the plot's scale and contour values are kept
across animation frames; design bodies are tessellated once per design; the
first content is framed where the display is rebuilt; field_values in one
match; probe reuses results::displacements.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* Results display for explicit steps: field frames, energy histories
The results view works on the engine's step results (static, frequency,
explicit): ResultsView.frame and results.frame {frame | time} pick an explicit
field frame; the deformed shape (true scale) and contours (element stresses
averaged to the nodes, velocity) follow it; animation runs through the frames.
The Results panel has a frame slider and plots the kinetic, internal, hourglass
and external energies with their values at the frame. sample::impact is a
steel bar flying onto a rigid wall.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* docs/features.md after the rebase
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
---------
Co-authored-by: jyviko <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
* fem: wedge6 element (triangle × interval, 3 × 2 points), patch and cantilever tests
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* solver: reference points, kinematic and distributing couplings, moments by constraint elimination
Coupled reference points get six degrees of freedom; kinematic couplings are rigid
links, distributing couplings least-squares interpolation constraints. Every
constraint row and prescribed value is eliminated master-slave (Abel-Shephard
pivot choice), so the reduced matrix stays SPD; no coupling gives the old
elimination bit for bit. Verified: rigid motion of a coupled block, clamping
through a point, distributing tension vs pressure, bending and tube torsion.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* engine: reference point results in solve.static and solve.probe
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* docs: couplings, moments and wedge6 in the solver docs, literature and roadmap
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* solver: simplify constraint elimination and reference point plumbing
Rows combine by one sort; T_f transposed once; K and M reduced in one pass on a
shared pattern; prescribed motion computed once; reference point loads and BCs go
through Resolved like node ones (Mesh::label_dofs); one dof-direction rule;
StepResult::locate for probes; dead Csr::restrict removed.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* solver: refuse oversized reductions before allocating them
A held distributing coupling point makes a dense (3N)^2 block; estimate it from
the constraint rows and refuse it against the factorization's memory budget
with a clear message, before anything is built. Other reductions with
dependents count their entries first (stopping at the budget) and reserve
fallibly.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
---------
Co-authored-by: jyviko <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Renumber the layer table for solver crates: fem at L1, solver at L2, fsi at L3
L0 model -> L1 geometry, mesh, fem, decks, render -> L2 solver (explicit, lbm later)
-> L3 fsi (later) -> L4 doc -> L5 io -> L6 engine -> L7 ui-egui, mcp -> apps.
UI_MIN_LAYER moves to 7, the testkit sits above it.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* modelcraft-fem: element kernels for tet4, tet10, hex8 (full, incompatible modes), hex20 (full, reduced)
Shape functions, Gauss and conical-product quadrature, stiffness, consistent and HRZ-lumped
mass, body forces, consistent pressure loads on faces, integration-point stresses and their
least-squares extrapolation to nodes, isotropic and orthotropic elasticity. The element table is
one row per kind, so wedges, shells and beams are added locally.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* modelcraft-solver: linear static and frequency analysis with an analytic verification suite
Assembly from the resolved model on the node graph, boundary conditions by elimination,
consistent pressure, gravity and body-force loads, nested-dissection ordering, sparse up-looking
Cholesky with a factor-size cap, reactions, integration-point and nodal-averaged stresses, von
Mises; shift-invert block Lanczos with full reorthogonalization for natural modes, participation
and effective mass. ADR 0001 records the own-linear-algebra decision (determinism, wasm).
Verified against closed forms: displacement and traction patch tests for every element, cantilever
tip deflection (Timoshenko) with convergence rates, Lamé cylinder, Kirsch hole (Kt -> 3),
cantilever and free-free beam frequencies with six rigid-body modes.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* Engine: solve.static, solve.frequency and solve.probe; results kept in the session
Results are stored per step, restored on a failed command and cleared when the model changes.
Caps (degrees of freedom, factor entries, Lanczos vectors) may only be lowered; hostile parameters
are tested against a solvable session.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* docs: implicit solver architecture, roadmap M4, parity and references
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* Simplify the solver: shared face integration, block assembly, cached extrapolation
- model: one face integrator (nodal::face_normal_weights) for tractions and pressures
- fem: mid-side tables from ElementKind::edges, orthotropic via inverse3, symmetric
accumulation without a symmetrize pass, incompatible-mode recovery without K_uu, model's
coordinate check, dead table rows removed
- solver: one pattern search per node pair (block, scalar and diagonal adds), extrapolation built
once per element type, reactions from the prescribed rows only, matrices dropped after
restriction, the free-body shift keyed on the factorization error, step selection in the solver,
Error::Element naming the element
- engine: the solver owns the default step and the shift check; the model comparison runs once
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
* docs: results are stamped with the session revision
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
---------
Co-authored-by: jyviko <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Owner decisions 2026-10-13: own solvers beyond linear (explicit dynamics with equations of state,
lattice Boltzmann CFD, fluid-structure interaction; blast fluids after), flow first; published
reference values allowed as a tagged benchmark suite beside the analytic gate; GPU backends
deterministic per device and checked against the bit-exact CPU path; layered crates (fem kernels,
solvers, fsi). ADR 0001 lists the GPL-family solver codes (Palabos, OpenFOAM, ...) and the
readable references.
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm
Co-authored-by: jyviko <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
