Sixteen connected workflows bring original evidence, chronological timber tests, nonlinear research models and independent checks into one retained campaign.
Explicit OpenAlex refresh, retained metadata, DOI deduplication, two-stage reviewer decisions, adjudication and a downloadable review flow diagram.
Each refresh is bounded to 50 metadata records. Reviewer identities and obtained full-text review are declarations; a metadata match is not a scientific review.
Original located excerpts, physical specimen identities, declared covariance and separate within-laboratory and between-laboratory variance estimates.
Supported inference uses a declared-comparability, two-stage random-effects method. It does not supply missing experiments or propagate all variance-component uncertainty.
Full chronological source histories fit eight pinching/degradation parameters, with held-out predictions, force/energy/residual errors and joint identifiability diagnostics.
Specimen-cluster uncertainty is withheld when the independent sample or numerical rank is insufficient. A fit is not physical validation of the constitutive law.
A downloadable fixed-schema NI-DAQmx worker records original numeric channels, calibration and timing configuration, sustained logs and explicit gap/overflow diagnostics.
Live operation needs supported local devices, drivers and physical commissioning. Hardware commissioning has not been demonstrated by this website release.
Track original cropped image planes with subpixel subset correlation, displacement vectors, texture/ambiguity and forward/backward rejection, strain cells and virtual gauges. Reviewed camera matrices support a bounded stereo quartet.
Exact cropped numeric planes and original-file digests are retained. Image texture, reviewed calibration and physical camera commissioning are required; this is bounded subset tracking rather than general dense stereo reconstruction.
Layered compression-plasticity studies now include an explicit crack plane inside one orthotropic ply: opening, slip, mixed-mode cohesive damage, contact closure, accepted solid fields and real step/mesh comparisons.
The declared plane is fixed in a bounded coupon. Arbitrary crack growth, full rolling-shear failure and experimentally validated capacity remain outside this calculation.
One to six storeys use staged constant gravity, modal and nonlinear pushover/transient calculations, drift/energy histories, exact reviewed hysteresis laws and a halved-time-step check.
This is a shear-building research model with one horizontal degree of freedom per floor. Torsion, flexible diaphragms and full spatial building mechanics are not represented.
Conforming layered panels with rectangular openings, targeted X/Z refinement, tetrahedral quality and conservation checks, followed by actual numerical convergence runs.
The supported geometry is bounded. General CAD import and through-thickness refinement are not newly implemented.
A downloadable worker executes a fixed supported cantilever model, retaining the job, solver/version, logs and checked receipt. OpenSees execution has been exercised against analytical references.
The supported external case is bounded. Abaqus needs a licensed executable and remains uncommissioned; verifying an imported receipt does not execute a fresh solver run.
Import actual IFC4 STEP sources, inspect source entity IDs, explicit units and nested placements, rectangular extrusions, materials, layers, properties and supported openings. Export and reimport the supported subset with a field-by-field loss report.
The original source archive stays separate from normalized exchange. Unsupported geometry and missing solver semantics block equivalence claims; arbitrary IFC/BIM and CAD conversion remain outside the supported subset.
Exact original source closures now run through private server jobs with per-case checkpoints, revision-checked controls, interrupted-job recovery and independently replayed completion receipts. Two browser workers remain available for local calculations.
Server jobs support eight selected cases, 1 MiB of complete sources and 1 MiB of results, with a 1,900,000-byte combined storage budget and bounded 20-second request bursts. Unfinished work needs explicit recovery; continuous external cloud execution is not commissioned.
Select exact records and their dependency closure, add specimen assignments, and share a checked native team document with role-based membership, expiry, revisions and independent review.
The source-complete selected handoff is bounded to 60 KiB. Larger native campaigns use the separate portable package; no invitation email is sent automatically.
Complete original campaign records, job checkpoints and record-bound pressure strokes use local IndexedDB revisions, cross-tab conflict detection and verified portable reopening.
Complete offline envelopes support 6 MiB. Browser-profile access can expose local copies. Physical iPad/Pencil acceptance remains unverified.
Actual native usage and team-seat reporting, private in-app support with owner replies, and configured-provider subscription checkout and signed lifecycle records.
Payments remain inactive until the owner supplies provider credentials, prices and a verified webhook. Existing access grants and storage quotas remain authoritative.