The Evidence Layer for Scientific AI

 Long Analysis Plan Copied

Analysis Plan: Wood Structure — Experimental + Computational Workflow

The supplied evidence supports a four-module plan: (1) quasi-static shear-wall performance modeling of prestressed CLT–concrete composites, (2) panel-scale validation of tensioned-wood (Tenso-Wood) systems, (3) timber-retrofit capacity/thermal simulation of masonry walls, and (4) cell-wall moisture-deformation analysis via in situ nanotomography. Each module pairs experimental data with FEM validation and Python-based analysis.

Module 1 — Prestressed CLT–Concrete Shear Walls

Three full-scale specimens (Spruce–Pine–Fir CLT, ρ ≈ 445 kg/m³, 15.2 mm strands, 23 vs 46 kN prestress) were tested under low-cycle reciprocating loading with ABAQUS models (20 mm mesh, tied contacts) reproducing skeleton curves within ~5% error. Doubling prestress raised stiffness ~1.11× and damping ~1.7×, while prestressed CLT–concrete walls carried 25.2–28.5% higher maximum load than comparable CLT–steel configurations Prestressed CLT-concrete shear wall study. Plan: reproduce hysteresis/skeleton curves in Python (numpy, pandas), fit bilinear backbone models, and run ABAQUS parameter sweeps on prestress (23–46 kN) and aspect ratio.

Module 2 — Tenso-Wood Panel System

Caribbean-pine column panels (P2) sustained 156.9–194.2 kN compression; beam panels B1–B3 failed at 11.7, 16.9, 22.5 kN in three-point bending; a full frame reached 63.7 kN with a ductility factor of 10.8 and ~108 mm lateral drift. ABAQUS/SAP2000 elastic models agreed within ~10% Tenso-Wood panel validation. Plan: parse load–displacement curves, compute secant stiffness and ductility ratios, and calibrate linear-elastic FEM inputs against the ~10% discrepancy noted.

Module 3 — Timber Retrofits for Masonry

Pushover FEM (Abaqus) showed CLT/LVL retrofits raised load capacity by 41–52% (rocking: 71 kN, +145% limited by anchors; shear: +52% load, +226% displacement), while strong-back/OSB mainly increased displacement (+265–300%). Externally bonded CLT/LVL cut U-values from 2.072 to 0.244/0.257 W·m⁻²·K⁻¹ Timber retrofit seismic-thermal simulation. Plan: replicate pushover curves in Abaqus/Python scripting, and validate the response-factor method (96-h triangular pulse, <5% deviation) as an in-situ U-value estimator.

Module 4 — Cell-Wall Moisture Behavior (In Silico / Image Analysis)

A windowless climate chamber (0–90% RH, 10–50 °C, ±0.2–0.3% RH stability) at DESY/P05 enabled in situ nanotomography of lignified cell walls; affine registration quantified hydro-adsorption swelling of ~7.6% (x), 7.5% (y), 0.2% (z), total volume ~15.9% Climate chamber nanotomography of wood tissue. Plan: implement TomoPy/gridrec + Hagemann phase-retrieval reconstruction and 3D affine registration (scikit-image, Amira workflow) to quantify anisotropic swelling across samples — note the demonstrated dataset is a single cell, so replicate across earlywood/latewood before generalizing.

Knowledge Graph of Workflow

Confidence: moderate — all quantitative values come directly from the supplied records; n = 3 walls, 6 panels, 1 demonstrated cell, so generalization is limited. FE models were elastic in places and gaps/slip were simplified; long-term creep, moisture cycling, and fire remain unaddressed in these studies. What would change conclusions: independent replication of wall tests at other prestress levels, and multi-cell tomography across moisture cycles.