Validation gallery
Reproducible 5-check internal-consistency verify pass for 7 canonical seeded cases spanning oil-unconventional, dry-gas, EGS, CBM, and offset parent-child regimes. Each row runs against its own stored inputs_hash fingerprint — any drift in the seed blocks the pass. Real-field blind matches live under /trust/validation-gallery.
One-click export of the 5 bundled public cases (HFTS-1 Wolfcamp, MSEEL MIP-3H, Bakken depleted offset, Barnett DFIT, Eagle Ford datum) with citations, DOIs, licenses, and the reconciliation snapshot fromevaluateClosureVsLayerStress. Both formats embed a run-fingerprint triple (solver build · inputs hash · results hash).
Compares the current bundled inputs and freshly-evaluated results against the FNV-1a-64 snapshot committed alongside the drift-guard test. Any mismatch expands a leaf-level diff so you can see which field changed.
Baseline case exercised throughout the engineer's Order #2 report. Oil-window unconventional, 40-stage lateral, mid-range treating.
- Fluid mass balanceΔ = 0.0 lb (tol 0.10% rel)
- Proppant mass balanceΔ = 0.0 lb (tol 0.10% rel)
- PKN half-lengthΔ = 0.0 ft (tol ±5 ft)
- Nolte-Smith modemode I vs expected I
- Screen-out marginΔ = 0.000 (tol ±0.02)
results_hash c6332299e568656f · build dev
Middle Bakken preset — 10,000 ft lateral, 880 ft parent spacing, PKN half-length ~380 ft, slickwater-dominant.
- Fluid mass balanceΔ = 0.0 lb (tol 0.10% rel)
- Proppant mass balanceΔ = 0.0 lb (tol 0.10% rel)
- PKN half-lengthΔ = 0.0 ft (tol ±5 ft)
- Nolte-Smith modemode I vs expected I
- Screen-out marginΔ = 0.000 (tol ±0.02)
results_hash 38f18fd1bdcedf67 · build dev
Eagle Ford preset — 7,500 ft lateral, 500 ft spacing, higher proppant intensity, PKN half-length ~260 ft.
- Fluid mass balanceΔ = 0.0 lb (tol 0.10% rel)
- Proppant mass balanceΔ = 0.0 lb (tol 0.10% rel)
- PKN half-lengthΔ = 0.0 ft (tol ±5 ft)
- Nolte-Smith modemode II vs expected II
- Screen-out marginΔ = 0.000 (tol ±0.02)
results_hash 4bb63bfbd191edc5 · build dev
Marcellus dry-gas preset — 8,500 ft lateral, 1,000 ft spacing, longer PKN half-length ~450 ft, Mode I dominant.
- Fluid mass balanceΔ = 0.0 lb (tol 0.10% rel)
- Proppant mass balanceΔ = 0.0 lb (tol 0.10% rel)
- PKN half-lengthΔ = 0.0 ft (tol ±5 ft)
- Nolte-Smith modemode I vs expected I
- Screen-out marginΔ = 0.000 (tol ±0.02)
results_hash 5dd739df814f0505 · build dev
EGS reference — proppant-free cold-water injection, KGD-dominant geometry, screen-out check is vestigial.
- Fluid mass balanceΔ = 0.0 lb (tol 0.10% rel)
- Proppant mass balanceΔ = 0.0 lb (tol 0.10% rel)
- KGD half-lengthΔ = 0.0 ft (tol ±5 ft)
- Nolte-Smith modemode III vs expected III
- Screen-out marginΔ = 0.000 (tol ±0.02)
results_hash 3d6b483ec4d03a5c · build dev
Coal-bed methane — low peak rate, KGD-thin geometry, moderate screen-out margin from soft-coal fines.
- Fluid mass balanceΔ = 0.0 lb (tol 0.10% rel)
- Proppant mass balanceΔ = 0.0 lb (tol 0.10% rel)
- KGD half-lengthΔ = 0.0 ft (tol ±5 ft)
- Nolte-Smith modemode II vs expected II
- Screen-out marginΔ = 0.000 (tol ±0.02)
results_hash 94d6716ade097d44 · build dev
Offset-parent-child stress-shadowed child well — PKN half-length trimmed by depletion asymmetry, elevated screen-out risk.
- Fluid mass balanceΔ = 0.0 lb (tol 0.10% rel)
- Proppant mass balanceΔ = 0.0 lb (tol 0.10% rel)
- PKN half-lengthΔ = 0.0 ft (tol ±5 ft)
- Nolte-Smith modemode IV vs expected IV
- Screen-out marginΔ = 0.000 (tol ±0.02)
results_hash 1df60c798977cc63 · build dev