← Back to the tool

A research prototype, with a question to test

Working title: Governing-layer transitions and resistance-information sensitivity in groundwater-dependent liquefaction screening.

Engineering question

When does the layer governing the onset of simplified liquefaction triggering change as groundwater varies, and can explicitly tracking that change improve investigation prioritization compared with a single nominal factor-of-safety profile?

What the prototype contributes

The interface combines bounded input scenarios, inverse threshold calculation, and idealized information sensitivity. These are proposed workflow diagnostics. Neither originality relative to all existing research nor practical superiority has been established. A new GUI and familiar correlations alone are insufficient for a research paper.

Relation to Dr. Dhakal’s work

Dr. Dhakal’s published work includes conventional CPT applicability, reclaimed soils, liquefaction settlement, and ejecta. This prototype addresses interpretation of conventional triggering results. It does not reproduce the 2025 settlement procedure or predict consequences. It is independent work with no claimed endorsement.

Hypotheses

  1. The layer with the lowest nominal FS at one groundwater depth need not dominate the onset envelope over a groundwater range.
  2. Resistance refinement in a non-governing layer can have negligible effect on the threshold envelope, while refinement in a layer controlling a subset of groundwater scenarios can matter appreciably.
  3. Ranking may change with the groundwater range, mesh, parameter bounds, and the selected engineering endpoint. That instability should be reported rather than hidden.

Study design

  1. Complete a targeted review of critical-layer selection, groundwater-sensitive liquefaction assessment, robust decision methods, and geotechnical value of information. Identify what existing methods already do.
  2. Benchmark the implemented normalized-CPT branch against independently verified software under matched assumptions. Add raw CPT normalization only with separate verification and reference investigation groundwater conditions.
  3. Use synthetic profiles to isolate thin lenses, interbeds, changing groundwater, and governing-layer transitions. Perform mesh, groundwater-grid, and bound-width convergence studies.
  4. Obtain permitted field profiles with repeat soundings and groundwater observations. Keep q uncertainty, spatial variability, groundwater variability, and triggering-model error conceptually distinct.
  5. Compare against nominal minimum-FS and thickness-based prioritization. A triggering-onset endpoint can overemphasize very thin layers, so evaluate consequence-relevant endpoints separately.
  6. Test investigation rankings using held-out soundings or withheld high-resolution information. Measure actual reduction in predictive uncertainty, decision reversals, and ranking stability.
  7. Before calling the metric value of information, introduce defensible joint distributions, imperfect observations, test costs, and a stated loss function. The current deterministic average contraction is not that calculation.
  8. Validate any settlement or ejecta extension against the original published equations and independent observations. Do not infer settlement from below-target thickness alone.

Current verification

Passed: hand-checkable total/effective stress balance, inverse threshold recovery, PGA scaling, dry-profile behavior, zero-bound collapse, input rejection, grid checks of resistance monotonicity, envelope ordering, and nonnegative idealized contraction. A separate Python transcription agrees with the JavaScript cell calculations. These are numerical checks, not field validation.

Limits that affect interpretation

Fixed normalized resistance, hydrostatic saturation, no capillary fringe, independent supplied surface PGA, level ground, no static shear, and no explicit probability distributions. The bounds are inputs chosen by the analyst. Ranking is a uniform groundwater-grid average, not an occurrence probability. Low-stress extrapolation and thin-layer discretization require particular care.

Source and development

Download browser source package. Unzip, open the folder in VS Code, run python -m http.server 8000, and visit http://localhost:8000. Run numerical checks with node test-liquefaction.cjs. No Node installation is needed to use the hosted app.

References

Boulanger and Idriss (2016), CPT-based liquefaction triggering procedure. Companion report: UCD/CGM-14/01 (2014).

Cubrinovski and Dhakal (2025), Simplified procedure for estimating liquefaction-induced settlement.

Dhakal and Cubrinovski (2025), Key considerations for a simplified evaluation of liquefaction-induced settlement.

Dhakal, Cubrinovski and Bray (2022), Evaluating the applicability of conventional CPT-based liquefaction assessment procedures to reclaimed gravelly soils.