Validation & method provenance

A pile-analysis tool is only as good as your trust in its numbers. New software earns that trust the way the profession has always demanded: by citing its method provenance and benchmarking against the codes engineers already rely on. Here is exactly how PileCalc is validated — and how you can check it yourself.

Method provenance

Every model in PileCalc traces to a named, published source — the same sources LPILE and RSPile cite. Nothing is a black box.

Governing equation: the beam-column on a nonlinear (Winkler) foundation, from COM624P (Wang & Reese, 1993, FHWA-SA-91-048). See the p-y method.
p-y curves: Matlock (1970); Reese, Cox & Koop (1974); Welch & Reese (1972); API / O'Neill & Murchison (1983); Reese (1997) weak rock.
Axial & shafts: NAVFAC DM-7.02 static methods; t-z / q-w load transfer; FHWA-IF-99-025 (O'Neill & Reese, 1999); Vesić (1977) settlement.
Footings: the general bearing-capacity equation with Meyerhof / Vesić shape, depth, and inclination factors.

Equation-level agreement

The RSPile Laterally Loaded Piles Theory Manual documents the governing equation and every p-y model. We checked the PileCalc formulation against it term by term — the Matlock pu expression, the y₅₀ = 2.5·ε₅₀·b relationship, the Reese sand wedge with its tan⁸ term, the API tanh form, the Reese–Nyman weak-rock reduction. The equation forms are identical. RSPile uses finite elements and PileCalc uses finite differences, but that is a discretization choice; the governing physics is the same.

One caveat: the empirical chart coefficients

A few Reese models read their coefficients from hand-drawn charts rather than closed-form equations — the cyclic-sand and sand B factors and the stiff-clay A factor. These were never published as numbers, so PileCalc digitizes them from the source figures; the labelled deep asymptotes are exact, but the intermediate points are chart reads that carry real uncertainty. The static-sand A array is sourced from a citable dataset. For design-critical work, supply site-specific p-y values rather than relying on the digitized intermediate points.

Numerical benchmarks, by tool

Every tool in PileCalc is anchored to two or three published cases with a known answer — a closed-form solution, a design-manual table, or an independent code (LPILE / RSPile). We run the engine on the exact same problem and compare. Every figure below is reproduced by the test suite (packages/core/test/validation-benchmarks.test.ts), so the numbers in this table are generated, not transcribed.

Lateral piles (p-y)

Closed-form + independent code

COM624P / LPILE p-y finite-difference method

Quantity	PileCalc	Reference	Agreement
Groundline deflection, long pile on constant subgrade (y₀ = 2Pβ/k)Hetényi (1946), Beams on Elastic Foundation — closed form	0.9999 ×	1.0000 × (exact)	within 0.01%
Maximum moment, same case (Mₘₐₓ = 0.3224·P/β)Hetényi (1946) — closed form	0.9998 ×	1.0000 × (exact)	within 0.02%
Head deflection, API-sand single layer (D 0.5 m, H 100 kN)RSPile 2018 / LPILE verification problem #1	7.33 mm	7.3 mm	within 0.5%
Max moment, elastic pile on linear subgrade (D 1 m, L 24.4 m)Liang et al. (2014) closed form / RSPile Verification 5	792 kN·m	≈ 800 kN·m	within 1%

Axial capacity

Published design table

NAVFAC DM-7 static method; Meyerhof Nq; Skempton Nc

Quantity	PileCalc	Reference	Agreement
Bearing factor Nq, φ = 32° (displacement pile)NAVFAC DM-7.02 Table 8-1 (Meyerhof)	29.10	29.1	exact
End bearing Qₚ = 9·cᵤ·Aₜᵢₚ (cᵤ 90 kPa, D 0.457 m)Das, Principles of Foundation Engineering 8e (2016), Ex. 9.7	132.9 kN	132.9 kN	exact
Deep end-bearing factor Nc (z/B > 4)Skempton (1951) / Das — limiting Nc	9.00	9.0	exact

Drilled shafts

Published design table

FHWA-IF-99-025 (O'Neill & Reese, 1999) α / β method

Quantity	PileCalc	Reference	Agreement
β depth function at the published cap depth (z ≈ 4.94 ft → 1.2)FHWA-IF-99-025 β = 1.5 − 0.135√z	1.199	1.20 (cap)	within 0.1%
Clay tip pressure Nc·Sᵤ (cᵤ 2000 psf, Nc = 9)FHWA-IF-99-025 end bearing	17,990 psf	18,000 psf	within 0.1%
Clay side resistance, α = 0.55 with FHWA exclusion zonesFHWA-IF-99-025 α-method (5 ft top / 1D base excluded)	331,750 lb	331,752 lb	exact

Shallow footings

Closed-form + independent code

General bearing-capacity equation, Vesić factors & shape/depth corrections

Quantity	PileCalc	Reference	Agreement
Bearing factors Nc / Nq / Nγ at φ = 30°Vesić (1973) / Das Table 3.3 (verified φ = 0–40°)	30.14 / 18.40 / 22.40	30.14 / 18.40 / 22.40	within 0.1%
Ultimate bearing pressure q_ult, square footing (B 2 m, Df 1.5 m, c′ 20, φ′ 25°)Das, Foundation Engineering 8e (2016), Example 3.2	1401 kPa	1373 kPa	within 2.0%

Moment–curvature

Closed-form / exact

Fiber integration of the section (elastic–plastic)

Quantity	PileCalc	Reference	Agreement
Solid-circle plastic shape factor Mp/My (= 32/6π)Closed-form section mechanics	1.696	1.698	within 0.1%
Pipe yield moment My = Fy·S (12.75 × 0.5 in, Fy 50 ksi)S = π(Dₒ⁴−Dᵢ⁴)/(32Dₒ) — closed form	2.836×10⁶ lb·in	2.836×10⁶ lb·in	exact
Pipe plastic moment Mp = Fy·Z, Z = (Dₒ³−Dᵢ³)/6Closed-form plastic modulus	3.756×10⁶ lb·in	3.754×10⁶ lb·in	within 0.1%

Pile groups (lateral)

Published design table

p-multiplier (row-shadowing) deduction factors

Quantity	PileCalc	Reference	Agreement
In-line (front) p-multiplier at 3D / 6D / 8D spacingAllPile Table 8-4 / FHWA-NHI-05-042	0.40 / 0.80 / 1.00	0.40 / 0.80 / 1.00	exact
Side-by-side p-multiplier at 1D / 2D / 3D spacingAllPile Table 8-5 / FHWA-NHI-05-042	0.30 / 0.60 / 1.00	0.30 / 0.60 / 1.00	exact

Slope stabilization

Independent code

Applied soil-displacement method (RSPile)

Quantity	PileCalc	Reference	Agreement
Displacement resolution at 5° slip (axial / lateral, 25 mm)RSPile multi-layer slope-stabilization example	2.18 / 24.91 mm	2.18 / 24.91 mm	exact
Lateral resistance at the 8 m slip surfaceRSPile / TZPile / LPILE	607 kN	582 kN	within 4%

Uplift anchors

Closed-form / exact

Grouted-anchor bond capacity; plate breakout (AllPile §8.4–8.5)

Quantity	PileCalc	Reference	Agreement
Grouted-anchor ultimate pullout π·D·Lb·Ca (100 kN/m, 12 m)FHWA-IF-99-015 (Sabatini et al., 1999) bond model	1200 kN	1200 kN	exact

How to read the table

The reference column draws on three kinds of evidence, strongest first:

Closed-form / exact — an analytical solution with no discretization (Hetényi's beam on an elastic foundation; plastic section moduli; the bearing-capacity factor formulas). Agreement here is a pure test of the solver, and it lands within a fraction of a percent.
Published design tables — values every engineer already uses: the NAVFAC Nq factors, the Vesić Nc/Nq/Nγ table, the FHWA β-method bounds, the p-multiplier deduction factors. PileCalc reproduces them to the printed precision.
Independent codes & textbook examples — LPILE, RSPile, and worked examples from Das. A few of these carry a chart-reading tolerance or a documented method convention (see the footing q_ult case, where a depth-factor convention accounts for the 2% offset).

An honest note on reading the references

Most published verification manuals present results graphically — the second curve is plotted over the first to show agreement, with no numbers table. The reference values for those cases are read off the charts and carry a few-percent reading tolerance. The criterion is “reproduces the reference curve,” and it does. Where a number is exact (a formula or a printed table), the agreement is exact too.

Two codes, not one

We deliberately benchmark against both LPILE and RSPile. They are independent implementations — different teams, different numerical methods, even opposite sign conventions for moment and soil reaction. If PileCalc matched only one, you could not rule out that it had inherited that program's idiosyncrasies. Matching both, plus closed-form solutions, is much stronger evidence that the solver is right. (It is also why sign conventions differ between tools — compare magnitudes and locations, not raw signs.)

What validation does & doesn't mean

Validation means the math is faithful to the published methods and reproduces accepted benchmarks. It does not remove engineering judgment. The p-y method itself has limits, soil parameters carry real uncertainty, and you remain responsible for reviewing results for design. PileCalc's job is to make every assumption and intermediate value visible so that review is actually possible — and then to get out of your way.

Reproduce it yourself

Every worked example in the app is a benchmark you can run. Open the lateral tool, leave the defaults, and you'll see the 7.33 mm API-sand case above compute live.