Interpretation report
Fit metrics appear after calculation.
Field curve and model
Log–log axes. Orange dots: measurements. Teal: candidate Wenner response. Dashed horizontal lines: model layer resistivities, not measured asymptotes. Plots autoscale independently; use the normalised plot for matching, not a physical overlay of printouts.
Normalised Wenner master curves
x = a/h1; y = ρa/ρ1. Each grey curve is a calculated ratio ρ2/ρ1. Teal is the candidate family; orange points are your shifted field data. The h1 and ρ1 match axes are at x = 1 and y = 1.
Displayed families: 0.01, 0.03, 0.1, 0.3, 1, 3, 10, 30, 100. Curves are calculated, not copied from the scanned graph.
Measurement-by-measurement review
| a (m) | Measured ρa (Ω·m) | Model ρa (Ω·m) | Residual (%) |
|---|---|---|---|
| No calculated results. | |||
Residual = 100 × (model / measured − 1). A positive value means the model is higher. No readings are silently dropped or automatically smoothed.
Method, source mapping and limitations
The supplied SANS 10199:2016 Edition 2.2 extract, clause 4.4 (printed page 9), describes shifting log–log field curves against theoretical curves to obtain ρ1, h1 and the ratio ρ2/ρ1. This tool implements that two-layer interpretation concept numerically for Wenner data only.
What the three attachments contribute
- INTERPRETATION OF A SOUNDING GRAPHS-SANS10199(1).pdf: clause 4.4 gives the matching method; clause 4.2.3.5 warns that lateral inhomogeneity can distort Wenner interpretation.
- INTERPRETATION OF A TWO LAYER FIELD CURVE-SANS10199(1).pdf: Figure F.3, printed page 29, illustrates ρ1 = 300 Ω·m, h1 = 2.3 m and ratio 0.09, giving ρ2 = 27 Ω·m. Its horizontal axis is AB/2. The synthetic demo uses these parameters in a Wenner model; it is not a digitisation or validation of the drawn field curve.
- TWO LAYER WENNER GRAPHS(1).pdf: Figure F.5, printed page 31, shows increasing and decreasing Wenner curve families against adjacent spacing a. The tool supports finite positive ratios 0.01–100, not perfect conductors or insulators.
Forward equation and optimisation
For equal-spaced surface point electrodes, k = (ρ2 − ρ1)/(ρ2 + ρ1) and
ρa(a) = ρ1 × {1 + 4 Σ[n=1…∞] kⁿ × [1/√(1+(2nh1/a)²) − 1/√(4+(2nh1/a)²)]}.
The image-source series is evaluated with a conservative absolute tail bound below 10⁻¹⁰ in ρa/ρ1. The difference of reciprocal roots is evaluated in a numerically stable form. This equation is the numerical forward model; it is not printed in the three supplied extracts.
Raw shallow-electrode resistance is converted as ρa ≈ 2πaR. Finite-electrode modelling is not included. Fitting minimises the mean squared natural-log residual. A 31 × 31 log-spaced ratio/thickness grid is searched; the ten best seeds receive bounded pattern-search refinement. For each ratio/thickness pair, the optimum log ρ1 is solved analytically. The result is the best candidate found, not a guaranteed global optimum.
Near-fit ranges use sampled models with mean squared log residual ≤ best value + ln(1.05)². They are sensitivity indicators, not confidence intervals; sparse sampling can miss other equivalent models. Review the thickness bounds.
Engineering use and unsupported cases
Use one sounding centre and one traverse per assessment. Constant-spacing readings collected at different locations describe lateral variation, not a vertical sounding. Electrode spacing is not an exact investigation depth. A numerical fit cannot resolve additional layers, lateral variation, steep topography, anisotropy or interference from buried metal.
The tool does not assess corrosivity, calculate an installed earth-system resistance, GPR, step/touch voltage, lightning risk or certify compliance. Confirm applicable standards and obtain independent engineering review before using an interpreted soil model in a design. The provided excerpts are not a complete standard.
Quality triggers (four unique points, spacing ratio ≥ 3, b/a ≤ 0.05, 10% relative RMS, etc.) are transparent tool rules, not SANS acceptance criteria. Missing asymptotes and large near-fit ranges should lead to additional measurements.
Supporting references and verification
US EPA — Resistivity Methods explains array-specific curves, normalised curve matching and ambiguity. AGI — Wenner Array describes equal electrode intervals and a fixed sounding centre. Neither link establishes SANS compliance of this software.
Release verification: 12 numerical test groups passed, including homogeneous-earth response, upper/lower limits, monotonic curves, spacing conversions, invalid-input rejection and recovery of five synthetic models. The forward series agreed with independent Bessel-integral quadrature in 30 cases (worst relative difference approximately 2.6 × 10⁻¹¹). Seven control-logic cases passed in a minimal DOM test harness, including stale-result clearing, manual evaluation, depth rejection and unresolved homogeneous thickness. These tests are not browser rendering tests.
Browser layout, downloads and PDF printing have not been end-to-end verified: the test browser could not be downloaded in the authoring environment. Test these in your browser before operational use. No independent field-data validation or SANS conformity assessment has been completed.
Release 1.0.0. Source graph readings checked visually. No raw client data, calibration certificates or completed field assessments are embedded. Review of numerical software does not replace independent validation on traceable field data.