WP4: sonic anisotropy (gamma from Stoneley and dipole logs) by ingehap · Pull Request #5 · ingehap/ES_PHYS_P011

ingehap · 2026-10-14T11:41:46Z

This PR implements work package 4 of docs/IMPLEMENTATION_PLAN.md: Thomsen's γ in a vertical well, from the dipole C44 and the Stoneley-derived C66 (Walsh et al., 2007). It is validated on the synthetic Wisting-like well, where the true γ is known.

What is added

iwt.sonic
- Tube-wave relation with a concentric tool, and its inverse. Following Norris & Sinha (1993), the tube wave senses an effective shear modulus, which is C66 for a vertical well in a VTI formation. The tool term comes from the annulus compliance (Norris, 1990).
- Exact Stoneley mode of a fluid-filled borehole along the axis of a TI formation, with a tool. It uses the formation fields of White & Tongtaow (1981) and handles leaky modes (complex slowness, outgoing radiation). It is solved for a whole log at once: 2400 depths take about 1.5 s.
- Dispersion-corrected C66 inversion. The C11 and C13 that a vertical well does not measure come from assumed Thomsen ε and δ. The isotropic shortcut is available for comparison.
- Dipole C44 matched to the Stoneley wave. C44 is moved to the Stoneley frequency (constant Q) and averaged over the Stoneley aperture before γ is formed.
- Uncertainty: the exact Jacobian of γ, with linear propagation.
- Mud-slowness calibration in a reference interval.
iwt.synthetic: a caliper log with washouts, and a Stoneley log computed with the exact mode in the Backus average of the model over the array aperture.
examples/wp4_sonic_anisotropy.py (about 50 s) and docs/figures/wp4_*.png.
README and plan: updated, with as-built notes and verified references. Version 0.4.0.

Results on the synthetic well

Check	Result
Exact mode vs tube-wave limit (with tool) and Scholte limit	match to 1e-6 and 1e-4
Round trip: forward exact mode → inversion	C66 recovered to 1e-7
Stoneley wave in the overburden shales	leaky: the C66 tube wave outruns the C44 axial shear wave. Radiation 1/Q ≈ 0.01 at 500 Hz; slowness 1.9 % above the tube wave
Isotropic substitute (μ = C66) vs TI with ε = δ = 0, at 500 Hz	slowness error 1.1 % vs 0.2 %; the measured C44 matters more than ε and δ
γ with all corrections and the true mud slowness	every unit within 0.014 of the truth
Effect of a 2.7 % mud-slowness error	+0.05 to +0.10 in γ, the dominant error, as the plan expected
Mud calibrated in the Stø assumed isotropic (true γ = 0.023)	mud velocity 1563 against 1540 m/s; γ errors −0.01 to −0.04
σ(γ) from the budget (2 % mud uncertainty)	0.033–0.055; Monte Carlo agrees to 0.001–0.003

Departures from the plan

Eccentricity is not modelled. Norris (1990) reports that eccentricity lowers the tube-wave speed. The plan had listed an eccentricity correction.
Added beyond the plan:
- the exact (leaky) Stoneley mode;
- matching the dipole's frequency and resolution to the Stoneley wave;
- the mud calibration.
Ambiguity near the leaky–trapped transition. There, the measured real slowness can be matched by two C66 values (1 random sample in 2400). The Stoneley attenuation would resolve it. This is documented.

Tests

18 new tests, 111 in all, passing on Python 3.13 with RuntimeWarning and DeprecationWarning raised as errors.
On Python 3.9 with numpy 1.21 and scipy 1.7: 107 pass and 1 module is skipped, because matplotlib is not installed there.
pyflakes is clean.

🤖 Generated with Claude Code

https://claude.ai/code/session_019vaUDUPicoFeQaUhfhnoYq

Generated by Claude Code

- iwt.sonic: tube-wave relation with a concentric tool and its inverse (C66 as the effective modulus, Norris & Sinha 1993); the exact Stoneley mode of a fluid-filled borehole along the axis of a TI formation, with a tool, including leaky modes on the outgoing sheet, solved for whole logs at once; dispersion-corrected C66 inversion (TI with assumed eps, delta, or the isotropic shortcut); dipole C44 moved to the Stoneley frequency and averaged to its aperture; exact gamma Jacobian and linear uncertainty propagation; mud-slowness calibration in a reference interval. - iwt.synthetic: caliper log with washouts and a Stoneley slowness log from the exact mode in the Backus medium of the array aperture. - examples/wp4_sonic_anisotropy.py and docs/figures/wp4_*.png; README and plan updated (WP4 done, as-built notes, references); version 0.4.0. - 18 new tests (111 in all). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_019vaUDUPicoFeQaUhfhnoYq

Copilot

🟡 Changes recommended

The leaky-mode determinant scaling and validation workflow contain correctness issues, and the uncertainty calculation omits dispersion dependence.

4 open findings

What changed in this PR

Implements WP4 sonic anisotropy, deriving Thomsen’s γ from dipole and Stoneley logs.

Changes:

Adds tube-wave and exact TI Stoneley modelling and inversion.
Adds synthetic caliper/Stoneley logs, uncertainty propagation, and validation.
Updates documentation and releases version 0.4.0.

File	Description
`iwt/sonic.py`	Adds sonic anisotropy modelling and inversion.
`iwt/synthetic.py`	Adds caliper and Stoneley logs.
`tests/test_sonic.py`	Adds WP4 tests.
`examples/wp4_sonic_anisotropy.py`	Demonstrates processing and uncertainty analysis.
`README.md`	Documents WP4 usage and results.
`docs/IMPLEMENTATION_PLAN.md`	Marks WP4 complete.
`iwt/__init__.py`	Exports the sonic module and version.
`pyproject.toml`	Bumps version to 0.4.0.

🧠 Review effort: Balanced

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

+        ea = np.exp(-f * (b - a))
+        c = 0.0 if np.isinf(kt) else pf * a / (2.0 * kt)
+        M = np.array([
+            [f * ive(1, f * b), -f * kve(1, f * b) * ea, col[0][0], col[1][0]],
+            [-pf * ive(0, f * b) / sc, -pf * kve(0, f * b) * ea / sc, col[0][1], col[1][1]],
+            [zero, zero, col[0][2], col[1][2]],
+            [(f * ive(1, f * a) + c * ive(0, f * a)) * ea,
+             -f * kve(1, f * a) + c * kve(0, f * a), zero, zero],
+        ])


+    c44_m = S.running_average(z, logs.rho * logs.vs ** 2, APERTURE, harmonic=True)
+    rho_m = S.running_average(z, logs.rho, APERTURE)
+    vs_m = np.sqrt(c44_m / rho_m)
+    qs = 0.5 * model.qp[model.layer_index(z)]
+    f_dip = logs.meta["f_sonic_s"]
+    base = (logs.vp, logs.vs, logs.rho, st.slowness)
+    matched = (logs.vp, vs_m, rho_m, st.slowness)


+    s_tube = st.slowness / final.dispersion
+    J = S.gamma_jacobian(s_tube, 1.0 / vs_m, rho_m, mud_cal.slowness, mud_cal.density, caliper,
+                         TOOL.tool_radius, 0.0)


+- **Tool term:** built as planned. **Eccentricity is not modelled.** Norris (1990) reports
+  that it lowers the tube-wave speed, but the size of the effect was not quantified here.


ingehap marked this pull request as ready for review October 14, 2026 14:13

Copilot AI balanced review requested due to automatic review settings October 14, 2026 14:13

ingehap merged commit cd67dd7 into main Oct 14, 2026
4 checks passed

Copilot started reviewing on behalf of ingehap October 14, 2026 14:13

View session

Copilot AI reviewed Oct 14, 2026

View reviewed changes

Navigation Menu

Repository navigation

WP4: sonic anisotropy (gamma from Stoneley and dipole logs) - #5

WP4: sonic anisotropy (gamma from Stoneley and dipole logs)#5
ingehap merged 1 commit into
mainfrom
claude/blissful-gauss-3y1u5w

ingehap commented Oct 14, 2026

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

		- Tool term: built as planned. Eccentricity is not modelled. Norris (1990) reports
		that it lowers the tube-wave speed, but the size of the effect was not quantified here.

Conversation

ingehap commented Oct 14, 2026

What is added

Results on the synthetic well

Departures from the plan

Tests

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🟡 Changes recommended

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants