ashfordeOU Eddy Current Inspection (manufacturing-quality/ndt/eddy-current-inspection)
Use when the task is executing eddy current inspection (ET) on a part: computing the standard depth of penetration from frequency, conductivity, and permeability, selecting the test frequency that keeps a surface or subsurface flaw within the usable penetration band, and interpreting the probe response on the impedance plane.
Domain quick reference
- Standard depth of penetration: delta = 1 / sqrt(pi * f * mu * sigma), with frequency f in Hz, mu = mu0 * mu_r in H/m (mu0 = 4 * pi * 1e-7), and conductivity sigma in S/m; delta is in meters. At one standard depth of penetration the eddy current density falls to 1/e, about 37 percent of its surface value. Higher frequency, higher conductivity, and higher permeability all shrink delta.
- Percent IACS to conductivity: 100 percent IACS equals the annealed copper reference of 5.8e7 S/m, so sigma = (percent / 100) * 5.8e7. Aluminum alloys sit near 30 percent IACS (1.74e7 S/m) and titanium alloys near 5 percent IACS.
- Frequency selection: choose the frequency so the flaw depth sits within one standard depth of penetration, where the current density is still strong. For a subsurface flaw at depth d, use a frequency whose delta is at least 1 to 2 times d (factor 2.0 is common), which keeps the current density at the flaw above 37 percent. For a surface crack use a smaller delta (factor below 1.0), concentrating the current near the surface for sharp response.
- Eddy current density ratio: J / J0 = exp(-depth / delta). At one delta the ratio is 0.368, at two deltas it is 0.135.
- Phase lag: the eddy current lags the surface current by depth / delta radians, so phase_lag_degrees = (depth / delta) * 180 / pi. At one standard depth of penetration the lag is one radian, about 57.3 degrees.
- Impedance plane: the probe coil response plots normalized reactance against normalized resistance. Lift-off moves the point along the lift-off trajectory, a conductivity change moves it along a curve toward the material point, and a crack rotates the trajectory with a characteristic phase angle. Reading which trajectory the point follows separates the indications.
- Skin effect: eddy currents concentrate at the surface and decay exponentially with depth; the standard depth of penetration is the characteristic length of that decay.
- Acceptance: indications are compared with reference standards and the engineering specification, recorded, and dispositioned under special-process control by NAS 410 qualified personnel.
Workflow
- Establish the material: its electrical conductivity in S/m (convert from percent IACS with conductivity_from_iacs) and its relative magnetic permeability (1.0 for aluminum, titanium, and austenitic steels; above 1 for ferromagnetic materials).
- Pick the inspection frequency: for a subsurface flaw compute select_frequency_for_flaw with a penetration factor of 2.0 or more; for a surface crack use a factor below 1.0.
- Compute the standard depth of penetration with standard_depth_of_penetration and confirm the flaw depth sits within one delta, or compute the frequency for a given depth directly with frequency_for_depth.
- Estimate the current density at the flaw depth with eddy_current_density_ratio and the phase lag with phase_lag_degrees; both support the impedance-plane reading.
- Interpret the probe response on the impedance plane: identify the lift-off trajectory, the conductivity trajectory, and the crack trajectory, and separate the crack indication from lift-off and conductivity drift.
- Compare the indication with the acceptance criteria in the engineering specification, record the results, and disposition the part.
Pitfalls
- Forgetting the mu0 factor: mu in the depth of penetration formula is mu0 * mu_r, not mu_r; skipping mu0 (4 * pi * 1e-7) makes delta come out far too small.
- Using percent IACS directly as S/m: 30 percent IACS is 1.74e7 S/m, not 30 S/m; always convert with the 5.8e7 reference.
- Choosing a frequency that buries the flaw: a subsurface flaw beyond one delta sees a density ratio below 0.37 and may be missed; lower the frequency or raise the penetration factor.
- Using one frequency for surface and subsurface flaws: the same setup cannot serve both; surface cracks need a small delta, subsurface flaws need delta at least 1 to 2 times the flaw depth.
- Reading lift-off as a crack: on the impedance plane lift-off follows its own trajectory and is a routine part of the scan, not a defect; separate the trajectories before dispositioning.
- Ignoring permeability: a ferromagnetic part with mu_r above 1 shrinks delta sharply, so a frequency that works on aluminum gives a shallow skin effect on steel.
- Treating the phase lag as a fixed 90 degrees: the lag is depth / delta in radians, one radian at one delta, and it is the angle that sizes the depth reading on the impedance plane.
- Skipping calibration and personnel qualification: ET results are dispositioned under special-process control with NAS 410 qualified personnel, and the setup is verified on reference standards before the scan.
Behavior contract (gate 3)
The inspection math is exercised by the gate 3 contract test: scripts/test_eddy_current_inspection.py against scripts/eddy_current_inspection_logic.py (stdlib unittest, offline). Run: python3 scripts/test_eddy_current_inspection.py
Compliance
- Standards referenced, not reproduced: AS9100 clause 8.5.1.3 frames NDT as a special process requiring controlled procedures and qualified personnel, and NAS 410 sets the qualification and certification requirements for the NDT personnel who execute eddy current inspection; the formulas and impedance-plane practice above are common ET methodology, summary-only per standards-map.yaml.
- compliance: STANDARDS-REF, gated: false.