ashfordeOU Ultrasonic Inspection (manufacturing-quality/ndt/ultrasonic-inspection)
Use when the task is executing ultrasonic inspection (UT) on a part: turning pulse-echo time-of-flight and echo amplitude measurements into discontinuity depth and size, and planning the transducer, frequency, and calibration that make those measurements valid.
Domain quick reference
- Pulse-echo principle: one transducer emits a burst and receives the echo; the sound travels to the reflector and back, so depth is half the product of time of flight and velocity: depth = tof * v / 2.
- Wave modes: longitudinal (compressional) waves are fastest and are used by normal probes; shear waves are slower and are generated by angle-beam probes through refraction. Typical velocities: steel longitudinal ~5920 m/s, steel shear ~3230 m/s, water ~1480 m/s.
- Wavelength: lambda = v / f. Higher frequency gives shorter wavelength, better resolution of small reflectors, and stronger attenuation, which shortens useful range. Aerospace inspection commonly uses 2.25 to 10 MHz.
- Near field: for a circular piston transducer the near-field length is N = D^2 / (4 * lambda). Inside the near field the beam has amplitude maxima and minima and has not yet diverged, so amplitude-based sizing there is unreliable. Beyond N the beam spreads with half-angle gamma where sin(gamma) = 1.22 * lambda / D.
- Calibration blocks: the IIW block carries reference reflectors (side-drilled holes) and radius surfaces for velocity, zero-point, and sensitivity calibration; the step wedge has thickness steps for depth and sweep-linearity checks. Calibrate velocity, zero point, and sensitivity before scanning.
- Reference reflectors: flat-bottom holes (FBH) give area-like echoes and are used with distance-amplitude correction (DAC) curves; side-drilled holes (SDH) give cylinder-like echoes. Echo amplitude is compared to the reference reflector level in decibels.
- Scanning patterns: raster or serpentine scans with index increments no larger than half the element width, with overlap so no volume is skipped; normal (longitudinal) and angle (shear) scans cover different defect orientations.
- Discontinuity depth and size: depth from time of flight; size from amplitude compared with the reference reflector, using the 6 dB (half amplitude) drop method to locate reflector edges.
- Dead zone: a near-surface region where the front-surface echo and the initial pulse mask small indications; its length is set by the pulse length and the transducer ringing.
- Acceptance criteria: defined by the engineering specification and the approved NDT procedure, usually as echo amplitude relative to a reference reflector at the same depth; results are recorded and dispositioned under special-process control.
- Decibels: dB = 20 * log10(A / A_ref), so 6 dB is a 2x amplitude ratio and 20 dB is a 10x amplitude ratio.
- Snell's law for angle probes: sin(theta2) / sin(theta1) = v2 / v1. When v2 * sin(theta1) / v1 exceeds 1 no refracted wave exists (total internal reflection).
Workflow
- Establish the material and the wave mode (longitudinal for normal probes, shear for angle probes) and the corresponding velocity.
- Select frequency and transducer: compute wavelength and near-field length with wavelength() and near_field_length(), and confirm the region of interest sits past the near field or plan for its amplitude effects.
- Compute beam-spread half-angle with beam_spread_half_angle() for coverage planning, and the refracted shear angle with snell_refraction_angle() for angle-beam setups.
- Calibrate on the calibration block (IIW or step wedge): zero point, velocity, and sensitivity against reference reflectors (FBH or SDH), building the distance-amplitude correction baseline.
- Convert each measured time of flight to depth with time_of_flight_to_depth(), and each echo amplitude difference in decibels to an amplitude ratio with db_to_amplitude_ratio().
- Size indications with the 6 dB drop method or by comparing the echo with the reference reflector at the same depth, then compare with the acceptance criteria and disposition the part.
Pitfalls
- Forgetting the round trip: depth is tof * v / 2, never tof * v.
- Sizing inside the near field: amplitude maxima and minima inside N corrupt echo comparison; prefer depth-position sizing or a higher frequency transducer that moves the reflector into the far field.
- Dead-zone indications: small near-surface flaws hide behind the front-surface echo; use a delay-line or immersion setup when the near-surface zone is critical.
- Mixing wave modes: using the shear velocity where a longitudinal wave is expected (or the reverse) misplaces depth by up to 45%.
- Decibel confusion: 6 dB is half amplitude (voltage), 20 dB is 10x amplitude; applying 6 dB as an energy halving corrupts sizing.
- Ignoring total internal reflection: a refracted angle computed by Snell's law beyond 90 degrees means no shear wave enters the part; the probe angle must be reduced.
- Comparing echoes at different depths without the DAC curve: the same reflector size at two depths gives different amplitudes; always compare at equal depth or apply the distance-amplitude correction.
- Skipping calibration: velocity, zero point, and sensitivity drift between setups; every scan session recalibrates on the block.
Behavior contract (gate 3)
The inspection math is exercised by the gate 3 contract test: scripts/test_ultrasonic_inspection.py against scripts/ultrasonic_inspection_logic.py (stdlib unittest, offline). Run: python3 scripts/test_ultrasonic_inspection.py
Compliance
- Standards referenced, not reproduced: AS9100 clause 8.5.1.3 frames NDT as a special process requiring controlled procedures, qualified personnel, and records; the formulas and calibration practice above are common UT methodology, summary-only per standards-map.yaml.
- compliance: STANDARDS-REF, gated: false.