Non-destructive testing methods

Why it matters

Castings, forgings, welds, pressure vessels, rails and aircraft parts are inspected for cracks, porosity and inclusions without scrapping them, both at manufacture and during service. Choosing the right non-destructive testing (NDT) method — one that can actually see the defect type, depth and material in question — is a routine quality-control decision, and getting it wrong lets dangerous flaws through.

Key ideas

NDT examines a part without impairing its future usefulness. Every method uses some physical interaction (light, liquid, magnetism, sound, radiation, induced current) that a defect disturbs. No single method finds everything, so the choice depends on the material (magnetic or not, conducting or not), the defect location (surface, near-surface or internal) and orientation, and the part geometry.

Visual inspection (VT). The first and cheapest step: eye, magnifier, borescope or camera. Finds only visible surface faults.

Liquid (dye) penetrant testing (PT/LPT). Steps: clean and degrease → apply penetrant (red visible dye or fluorescent) → allow dwell time for capillary action into cracks → remove excess → apply developer, which draws penetrant back out → inspect (white light or UV) → clean.

• Any non-porous material: metals (magnetic or not), ceramics, plastics.
• Detects only defects open to the surface. Not for porous materials. Surface must be clean; paint and smeared metal from machining can close cracks.

Magnetic particle testing (MT/MPI). The part is magnetised; a surface or near-surface crack disturbs the flux and creates a leakage field that attracts fine iron particles (dry powder or wet suspension, often fluorescent).

• Ferromagnetic materials only (carbon and low-alloy steels, not austenitic stainless steel, aluminium or copper).
• Finds surface and slightly sub-surface defects (a few millimetres).
• Most sensitive to cracks perpendicular to the magnetic field, so parts are magnetised in two directions (circular and longitudinal). Demagnetise afterwards.

Ultrasonic testing (UT). A piezoelectric probe sends short pulses of 0.5–25 MHz sound into the part through a couplant (gel, oil or water). Reflections from flaws and from the back wall return to the probe.

• Pulse-echo: one probe sends and receives; the A-scan display shows echo amplitude against time. Depth = velocity × time / 2.
• Through-transmission: separate probes on opposite sides; a flaw reduces the transmitted signal.
• Angle-beam (shear-wave) probes: for welds, where flaws are not parallel to the surface.
• Phased-array UT steers and focuses the beam electronically and images a sector.
• Strengths: deep penetration (metres in steel), accurate depth location, one-side access, portable, safe, thickness gauging. Limitations: couplant needed, operator skill, difficult on coarse-grained, rough, thin or complex parts; planar flaws parallel to the beam are missed; a near-surface dead zone.
• Higher frequency = shorter wavelength = better resolution of small flaws but more attenuation (less penetration). The smallest detectable flaw is roughly half a wavelength.
• Reflection at an interface depends on the acoustic impedance mismatch Z = ρ·v; a steel–air interface reflects almost 100 %, which is why couplant is essential and why cracks reflect strongly.

Radiographic testing (RT). X-rays (from a tube) or gamma rays (from isotopes such as Ir-192 or Co-60) pass through the part onto film or a digital detector. Less material (a void, gas porosity, crack along the beam) lets more radiation through and appears darker on film; dense inclusions (tungsten) appear lighter.

• Gives a permanent image; good for volumetric flaws in welds and castings (porosity, slag, shrinkage, lack of penetration).
• Planar cracks are seen only if roughly parallel to the beam. Needs access to both sides, radiation safety and licensing; thick sections need high energy.
• Image quality is checked with an image quality indicator (IQI or penetrameter, wire or hole type). Sharpness is limited by geometric unsharpness, which falls with a smaller source, larger source-to-object distance and the film close to the part.

Eddy current testing (ET). An AC coil induces eddy currents in a conducting part; flaws, conductivity changes or thickness changes alter the coil impedance.

• Conductive materials only; surface and near-surface defects because currents concentrate within the skin depth.
• Fast, no contact or couplant, automatable: tube inspection in heat exchangers, aircraft skins and bolt holes, sorting alloys and heat-treatment conditions, coating-thickness measurement.

Other methods. Acoustic emission (listens for stress waves from growing cracks during loading of pressure vessels), infrared thermography (composites, bond lines), leak testing, and computed tomography (3-D X-ray).

Selection summary.

• Surface-breaking crack in aluminium casting → PT or ET.
• Surface or near-surface crack in a steel shaft → MT.
• Internal crack or lamination in a thick forging or plate → UT.
• Porosity and slag in a butt weld, with permanent record → RT (UT increasingly replaces it).
• Wall-thickness loss in heat-exchanger tubes → ET or UT.

Formulas

λ = v / f Ultrasonic wavelength (m): v = sound velocity in the material (m/s; about 5900 m/s longitudinal in steel), f = frequency (Hz).

d = v·t / 2 Depth of a reflector in pulse-echo UT (m); t = round-trip time (s).

Z = ρ·v and R = [(Z2 − Z1)/(Z2 + Z1)]² Acoustic impedance (kg/(m²·s)) and intensity reflection coefficient at normal incidence.

I = I0·exp(−μ·x) and HVL = ln 2 / μ Radiation intensity after thickness x (m) of material with linear attenuation coefficient μ (m⁻¹); HVL = half-value layer thickness.

Ug = F·t / D0 Geometric unsharpness (mm): F = source size, t = object-to-film distance, D0 = source-to-object distance (all in mm).

δ = 1 / √(π·f·μ·σ) Eddy-current standard depth of penetration (m): f in Hz, μ = μ0·μr (μ0 = 4π × 10⁻⁷ H/m), σ = electrical conductivity (S/m).

Worked examples

Example 1 (standard): ultrasonic pulse-echo. Given: steel plate, v = 5900 m/s, 5 MHz probe. Back-wall echo at 34 µs; a flaw echo at 20 µs.

1. λ = 5900 / (5 × 10⁶) = 1.18 × 10⁻³ m = 1.18 mm (flaws down to about 0.6 mm can be resolved).
2. Thickness = 5900 × 34 × 10⁻⁶ / 2 = 100.3 mm.
3. Flaw depth = 5900 × 20 × 10⁻⁶ / 2 = 59.0 mm below the probe surface.

Example 2 (GATE level): couplant, radiography and eddy currents.

1. Water (ρ = 1000 kg/m³, v = 1480 m/s) to steel (ρ = 7850 kg/m³, v = 5900 m/s): Z1 = 1.48 × 10⁶, Z2 = 4.63 × 10⁷ kg/(m²·s). R = [(46.3 − 1.48)/(46.3 + 1.48)]² = 0.88, so only 12 % of the energy enters the steel; without couplant (air, Z ≈ 400) almost nothing would.
2. Radiography with μ = 0.05 mm⁻¹ in steel: HVL = 0.693/0.05 = 13.9 mm; transmission through 25 mm = e^(−1.25) = 0.287.
3. Unsharpness for a 3 mm source, 25 mm object-to-film distance and 700 mm source-to-object distance: Ug = 3 × 25 / 700 = 0.107 mm.
4. Eddy-current depth in copper (σ = 5.8 × 10⁷ S/m, μr = 1) at 2 kHz: δ = 1/√(π × 2000 × 4π × 10⁻⁷ × 5.8 × 10⁷) = 1.48 mm; at 50 kHz it falls to about 0.30 mm.

Common mistakes

• Using penetrant testing to look for internal porosity; it only finds surface-open defects.
• Applying magnetic particle testing to austenitic stainless steel or aluminium.
• Forgetting the factor of 2 in pulse-echo depth (sound travels down and back).
• Expecting radiography to show a tight crack lying perpendicular to the beam.
• Thinking higher UT frequency always helps; it improves resolution but reduces penetration.
• Magnetising in only one direction and missing cracks parallel to the field.

For GATE PI

Expect matching questions (method ↔ principle ↔ detectable defect ↔ material limitation), sequences of steps in penetrant testing, and which method suits a given component. Numericals use λ = v/f, pulse-echo depth, attenuation or half-value layer in radiography, geometric unsharpness and eddy-current skin depth.

Quick check

1. Which NDT methods can find a surface crack in an aluminium alloy?
2. A UT echo returns after 8 µs in steel (v = 5900 m/s). How deep is the reflector?
3. Why is a couplant used in UT?
4. In MT, how should a crack be oriented relative to the field for best detection?
5. What does an IQI (penetrameter) check?

Answers: 1. Penetrant and eddy current testing (also UT surface waves); 2. 23.6 mm; 3. To exclude air, whose impedance mismatch would reflect almost all the sound; 4. Perpendicular to the magnetic field; 5. The sensitivity and quality of the radiograph.