Nondestructive testing (NDT) examines a component for discontinuities without intentionally destroying its ability to remain in service. Different methods probe the material through different physical interactions, so no single method can reveal every defect.

Common methods include:

• visual testing: light reflected from accessible surfaces reveals visible cracks, corrosion, misalignment or geometric damage;
• liquid penetrant testing: a liquid that readily wets the surface is drawn into openings that reach the surface. After excess liquid is removed, penetrant emerging from a crack makes the opening easier to see;
• magnetic-particle testing: a ferromagnetic component is magnetized. A crack can interrupt the path of magnetic flux and create a leakage field near the surface, attracting fine magnetic particles;
• ultrasonic testing: short pulses of high-frequency mechanical waves travel through the material. Part of a pulse is reflected when it reaches a boundary or region with different acoustic properties, so echo time and strength can reveal an internal discontinuity;
• radiographic testing: X-rays or gamma rays pass through the component and are attenuated by the material. Differences in thickness or density change how much radiation reaches a detector, producing image contrast.

Method selection must match the expected defect. A closed internal void cannot be found by penetrant testing because no liquid can enter it from the surface. Magnetic-particle testing requires a material that can be strongly magnetized. A planar internal crack may be well suited to ultrasound, but its detectability depends strongly on the angle between the crack and the sound beam.

For example, consider an aluminium plate suspected of containing a crack. If the crack is known to reach an accessible surface, penetrant testing can provide a simple sensitive test. If the suspected crack is internal, penetrant testing cannot reveal it; ultrasound or radiography may be appropriate depending on geometry, thickness and crack orientation.

Detection is therefore probabilistic rather than absolute. Accessibility, surface condition, equipment, calibration, operator procedure, defect size and orientation all affect the probability of detection.

The useful question is not “Which NDT method is best?” but which physical interaction can make the relevant discontinuity observable, under the conditions of this inspection?