Nondestructive testing is a combination of analysis techniques, used both in science and industry, for the detection of faults and defects in a given component without causing damage. The use of Nondestructive Tests (NDT) in the industry, allow cost reductions, increased product quality and higher quality standards.
The majority of NDTs used nowadays, such as visual inspections, fluorescent penetrating ink, magnetic particle inspection ink and eddy current testing, amongst others, require trained operators to track defects and subjectively interpret results .
The outcome is a test subject to human error, which requires specially trained operators, and represents significant time loss in testing each individual component and in producing results. Although there are other NDT methods (for further information review Tur and Aygun’s work ), many are not widely known due to their higher installation and maintenance costs.
Acoustic Resonance as a method to detect defects is not a novelty; it has been used since ancient times to diagnosed defects. That is because when faulty objects are impacted, they produce a different sound that is usually detected by the human ear . This means that when hitting an object, the mechanical vibrations produced in its structure (sound transmitted by solids) and the vibrations transmitted to the surrounding air (audible sound) carry information. Resonance is a clear physical criterion that describes an object in its totality (components, in this case), both internally and externally, with information on the object’s material, structure and geometry. Differences in these vibrational characteristics allow us to find defects in the object , , . Acoustic Resonance Testing is also known as ART.
Conventional defect diagnosis methods are not comprehensive, in other words they require numerous measurements in small areas, increasing difficulty and costs . On the other hand, ART covers the totality of a component, since measurements performed in one area of the component can capture defects from other parts. Also, the test is a quick, it can be completely automated and allows for objective and measurable results, removing human error and avoiding the use of ink-based aerosols that contain toxic metallic particles. This allows us to perform quality control over 100% of components in a production line.
Typical defects detected included cracks, cavities, porosity, nodules, hardness, out of tolerance dimensions, density variations, amongst others . Each imperfection will change the acoustic response in some way. But depending on the magnitude of the imperfection, these changes can be very subtle, making acoustic signal processing a key element .
Procedure for acoustic response measurements
The AFENsis system procedure for the detection of defects is detailed below. First, an impact hammer stimulates the component under analysis.
The impact acts like a white impulse providing constant energy to all frequencies, producing vibration in the component. The frequencies that do not correspond to resonance modes are rapidly diminished; on the other hand, the oscillations corresponding to the objects vibration modes persist in time producing sound waves. A microphone captures the acoustic response of the object and data is stored digitally on a computer. High sample frequencies are used (44.1 kHz or 48 kHz) to ensure that all relevant resonance modes are captured. A Hann window is applied to the data collected to reduce the spectral leakage and to minimize any distortion produced by the sound of the hammer’s impact. To identify the resonance frequencies (spectral peaks), a Fourier transform is applied to the data.
In components with multiple vibration modes, not all modes are affected by the presence of a defect. The modes affected are only those where the faulty section suffers a malformation, this modifies the resonance frequency. Consequently, a faulty object can be identified by comparing its specter with the specter of a fault free component, and locating a displacement in the resonance frequency  (figure 2).
Before running the system, there’s a consideration that should be taken into account. The vibrational behavior must first be determined for a referential or standard component without defects and with the smallest error possible. This is commonly accomplished by recording the acoustic response of several components presumably without defects, and then averaging all measurements.
 E. Coffey, “Acoustic resonance testing,” in Future of Instrumentation International Workshop (FIIW), 2012, pp 1-2. 8
 K. Tur and H. Aygun, “The Role of Nondestructive Inspection as a Part of Quality Assurance in Casting Industry.,” Indian Foundry Journal, vol no. 1 pp. 33-41, 2003.
 A. Zapico, L. Molisani, R. O’Brien, J. C. Del Real, Y. Ballesteros, N. Ponso, “Diagnóstico Global de Fallas en Vigas de Aluminio Usando Niveles de Presión Sonora,” Mecánica Computacional, vol. XXX, no. 42, pp. 3271-3276, 2011.
 I. Hertlin and D. Schultze, “Acoustic Resonance Testing” the upcoming volume-oriented NDT Method,” in III Pan-American Conference for Nondestructive Testing, 2003.
 G. R. Stultz, R. W. Bono, and M. I. Schiefer, “Fundamental of Resonant Acoustic Methods NDT,” in Advances in Powder Metallurgy and Particle Materials, Montreal, Canada, 2005, vol. 3, pp. 1-11.
 R. W. Bono and S. Sorensen, “Resonant Acoustic Method Delivers Defect-free Parts, “Advanced Materials & Processes, pp. 25-28, Jul-2008.
 M. I. Schiefer and L. Sjoeberg, “Physical Basis of the Resonant Acoustic Methos for Flaw Detection,” in Advances in Powder Metallurgy and Particle Materials, Montreal, Canada, 2005, vol. 3 pp. 87-97.