A 2D multiphysics model for vibration induced heating at closed defects
Résumé
Ultrasonic vibrothermography is known as a promising Non-Destructive Testing (NDT) technique for the detection of closed defects. The technique uses ultrasonic wave excitation to induce frictional behavior at the defect’s interfaces. As a result, the vibrational energy of the elastic wave will be locally converted into thermal energy and heat will be generated around the defect, which subsequently dissipates through the medium by way of thermal diffusion. In practice, the generated heat can then be recorded by an infrared camera, creating thermograms that allow to detect the defect. Apart from defect detection, defect characterization has become one of the major research interests in NDT, e.g. to anticipate the severity of the defect. To obtain reliable and high quality defect classification by ultrasonic thermography, theoretical models capable of interpreting the obtained thermograms are necessary.
In this study, a previously developed two dimensional model, consisting of a mechanics/dynamics module to describe the (nonlinear) interaction between elastic waves and a closed defect with rough surfaces, is extended with the calculation of the instantaneous friction-induced energy loss at the contact interface. The model allows to monitor the evolution of both the normal and tangential contact stresses and the temperature at each point on the defect interface. Figure 1 shows an example of the stresses and the temperature increase for a particular location on a surface breaking crack, simulated during 5 ms, after being periodically excited at 100 kHz. The simulated data clearly illustrates periodic temperature increases that are directly linked to the dynamic frictional contact behavior of the defect interfaces. This locally generated heat subsequently propagates through the sample, allowing for thermographic measurements to be performed on the surface of the sample. In order to confirm the reliability and accuracy of the model, a qualitative analysis will be performed by comparing the simulated data with numerical and experimental results described in literature.