Reliable applications of adhesively bonded joints require an effective nondestructive evaluation technique for their bond strength prediction. To properly evaluate factors affecting bond strength, effects of defects such as voids and disbonds on stress distribution in the overlap region must be understood. At the same time, in order to use the acousto-ultrasonic technique to evaluate bond quality, the effect of these defects on the dynamic response of single lap joints must be clear.
The Stress distribution in a single lap joint with and without defects (void or disbond) is analyzed. A θ parameter that contains adherend and adhesive thickness and properties is introduced. It is shown for bonded joints with θ ≥ 10 that a symmetric void or disbond in the middle of overlap up to 70% of overlap length has a negligible effect on bond strength. In contrast, frequency response analyses by a finite element technique showed that dynamic response is affected significantly by the presence of voids or disbonds. These results have direct implication in the interpretation of acousto-ultrasonic results.
Through-transmission attenuation and a number of acousto-ultrasonic parameters for various specimens with and without defects are evaluated. It is found that, although void and disbond have similar effects on bond strength (stress distribution), they have completely different effects on wave propagation characteristics. For steel-adhesive-steel specimens with voids, attenuation changes are related to the bond strength. However, the attenuation changes for specimens with disbond are fairly constant over a disbond range. In order to incorporate the location of defects in acousto-ultrasonic parameters, a weighting function is introduced. Using an immersion system with focused transducers, a number of acousto-ultrasonic parameters are evaluated. It is found that, by incorporating weighting functions in these parameters, better sensitivities (acousto-ultrasonic parameters versus bond strength) are achieved.
Acoustic emission activities of steel-adhesive-steel specimens with θ = 3.4 are monitored. Two different formats of energy versus time have resulted, each corresponding to the perfect specimens or the specimens with void or disbond. The relative acoustic energy and number of events at failure are found to be a means of predicting bond strength.