Acoustic source localization without knowing the velocity profile in anisotropic plates is still one of the most challenging areas in this field. The current time-of-flight based approaches for localization in anisotropic media, are based on using six high sampling sensors. The number of sensors and the corresponding large amount of data, would make those methods inefficient in practical applications. Although there are many different non-time-of-flight based approaches such as machine learning, or soft computing based methods that can be used for localization with a less number of sensors, they are not as accurate as timeof- flight based techniques. In this article, a new approach which requires only four low sampling rate sensors to localize acoustic source in an anisotropic plate is proposed. In this technique, four electret low sampling rate sensors in two clusters are installed on the plate surface. The presented method uses attenuation analysis in a suitable frequency band to decrease the number of sensors. The approach is experimentally tested and verified on an airplane composite nose by applying artificially generated acoustic emissions (Hsu–Nielsen source). The results reveal that the accuracy of proposed technique depends on distinction of dominant frequency band. A stethoscope as a physical filter is employed to reduce the sensitivity of the technique and delineation of frequency band. The suggested technique improves the accuracy of localization prediction.

Keywords