Abstract:
Ultrasonic inspection as a non-destructive technique is used to inspect various materials for subsurface defects caused by air, cracks, material variation etc. Sometimes particles are added to clean materials to improve the strength and thermal conductivity of the material. While this improves the material properties, the presence of these particles causes attenuation of the ultrasonic signal, loss of signal strength and also signal scattering. Due to this, imaging subsurface interfaces with the complex matrix material as a primary interface is challenging. The images from subsurface interfaces are noisy and the features are not well resolved. This dissertation addresses the development of a transfer function for the noise caused by the filler particles that corrupts the Ultrasonic signal. The transfer function is based on frequency domain transform of modeled ultrasonic reflection from clean material and modeled ultrasonic reflection from the complex matrix material with particles in the size range of 150um-250um. A real sample with particles was made and imaged ultrasonically. The transfer function was applied to the A-scan data from the sample with particles to assess the performance of the transfer function for this application. The transfer function approach was extended to a case where the sample has a pattern on the back wall to assess the performance of this method.
