Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.636644
Title: Quantitative sediment identification using remote acoustic techniques
Author: Williams, J. P.
Awarding Body: University College of Swansea
Current Institution: Swansea University
Date of Award: 1978
Availability of Full Text:
Access from EThOS:
Abstract:
A laboratory study has been carried out on the reflection/scattering characteristics of unconsolidated clastic sediments. Broad-band acoustic measurements (100 klhz - 600 klhz) have been made at near-normal incidence on samples ranging from those with a mean grain size of 90pm to those containing particles up to 13mm in diameter. The measured acoustic returns have been deconvolved with smooth surface calibration traces, enabling the roughness of the sediment surface and, hence, mean grain size of the sample to be determined. These estimates of surface roughness were used in comparing experimental scattering levels with those obtained using a modified form of Eckart's theory of scattering from a rough surface. The same theory has been used to predict surface parameters of the coarser samples. The level of the mean amplitude reflection coefficient for each sediment is determined by the impedance mismatch at the water/sediment interface and by scattering due to the rough sediment surface. The form of the distribution of the reflection coefficients has been found to change from near-Gaussian for low values of surface roughness parameter g, to near-Rayleigh as the rough surface limit is approached (g = 10). Quantitative discrimination between sediments has been achieved by applying pattern recognition techniques to sequences derived from the measured acoustic returns. The primary classification parameters have been identified as being the frequency content, time-spreading and absolute level of the reflected returns. Correct classification levels approximately 40% above random have been achieved when the data are normalised for both 3-way and 5-way classification tests, while calibrated data yield levels approximately 50% above random.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.636644  DOI: Not available
Share: