Michigan Technological University
SPECTRAL DECOMPOSITION OF SEISMIC DATA USING SPARSE S-TRANSFORM COMPARED TO CONVENTIONAL METHODS
Abstract
dc:description.abstractSpectral decomposition is a technique used to identify the component frequencies of a signal. Since seismic signals are nonstationary, spectral decomposition can be used as a tool for detecting spectral anomalies and interpreting geological features, making it valuable for identifying thin layers and potential hydrocarbon reservoirs. This research project presents the application of a newly developed spectral decomposition method known as the sparse S-transform (SST). This method calculates areas with the highest energy concentrations and eliminates areas containing unimportant features. It is based on inverse theory and uses an optimization problem to create sparse coefficients to prevent overfitting of data. This study performs two synthetic tests and a real seismic data example using the F3 North Sea dataset to compare the SST to conventional methods: short-time Fourier transform (STFT), continuous-wavelet transform (CWT), and S-transform (ST). Two synthetic wedge models were also created to demonstrate how thin layers are identified through spectral decomposition. The SST proves to generate solutions with superior time-frequency resolution and has the benefit of optimized window parameters. It also has advantages for interpreting geological features found in the F3 dataset, such as faults, sigmoidal bedding, thin layers, and unconformities.
Degree
thesis:*- Name thesis:degree_name
- Master of Science in Geophysics (MS)
- Level thesis:degree_level
- Campus Access Master's Thesis
- Discipline thesis:degree_discipline
- Department of Geological and Mining Engineering and Sciences
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Donahue, Kara
- Committee members dc:contributor.committeemember
-
- Wayne D. Pennington
- Mir Sadri-Sabet
- Contributors dc:contributor
-
- Roohollah Askari
Subjects
dc:subject × 7Identifiers
dc:identifier.*- Identifier
- https://digitalcommons.mtu.edu/etdr/870