dc.description.abstract | High-quality seismic imaging contributes to geological interpretation, rock property analysis, reservoir characterization and hydrocarbon exploration. Shear wave imaging has been significantly utilized with the development of advanced processing techniques, which improve the quality of seismic stacked imaging. Multi-component data used in this research provide an opportunity to study both compressional waves and converted shear waves. However, the acquisition system in the shallow-water Arabian gulf with the hard seafloor gives rise to dispersive surface waves, which contaminate reflection signals severely. In this thesis, a processing workflow is designed for the 4C OBC seismic data to obtain P-P wave, converted PP-S and PS-S wave imaging in four-component data.
In this study, by considering their characteristics in linearity, surface waves can be separated from reflection signals. Three methods are compared and implemented to remove surface waves, including a bandpass filter, a F-K filter and a F-X filter (named as Surface Wave Noise Attenuation in ProMAX). According to surface wave analysis and comparison of filtering results, the Surface Wave Noise Attenuation methodology can lead to the best result of surface wave removal in shot and receiver domain. Thenceforth, residual noises at far offsets still destroy the reflected signals. FK filters are applied in shot, receiver and CDP domain to suppress residual noises and enhance signals. After several rounds of noise attenuation, the signal-to-noise ratio has a remarkable increase in 4C seismic data, and clear reflection events appear in the CDP records.
Besides compressional waves, two kinds of converted waveforms exist in the shallow-water environment: PP-S wave (converted from subsurfaces) and PS-S wave (converted at the sea bottom). All wave modes are recorded in the four-component seismic data, so the enhancement of each wave mode is essential to achieve their stacked images. By using the P stacking velocity to do NMO, only the reflection events with P velocities are corrected to flat. After filtering out up-dipping and down-dipping events via FK filters, P wave reflection signals are enhanced, and the velocity analysis of P waves is more accurate. In the same way, converted PP-S and PS-S waves are strengthened and their stacked sections are attained separately in four-component data.
To evaluate these stacked images, synthetic traces are created by convolving the Ricker wavelet with reflection coefficients, which relate to the sonic log and density log data. By plotting synthetic traces in seismic images at CDP702, first, assess how well stacked sections match seismology traces. Then, the energy distribution of various wavefields is analyzed in four components. These two criteria decide the better stacked images to describe the real geological structures. Ideally, these images should reflect the same subsurface structures. However, because of various wavefields and different measurement, there exist slight variations between P-P wave, PP-S and PS-S wave imaging. In summary, the processing sequence developed in the thesis achieved signal enhancement, P wave and converted shear wave imaging, which can improve the geological interpretation and hydrocarbon exploration. | en |