DATA CONDITIONING AND ITS IMPROVEMENTS ON THE PRESTACK INVERSION: A CASE STUDY FROM THE XINGMA AREA, CHINA

摘要

Seismic data used for prestack inversion and corresponding reservoir characterization should be of high quality, while any undesired prestack seismic phenomena that need to be diminished prior to inversion could possibly suppress the real geology. In an effort to improve the seismic data quality of Xingma area, Liaohe Oil field, China, data conditioning is used in this paper to optimize the match between the synthetic data, used in the prestack inversion, and the real data. The conditioning processes include multiple removal, random noise attenuation, and gather flattening. As the offset increases, the raw amplitudes of target area initially change from weak to strong, and subsequently from strong to weak. But after data conditioning, the amplitudes change from weak to strong. This kind of AVO anomaly is coincident with the synthetics. A comparison of wavelets extracted from four angle stacks found the amplitude and phase spectra of conditioned data to be much more stabilized in 0-42Hz. After data conditioning processing, the average RMS amplitude ratios of inversion residual to seismic demonstrate at least 20% drop in the amplitude, while the inverted Vp/Vs attribute shows a much more compact signature, providing a significant improvement on the accuracy of prestack inversion and corresponding reservoir characterization.