Song, Jianyong (RIPED, CNPC) | Li, Jinsong (RIPED, CNPC) | Zheng, Xiaodong (RIPED, CNPC) | Xie, Zhanan (RIPED, CNPC) | Wang, Daxing (Research Institute of CPOCC) | Wang, Guanchao (China University of Petroleum-Beijing)
Summary The elastic reflection travel-time inversion (ERTI) based on conventional elastic wave equations obtains low wave number components in the model parameters by using travel-time and reflected wave information. However, the coupling effect between the PS waves and the sensitivity of velocities to wave fields enhance the nonlinear problem of inversion. Therefore, in this paper, the reflection traveltime inversion based on the decoupled elastic wave equations is studied, and an improved time-shift crosscorrelation objective function is proposed to implicitly calculate the relative time shift in image domain respectively. The coupling between PS waves is greatly reduced and the inversion results of low wavenumber components are improved. Finally, the Marmous II model tests prove the correctness of this method.
Duan, Yanting (Institute of Oil & Gas, Peking University, Beijing, China) | Zheng, Xiaodong (Research Institute of Petroleum Exploration & Development, PetroChina, Beijing, China) | Hu, Lianlian (Research Institute of Petroleum Exploration & Development, PetroChina, Beijing, China)
Unsupervised seismic facies are a convenient and efficient method for interpretation. Current seismic facies analysis mostly focuses on the improvement of the precision of seismic facies belt prediction. In this paper, we propose a clustering method that simultaneously learns feature representations and cluster assignments by using deep auto-encoder network, which improves the clustering result. First, this method learns a mapping from the high-dimensional data space to a low-dimensional feature space. Then, it iteratively optimizes the clustering objective by minimizing the error between cluster centers and embedded points as the constraint condition. The application with actual data demonstrates that the proposed method is helpful for improving the precision of seismic facies prediction, which may be conducive to the seismic interpretation.
Presentation Date: Tuesday, October 16, 2018
Start Time: 9:20:00 AM
Location: Poster Station 1
Presentation Type: Poster
Summary This paper introduces a new time-frequency analysis method, which is called General Linear Chirplet Transform (GLCT), into seismic data interpretation. The GLCT method is an extended form of the linear chirplet transform (LCT) by scanning the chirp rates of chirplet atoms at each time-frequency point and picking the one with the maximum projection upon seismic signals under analysis. The chirplet is a generalized wavelet with time-varying frequency and LCT is a subspace of chirplet transform encompassing short time Fourier transform (STFT) and continuous wavelet transform (CWT). Therefore, GLCT, being a generalized form of STFT and CWT as well, could adaptively fit the instantaneous frequency of seismic signals and produce higher energy concentration in the time-frequency plane without cross-terms interferences, enabling a better time-frequency resolution than STFT and CWT, particularly at the low and high frequency ends. The synthetic and the field data examples demonstrated that GLCT could produce spectral decomposition results with considerably robust and improved performance in stratigraphic visualization and hydrocarbon detection in contrast with conventional methods, making it a potential tool in reservoir characterization.
Seismic AVOA method, which employs P-wave amplitude variation with incident angle and acquisition azimuth, is a powerful tool for fracture characterization in HTI medium. However, this azimuthal AVO based fracture characterization approach is only applicable to fracture set with single dominant orientation, which is not always true in real applications. Ant tracking is able to delineate small scale discontinuities, which has a close relationship to fractures. Curvature is also a popular method for fracture characterization. In this paper, we perform fracture characterization by comprehensively utilizing azimuthal AVO, ant tracking and curvature analysis. We select optimal incident angle range for azimuthal AVO based fracture inversion. To make full use of wide azimuth seismic data, we perform ant tracking on six azimuthally sectored seismic volumes and then combine the results together to better delineate fractures set with multiple orientations. Curvature attribute is also introduced as a useful aid to AVOA and ant-tracking for fracture characterization. We applied this integrated fracture characterization scheme to Tarim Basin, Northwest China. Application results showed that there is a good accordance in the results produced by AVOA, multi-azimuth seismic volume ant-tracking, and curvature. The estimated fracture orientation was verified by image log data. These results indicate that the proposed method is effective in fracture characterization.
Liu, Xingfang (Research Institute of Petroleum Exploration and development) | Zheng, Xiaodong (Research Institute of Petroleum Exploration and development) | Xu, Guangcheng (Research Institute of Petroleum Exploration and development) | Yang, Hao (Research Institute of Petroleum Exploration and development) | Song, Jianyong (Research Institute of Petroleum Exploration and development)
Li, Yandong (Research Institute of Petroleum Exploration and Development, PetroChina Company Limited) | Li, Jinsong (Research Institute of Petroleum Exploration and Development, PetroChina Company Limited) | Zheng, Xiaodong (Research Institute of Petroleum Exploration and Development, PetroChina Company Limited)
No preview is available for this paper.
No preview is available for this paper.