This is a repository copy of Seismic waveform classification and first-break picking using convolution neural networks.

No abstract

No abstract

Deep-formation oil/gas exploration is a key objective in the geophysical field, and structural and stratigraphic discontinuities, such as faults and channels, usually contribute to the construction of traps and reservoirs. Coherence has been used successfully to identify these abnormal features in seismic amplitude volumes. However, the current coherence algorithms seldom involve the geologic concept. We propose geosteering coherence attributes by implementing the coherence calculation perpendicular to the direction of the structural trend in a 3D curved plane. We estimate a group of time lags between the original analysis trace and each original neighboring trace along a certain spatial direction by using dip scanning. For each spatial direction, we subsequently construct two new model traces by weighting phase traces derived from the complex seismic traces, in which time lags are eliminated. We then use the new model traces to compute the crosscorrelation coefficients for each spatial direction. We finally obtain the 3D geosteering coherence attributes by taking the minimum values among the modulus of the crosscorrelation coefficients along different spatial directions to approximately characterize the coherence perpendicular to the structural trend in a 3D curved plane. An example of the 3D physical modeling involving fracture groups and faults embedded in the deep formation is used to demonstrate the effectiveness of the 3D geosteering coherence attributes. The applications on two real 3D seismic data sets of sand reservoirs from western deep formation illustrate that our method can alleviate the influence of dipping strata and can highlight subtle structures. Compared to the conventional coherence method, our method can highlight subtle geologic structures more and better, suggesting that it may be serve as a future tool for detecting the distribution of geologic abnormalities in deep exploration.

Abstract-Fracture monitoring is crucial for many geoindustrial applications, such as carbon dioxide storage and hydrocarbon exploration in tight reservoirs, because fractures can form storage space or leaking paths for geological sealing. We propose a fracture identification framework for geo-industrial applications by exploiting seismic reflection anisotropy and automatic multi-sensitive attribute fusion. Anisotropy maps extracted from different seismic attributes are automatically selected and fused according to the correlation between the predicted anisotropy strengths and the measured fracture densities at well locations. Through seismic anisotropy extraction and automatic multisensitive attribute fusion, we can acquire a more comprehensive evaluation of different fracture types in a reservoir. The proposed fracture identification framework is successfully applied to a deep, tight sandstone reservoir in southwest China. The predicted fracture distribution is closely related to the local structures in the target reservoir. The orientations of most predicted fractures are consistent with the local maximum principal stress direction in this area, which is good for the opening and fluid filling of fractures. The fracture identification results will be used to guide hydrocarbon exploration activities in this region, such as exploration well deployment.

No abstract

Detection and identification of subsurface anomalous structures are key objectives in seismic exploration. The coherence technique has been successfully used to identify geologic abnormalities and discontinuities, such as faults and unconformities. Based on the classic third eigenvalue-based coherence ([Formula: see text]) algorithm, we make several improvements and develop a new method to construct covariance matrix using the original and Hilbert transformed seismic traces. This new covariance matrix more readily converges to the main effective signal energy on the largest eigenvalue by decreasing all other eigenvalues. Compared with the conventional coherence algorithms, our algorithm has higher resolution and better noise immunity ability. Next, we incorporate this new eigenvalue-based algorithm with time-lag dip scanning to relieve the dip effect and highlight the discontinuities. Application on 2D synthetic data demonstrates that our coherence algorithm favorably alleviates the low-valued artifacts caused by linear and curved dipping strata and clearly reveals the discontinuities. The coherence results of 3D real field data also commendably suppress noise, eliminate the influence of large dipping strata, and highlight small hidden faults. With the advantages of higher resolution and robustness to random noise, our strategy successfully achieves the goal of detecting the distribution of discontinuities.

No abstract