Abstract The Y reservoir in the Tarim Basin is a fault-controlled buried hill reservoir. Affected by factors such as deep embedding depth and high degree of stratum fragmentation, the signal-to-noise ratio of is low and the imaging effect of faults is poor. Therefore, it is very difficult to identify faults accurately. The conventional geometric attribute plane is extremely messy, the interference information is serious, and the fault information is difficult to be effectively identified, which cannot solve the problem of fine identification of buried hill reservoir faults. Aiming at the demand of fault characterization in reservoir area, combined with the artificial intelligence analysis method of big data, a fault identification technology under supervised pattern is proposed. Firstly, through cascade denoising and fault space imaging enhancement processing, the signal-to-noise ratio of seismic data and the fault imaging effect is improved. Secondly, the fault types are divided according to the discontinuous characteristics of the seismic axis, and different types of faults are characterized by multi-type sensitive attributes. Finally, the training samples are established according to the fault type. Through the iterative constraint training method, the fault identification work under the supervised pattern is carried out, and the fine characterization of different types of faults in Y reservoir is completed. Relying on the artificial intelligence big data analysis method, this technology has the innovation of full-type fault information utilization and supervised constraint training, which plays an important guiding role in the fine identification of faults in low signal-to-noise ratio areas.
Read full abstract