| The bright spot technology in seismic exploration has been directly used for hydrocarbon detection in reservoirs for more than fifty years and has achieved great success in practice.However,not all oil and gas reservoirs show bright spots on seismic profiles,and some oil and gas reservoirs formed in special depositional environments show non-bright spots.Such reservoirs have weak amplitudes on the seismic profile and are extremely hidden and easily overlooked.Exploration practice shows that non-bright spot sandstone reservoirs widely exist in major petroliferous basins and have great exploration potential.So far,the non-bright spot sandstone reservoirs discovered are mainly encountered by drilling,and then the interpreters look for evidence of their existence from the profile.Therefore,it is of great significance to research and develop more effective seismic identification and characterization methods for non-bright spot sandstone reservoirs.Based on previous research work,combined with the field data of the Zhongjiang gas field in the Sichuan Basin,we develop a set of identification and characterization methods for non-bright spot sandstone reservoirs,which specifically includes the optimization processing and interpretation of pre-stack gathers,automatic identification of reservoirs on post-stack data,and fine characterization.Finally,applying this method,two hidden non-bright channel sandstone reservoirs were discovered in the Shaximiao Formation of Zhongjiang Gas Field.The main research contents and achievements of the paper are as follows:(1)Based on the analysis of the seismic response characteristics of the non-bright spot sandstone reservoirs in the Shaximiao Formation of the Zhongjiang Gas Field,a method for highlighting the seismic response characteristics of the non-bright spot sandstone reservoirs was developed,including the pre-stack common reflection point gather based on tensor low-rank constraints.Random noise suppression processing,anisotropy and residual velocity correction,polarity reversal angle analysis,optimal stacking angle optimization,etc.,to avoid the pitfall of amplitude identification in non-bright spot sandstone reservoirs.(2)A method and technique for automatic identification of non-bright spot sandstone reservoirs were developed by using the deep transfer learning method.Transfer pre-trained classification network models that perform well under large-scale annotations in the natural image domain to the domain of reservoir identification under weakly-labeled conditions.By fine-tuning and training the network to make it meet the current task requirements,and finally using the trained network to identify the non-bright channel sandstone reservoirs successfully.(3)Research and development of non-bright spot sandstone reservoir comprehensive characterization method.Including methods and techniques for analyzing channel development stages and sedimentary evolution laws based on instantaneous amplitude in the Wheeler domain,methods and techniques for predicting sandstone reservoir thickness by using sub-band principal component RGB color space fusion,and using normal map technology to shade coherent energy attributes,In this way,the method and technology of describing the channel boundary can be more clear,and based on the sandstone reservoir results identified by deep transfer learning,the seed region growth algorithm can be used to pick up the cloud point data of the connected non-bright spot sandstone reservoirs,and then complete the methods and techniques for 3D characterization of non-bright spot sandstone reservoir,etc.(4)The seismic data of the Shaximiao Formation in the Zhongjiang Gas Field were reinterpreted by using the research and development of the non-bright spot sandstone reservoir identification and characterization method,and the new non-bright spot sandstone reservoirs were discovered.First,AVO forward modeling shows that the "positive and negative cancellation" of the near-and far-angle traces in the pre-stack gathers is the root cause of the Channel hidden.Then the method of the far-angle stack was used to avoid the seismic amplitude pitfall of the non-bright spot sandstone reservoirs,the original discontinuous channel sandstone became more continuous on the horizon,and at the same time,two hidden non-bright channel sandstone reservoirs were successfully identified,and their development stage,thickness,width,and three-dimensional spatial geometry were characterized.The application results show that the developed non-bright spot sandstone reservoir identification and characterization method is reliable and effective. |