Fracture bodies usually develop in the slip fault systems in sedimentary basins,and the fractures have strong concealment,and traditional fracture identification techniques do not perform well.Existing fracture inter-pretation strategies for target layers usually adopt a local view,ignoring the overall characteristics of fracture body fractures.Through the use of the U-ResNet deep learning model,all strata of the Zhenjing Block in Ordos Basin were identified for fractures.Combined with dip-guided seismic attribute slices,the formation mecha-nism,periods and levels of fractures were revealed,and for the first time,the NWW slip distance was estimat-ed.By analyzing the extension characteristics of deep fractures,it was confirmed that the three groups of NWW-oriented slide fractures in the southern part of the block are essentially flower structures,and their rootstock ex-tended to the basal fractures,confirming the reactivation of basal fractures in multiple tectonic movements.The cross-section and plan style of Chang 8 fracture body were divided,and a diamond-pulling rift was identified on the NEE fracture,providing seismic evidence for the NEE fracture slip movement.In addition,three spindle-de-pression fracture combinations were found,explaining the formation reasons of staggered-step faults,and giving the favorable combination style of fracture bodies and their distribution positions on the plane.The study shows that the application of deep learning fracture technology and a full-view interpretation strategy helps to reveal the development characteristics and evolution rules of complex high-angle slip fracture bodies.
NETL
NSTL
万方数据
