The Study Of Carbonate Fault Zone Identification Based On Seismic Attributes

With the deep development of the world oil and gas exploration, the carbonate reservoirbecomes more and more important. The oil reserves discovered in carbonate reservoiraccounts for about half of the world’s total oil reserves and production has reached more than60%. Most of carbonate regions are fractured reservoirs. Geological features such as Faults,channels and cracks are the basis to performance and describe reservoir in fractured reservoirs.Fault is the root cause of control petroliferous basins in a variety of geological process and oiland gas produces, migration, accumulation, bulk and distribution. Fault either as hydrocarbonmigration pathways, and can block the surface for hydrocarbon accumulation, and thus thefine identification of faults and fractures become a key factor in identification of carbonatereservoirs.There is plenty of geological information in the seismic data, seismic attribute analysiscan extract useful information from the seismic data hidden. After decades of development,seismic attributes that can be extracted reached hundreds of species. In many seismicattributes, we mainly study attributes that can reflect lateral changes in reflector time, reflectlateral changes of reflector waveform, and reflect lateral changes in amplitude.There are usually two problems in post-stack seismic data, contaminated with randomand coherent noise, another is their relatively low bandwidth, so to precondition the seismicdata is needed. Structure-oriented filtering is used to suppress noise, inverse Q filtering,spectral blueing and colored inversion are used to improve seismic resolution. Volumetric dipand azimuth, coherence and curvature attributes on preconditioned seismic data can describefault and fracture in different ways, to realize identification and characterization carbonatefault zone. At last, there are problems for example huge computation and spend too muchtime in3D seismic attributes extracted, with GPU parallel computing, greatly accelerate thespeed of the attribute extraction.