## Identification of deep reservoir fluids based on basis pursuit inversion for elastic impedance

**LIU Xiao-Jing;YIN Xing-Yao;WU Guo-Chen;ZONG Zhao-Yun;School of Geosciences,China University of Petroleum(Huadong);**

Fluid discrimination is an important technology for reservoir characterization in seismic exploration.Despite many years of studies on this tool,many problems remain unsolved.The conventional pre-stack inversion method is implemented based on three-parameter approximation.Usually,the density information is contained in three-parameter inversion. While for deep reservoirs,the maximum angle of incidence is not enough to invert the density information,and the low signal-to-noise ratio(SNR)of the data makes the inversion unstable.With the unstable pre-stack inversion,there will be more uncertainty in fluid identification.In order to stabilize the inversion and improve the accuracy of fluid identification for deep reservoirs,we present an identification method for deep reservoir fluids based on basis pursuit elastic impedance inversion.According to the theory of porous media,the f-μtwo-term AVO approximate equation is derived in terms of Gassmann fluid term(f)and shear modulus(μ).We make a comparison of the reflection coefficients calculated by Zoeppritz equation,f-μ-ρthree-term approximation and f-μtwo-term approximation.As elastic impedance(EI)inversion is more reliable thanAVO/AVA inversion,we derive the two-term elastic impedance equation based on the idea proposed by Connolly.The model constrained basis pursuit inversion is proposed for elastic impedance inversion,which can improve the stability of inversion for elastic impedance.After implementing the inversion for two-angle elastic impedance,we present the fluid factor and shear modulus extraction method by utilizing the two-term elastic impedance equation.We use the well log to confirm the validity of our method.Finally,we apply the method to a real deep reservoir data to identify the fluid saturated in the reservoir.The derived two-parameter equation is close to the Zoeppritz equation when the angle of incidence is less than 20°,and it has almost the same accuracy as f-μ-ρthree-term approximation equation.Therefore,the f-μtwo-parameter approximation can be used for fluid factor estimation.It is clear that the method of fluid factor and shear modulus extraction is stable and the estimates match well with the real model from the model test.The results of real data application match well with the well data interpretation and the blocky inversion results generated by the model constrained basis pursuit method have a good resolving power to the layers.We simplify the f-μ-ρthree-term approximate equation into f-μtwo-term approximate equation which is suitable for deep reservoirs.The accuracy of the new approximate equation is almost the same as the f-μ-ρ approximate equation.The model constrained basis pursuit inversion method yields the blocky layer which has the advantage in interpretation.The tests on synthetic data and field data show that the estimates are reliable and can be used for fluid discrimination for deep reservoirs.

【Fund】：
国家自然科学基金石油化工联合基金重点项目(U1562215);;
国家重点基础研究发展计划(973)项目(2013CB228604);;
国家科技重大专项(2011ZX05030-004-002 2011ZX05006-002 2011ZX05009-003);;
中国博士后基金;;
青岛市博士后启动项目;;
中石化重点实验室基金(WTYJY-WX2014)联合资助

【CateGory Index】： P631.4

【CateGory Index】： P631.4

CAJViewer7.0 supports all the CNKI file formats; AdobeReader only supports the PDF format.