Application Of Imaging Logging Data In Identification Of Fracture

The evaluation of sandstone reservoirs fractures in Anpeng deep zone haslong been difficult in logging data interpreta tion. The conventiona l logs havepla yed very important role in identifying fractures. Because of heterogeneity andanisotropy of fractures reservoirs have resulted in multi-resolution and nondefinition.Now the occurrence of imaging logging leads directly and effectivelyrecognizing fractures to reality.The following studies are carried out on the basis of acoustoelectric andcrossed dipole acoustic imaging logging and integrating with core, conventiona llogging toward the reservoirs fractures of Anpeng deep zone.1.The paper studies the analysis of factors influencing of acoustoelectricimaging logging and builds a model to expla in geological features and gets themethod to evaluate the authenticity, valid ity, filling and development occurrenceof fractures.2.On the basis of conventiona l logs and crossed dipole acoustic logs data,according to the fracture characteristics of the region, it gets ma ny fractureinstructions parameters and establishes a comprehensive probabilistic model.3.The paper combines dual laterolog and imaging logging to establish asuitable model of fracture parameters in this region. The use of spectral analysismethods achieve effective porosity Quantitative evaluation of reservoir of a double porosity.4.The effective ima ge segmentation algorithm gets rid of the backgroundvalues in the imaging map which have nothing to do with the fractures. Thefractures were automa tic identified and got the parameters to use the HOUGHtransformation.5.On the basis of comprehensive selecting conventiona l and imaging logsdata, The paper establishes a fracture identifica tion model to use the BP neura lnetwork.The comprehensive fractures probability model which synthesizes theconventiona l and crossed dipole acoustic imaging logs data modifies thedeficiency of conventiona l comprehensive model and be able to preferablequalitative indicate development state of fractures. It's result has a betterconformability with core and acoustoelectric ima ging. The fractures parametersof ten wells were computed by the established model have comparativelyapproxima te to parameters of core. It shows that the accuracy to meet productionrequirements. The establishment neura l network model of fracture identifica tionhas better processed results. It can resolve the present situation that other wellshave less specia l data and fracture identifica tion has comparatively difficulties.