Fracture Characteristics,Identification And Evaluation Of The Distribution For Deep Tight Sandstone Gas Pool

Natural fractures as a structure are common in tight sandstone reservoirs， especially inthe strong tectonic deformation or regional fault systems development area. Tight sandstonereservoirs usually presents low porosity and permeability, but the development level ofnatural fracture systems can greatly improve the capacity of reservoir and interstifial flow. Inthis paper, the second member of Xujiahe formation in Xinchang gas field is deep tightsandstone gas reservoirs，in4700~5200m depth. Years of exploration and development andpractice have confirmed the fracture system is the key geological factors to the high-yieldgas reservoir. In this paper, fracture systems and faults associated fracture systems are themain object of study, by carring out the study of fracture characterization, fractureidentification, cause and the prediction and evaluation of the distribution. And it set up asuitable comprehensive fracture recognition standards and fracture prediction methods fordeep tight sandstone gas reservoirs，which can ensure the accuracy of fracture identificationand fracture prediction.According to a large number of core observation, thin and imaging logging analysis,summed up the fracture characteristics,and it shows that the fracture is high-angle and higheffective.The degree of effectiveness in the sandstone is above the shale, and fracture canextend longitudinally from0.1m to5m lengths.The length of fracture can be maximum to10m and it can be seen that fracture run through layers which is common phenomenon. Thefracture in E-W is the most developed cracks，followed by the NE and approximate EW,while the NW-trending fractures developed relatively weak. The fracture in EW is consistentwith the current maximum principal stress orientation. The large tensile breaking fracturesare developed in sandstone, while the shear fractures are mainly visible in mudstone. Thefracture width and porosity are explained by conventional logging data, and the greater thecrack width and porosity,the fracture is more developed on core or imaging logging. Using14wells imaging logging data, fracture identification based on multi-methods (linear andnon-linear),like BP and PNN neural networks, K neighboring nodes, support vectormachine method and stepwise discriminant analysis.The conventional logging comprehensive identification result is consistent with core fractures by92.86%, andconsistent with imaging logging by76.11%. In the end,we establish a comprehensivemulti-parameter fracture identification standards.Integrating acoustic emission testing data, fracture fillings stable carbon and oxygenisotope testing, fracture filling inclusions tests and other information, it shows that fractureformation stage exist at least four times, that is, Indosinian fracturing (relatively weak),Yanshan fracturing (also relatively weak) and Himalayan fracturing (faults, joints andregional tectonic fractures are mainly formed during this period). Integrating core, logfracture identification, testing data analysis, it thinks that the controlling factor fractures arebroken, thickness of layer and structural deformation, while the lithology’s influence isrelatively weak.Acordding to fault-controlled fracture growth mechanism,we evaluate thefault-controlled fracture distribution density,and the evaluation results are consistent withcore, conventional logs and image logging. Evaluation methods of seismic data includegeneral properties, P-wave azimuth property and3D3C data, in which fracture coherentwave splitting+forecast is much better.Using acoustic emission test data, consideringvertical stratification and flat lithology changes,we simulate ancient stress use with FlAC3D,and the stress simulation results show that the distribution of stress field is obviouslycontrolled by faults. Finally, in considering the lithology, thickness, faults and tectonicdeformation,and combinating logging and seismic fracture prediction results,we predictefracture in comprehensive evaluation methods.The predicting results shows that TX22relatively developed,and the results coincide with the production test data at the rate of90.91%;while the fracture in TX27is relatively less developedand its results coincide withthe production test data at the rate of75.00%.The research results are strongly targeted,and highly fit with the production anddrilling.Fracture identification and prediction results has highly reliability.This study providestrong geological evidence for reservoir construction renovation and preferably favorableareas. The fracture comprehensive recognition standards and fracture prediction system fordeep tight sandstone gas reservoirs,is a good reference for the same type of gas field homeand abroad.