川东地区沙罐坪气田石炭系储集层特征及预测研究  

THE RESERVOIR CHARACTERISTICS AND PREDICATION OF CARBONIFEROUS IN SHAGUANPIN GAS FIELD,EASTERN SICHUAN PROVINCE

作　　者：张瑞[1] 徐国盛[1] 杨斌[1] 周存俭[1] 张萍[1] 

机构地区：[1]成都理工大学"油气藏地质及开发工程"国家重点实验室,四川成都610059

出　　处：《物探化探计算技术》2009年第4期388-392,291,共5页Computing Techniques For Geophysical and Geochemical Exploration

基　　金：中国石油西南油气田分公司2005年科研项目(20050305)

摘　　要：沙罐坪气田石炭系黄龙组主要储集空间为孔隙和裂缝,属于碳酸盐岩低孔低渗型储层,其中,裂缝在改善储层渗透率方面发挥着重要的作用。以测井信息为基础,利用神经网络算法对该区未取芯井储层的孔隙度、渗透率、含水饱和度参数,以及裂缝发育程度进行了预测。使用误差统计法对储层参数预测模型效果进行了评价,其预测效果满足本区所需储层参数计算的精度要求,裂缝预测总体回判率达97.53%。证明了神经网络算法是在测井信息较少的情况下,预测储层的有效手段,为气田评价井、开发井的部署,储量计算及编制气田开发方案,提供了可靠的地质依据。Pore and fracture are principal reservoir space in the Huanglong formation of the Carboniferous in Shaguanping Gas Field,and the Huanglong Formation belongs to low porosity and permeability reservoir but the fracture plays an important role in improving the permeability.Based on the Logging information,the porosity,permeability,aqueous saturation and fractures growth degree of the non-coring well had been forecasted by using the neural network algorithm.Using the method of erroneous statistics to evaluate the results of prediction-model of the reservoir-parameter,the results meet the requirements to calculate the reservoir-parameter.The returns to sentencing rate of the fracture achieve 97.53%,which proved that the neural network algorithm is effective mean of reservoir prediction in the case of less the logging information.The method can provide a reliable basis in geological for the deployment of appraisal wells and development well in the gas fields,calculation of reserves,and establishment development plan of the gas field.

关 键 词：碳酸盐岩 神经网络算法 储层预测 石炭系 沙罐坪气田 

分 类 号：P588.15[天文地球—岩石学]