摘要
低渗砂岩具有低孔、低渗特征,气体渗流过程存在启动压力梯度。目前,低渗储层启动压力梯度测定方法尚不统一,且定量描述方面研究较为欠缺。采用渗流法和气泡法各对一组低渗岩心进行了启动压力梯度测试,并对测试结果进行了对比分析。运用多元回归方法,提出了一种考虑多因素的启动压力梯度预测模型。利用预测模型建立了某低渗区块启动压力梯度经验公式,并通过另一组岩心实测值进行了公式验证。结果表明:启动压力梯度分别与岩心渗透率和含水饱和度呈幂函数关系,其值随渗透率的增加而减小,随含水饱和度的增加而增加;气泡法启动压力梯度测试值较渗流法测试值大,且差值随渗透率的增加而减小,随含水饱和度的增加而增加;含水饱和度对气泡法测试结果影响更为明显;提出的模型预测精度较高,对启动压力梯度定量描述具有一定的参考价值。
Because of the severe heterogeneity of low-permeability gas reservoirs, there is threshold pressure gradient（TPG） in some of these reservoirs. At present, methods for determining TPG in low-permeability gas reservoirs have not been unified. Multiple regression equations that describe TPG quantitatively are not available. In this article, tests of threshold pressure gradient were conducted on typical cores of low-permeability gas reservoirs by both percolation method and bubble method. Test results of the two experimental methods were analyzed and compared. Using multiple regressions, a regression model for determining TPG with permeability and water saturation was established. A field case study was performed to validate the established model. Results indicate that TPG presents power function relationship with permeability or water saturation respectively, and TPG increases with the decrease of permeability or increase of water saturation. TPG values by bubble method are significantly higher than those of percolation method, and the differences increase with the decrease of permeability or increase of water saturation. Bubble method is more sensitive to change of water saturation. The predicted TPG values from the proposed equation have good agreement with the experimental results in the case study, which can provide a reference for determining TPG quantitatively.
出处
《断块油气田》
CAS
北大核心
2016年第5期610-614,共5页
Fault-Block Oil & Gas Field
基金
国家自然科学基金委联合项目"煤层气在基质孔隙中解吸传递机理及产能预测"(U1262113)
国家自然科学基金委青年科学基金项目"页岩气扩散渗流机理及产气规律研究"(51504269)
中国石油大学(北京)科研基金资助项目"煤层气藏开发扩散渗流机理及产气规律研究"(YJRC-2013-37)
关键词
低渗气藏
启动压力梯度
实验研究
多元回归
预测模型
low-permeability gas reservoir
threshold pressure gradient
experimental study
regression equation
prediction model
