摘要
通过已有的少数井的试井资料分析得出压裂裂缝参数,以现有的参数为样本建立人工神经网络系统.以影响压裂结果的地层厚度、孔隙度、泥质含量、压裂施工参数、工作压力加砂排量为输入参数,以裂缝导流能力和裂缝半长为输出参数,用BP神经网络训练,推断出所有井的压裂裂缝参数,从而得到整个油藏的压裂裂缝分布特征,对压裂措施的效果有了直观的评价.
The fracture parameters are obtained through the well test data from a small number of wells. With the existing parameters as samples, the artificial neural network system is set up. The input parameters that influence the outcome of fracturing, such as the formation thickness, porosity, clay content, working stress and sand displacement, are selected. The ability to fracture conductivity and fracture half-length are the output parameters. With training by BP neural networks, the fracturing parameters of the whole wells are inferred, resulting in the fracturing distribution of the reservoir
出处
《地质与资源》
CAS
2009年第3期217-221,共5页
Geology and Resources
关键词
人工神经网络
试井分析
裂缝
artificial neural network
well test analysis
fracture
