School of Energy,China University of Geosciences( Beijing) China United Coalbed Methane Corporation Limited
Engineering and Technology Branch,CNOOC Energy Development Corporation Limited

为了解决试井资料有限、无法完全掌握煤储层渗透率分布的难题,提出了把测井信息作为煤储层渗透率的影响因素,并据储层渗透率与已知测井信息的关系建立BP神经网络模型,在一定的学习前提下对未知样本进行预测。综合应用BP神经网络模型和模糊排队算法,以柳林区块56口井的试井和测井资料为基础,以模糊曲线分析方法优选输入变量,建立BP神经网络煤层渗透率预测模型,并对渗透率进行了预测。预测结果与实测结果之间具有较高的吻合度,结果表明该方法对煤储层渗透率预测具有较好的适用性。对柳林区块煤层渗透率分布进行预测,摸清了全区渗透率平面展布的样式,指明柳林区块东北部渗透率较高,是勘探开发的有利区。

In order to solve the limited information of the test well and not to fully hold the difficult permeability distribution of the coal reservoir,the logging information was provided to be the factors affected to the permeability of the coal reservoir. With the relationship between the reservoir permeability and the known logging information, a BP neural network model was established. On the premise of certain study,a prediction method was conducted on the unknown samples. With comprehensive application of the BP neural network model and the fuzzy queuing algorithm,based on 56 test wells and logging well information in Liulin Block,with a fuzzy curve analysis method was applied to optimize the input variation,a prediction model of the seam permeability was established with the BP neural network and a prediction was conducted on the perm eability. There was a high ftting between the predicted results and the measured results and the results showed that the method could have a good adaptbility to predict t he permeability of coal reservoir. With the prediction on the permeability distribution of seam in Liulin Block,the plane layout mode of the permeability in the full block was C leared. The paper pointed out that the permeability in the northeast part of Liulin Block was high and the northeast part could be a favorable block of exploration and devel opment.

国家科技重大专项资助项目(2011ZX05038-002；2011ZX05062-01)；