用时间序列分析方法识别废弃河道

IDENTIFYING ABANDONED CHANNEL BASED ON TIME SERIES

时间序列分析方法是随机数学的一个重要分支,应用到识别废弃河道沉积微相中,将地层垂向测井序列看做是沉积事件在测井相上的反映,每一个测井点都是沉积历史中相应沉积事件的反应,提出用时间序列分析方法中的自回归模型对垂向测井序列进行描述。利用最小二乘的辨识方法对模型进行辨识,用辨识的模型对相邻井相同层段的测井曲线进行预测。将这种方法运用到大庆油田的三次加密井的废弃河道沉积微相的识别中,结果表明:对于位于同一河道砂体的测井曲线,其预测结果和实际值的误差较小;对于废弃河道上的砂体的测井曲线,其砂体上部的预测结果和实际值的误差较大,由此可以识别废弃河道砂体实现。实验结果显示出这种方法能很好地区分废弃河道砂体和河道砂体。

Time series,which is an important branch of stochastic process,is applied to the analysis of sedimentation sequence of rocks.The vertical log profile is treated as time sequence data,so the well logs can be formulated by the auto-regression model of time series.Also,the parameters of the model can be estimated by the least squares estimation with well log data.That means the estimated model has learnt the deposition sequence on the spot of the corresponding well.At last,the model is used for verifying other wells whether they are of the same facies as the first well.This method is applied to identification of abandoned channel facies in Daqing Oilfield.The results show that,to the logging curves of the wells which are located at the same depositional facies,the error between estimated value and practical value is smaller and to the curves of the well located at abandoned channel,the error is greater.The conclusion that time series can be applied on the whether the wells are of the same depositional facies is drawn.