测井数据分维计算及其应用条件研究

分形几何已成功地应用于储层表征，分维是储层一征和随机建模的控制参数，为了准确表征储层，必须精确地估计分维值。本文介绍了利用测井数据计算储层分维值的三种方法：变尺度分析法（Ｒ／Ｓ分析），变异函数分析法、功率谱分析法，研究发现分维计算中对数据的分级处理实际上就是对数据进行灰色累加生成。研究表明影响分维估计精度的主要因素有四种：数据点数、数据概率分布、数据平稳性，无标度区，研究认为，现有的３种方法不能用

STUDY ON THE APPLICATION CONDITION OF ESTIMATING FRACTAL DIMENSION OF LOG DATA

Fractal geometry has been used to reservoir characterization successful.Fracral dimension is a control parameter of reservoir and random modeling.To characterize reservoir accuracy,fractal dimension must be estimated accuracy.This paper introduces three methods calculating fractal dimension of log date:rescaled analysis(R/S),variogram analysis,power spectral analysis. This study finds that reservoir parameter or vertical log data is a typical gray system,and actually graded sequence operation is grey accumulated generating (AGO) operation of data. Main factors influencing precision of fractal dimension computation include number of data,distribution of data,stationarity of data,no scale area. Three methods computing fractal dimension of log data can not been used to thin reservoir. A reasonable procedure computing fractal dimension of log data has been given.