地震多属性分析在煤田拟声波三维数据体预测中的应用

Application of Seismic Multi-Attribute Analysis in Predicting 3D Pseudo Sonic Logging Cube of Coalfield

介绍了利用地震属性分析技术预测三维拟声波测井曲线数据体的方法。对拟声波测井曲线和过井点处的地震属性使用线性回归和神经网络等方法进行了统计,将得到的统计关系应用到非井点处的地震属性上,得到三维拟声波测井曲线数据体,进而研究地层物性参数的变化和空间分布规律。与常规资料相比,三维拟声波测井曲线数据体分辨率有了很大提高,有助于煤层的识别和预测。

The seismic multi-attribute analysis technology to predict 3D pseudo-sonic logging datum was introduced. The input seismic attributes were derived via mathematical transformation of post- and/or pre-stack seismic data. These attributes are characteristics of seismic waves, concerning geometrical, kinematics, dynamic and statistical properties. The seismic multi-attribute analysis technology transforms a series of seismic attributes which possess ambiguous geological meaning only into new attributes whose geological meaning is clear and definite. Then it employs algorithms, such as linear regression and neural network, etc. , to statistically relate the new seismic attributes to the pseudo-sonic logging at well locations. Next, the statistical relation is applied to the attributes so as to get a 3D cube of the pseudo-sonic logging. Finally, the variation and distribution of parameters of the strata are determined via the 3D cube of the pseudo-sonic logging. Compared with the seismic profile, the 3-D cube of pseudo-sonic logging is of higher resolution and can be helpful for coalbed identification and prediction.