论文部分内容阅读
本文借鉴语音识别技术中的线性预测倒谱系数(LPCC系数)特征参数提取方法对地震数据进行分解,这种方法的优点是:可以获得将子波和反射系数信息分离的地震语音特征参数,对地质现象边界具有较好的描述能力,使我们可以从不同维度更细致地观察隐藏在地震数据中的地质特征.理论模型分析表明,基于LPCC系数的地震分析具有较高的地震相划分能力.实际地震资料应用表明,LPCC系数对储层特征的描述比常规三瞬属性更为细致,不同阶次LPCC系数在描述储层不同特征时也保持了内在的联系.采用K均值聚类方法对提取的12阶和24阶LPCC系数进行聚类分析,聚类结果与目的层段古地形较为吻合,较好地反映了研究区的断裂、礁滩相带、深水扇和储层的分布特征,说明在地震相分析中采用LPCC系数作为特征参数是可行和有效的.
In this paper, we use the linear prediction cepstrum coefficient (LPCC coefficient) feature extraction method in speech recognition technology to decompose the seismic data. The advantage of this method is that we can get the seismic speech feature parameters that separate the wavelet and reflection coefficient information, The boundary of geological phenomenon has good descriptive ability, so that we can observe the geological characteristics hidden in the seismic data more carefully from different dimensions.The theoretical model analysis shows that the seismic analysis based on the LPCC coefficient has higher seismic facies division ability. The application of seismic data shows that the LPCC coefficients describe the reservoir characteristics more carefully than the conventional three-moment attributes, and the LPCC coefficients of different orders maintain intrinsic relationships when describing different reservoir characteristics.Using K-means clustering method, 12th, and 24th order LPCC coefficients. The clustering results are in good agreement with the paleogeomorphology of the target interval, reflecting well the distribution of faults, reef banks, deep-water fans and reservoirs in the study area, It is feasible and effective to use LPCC coefficients as characteristic parameters in seismic facies analysis.