小波变换与Hilbert-Huang变换应用于层序划分的比较

Comparative study of the application of wavelet transform and Hilbert-Huang transform to the demarcation of sequences

测井资料应用于层序地层研究的实例很多,一般选取能够敏感反映沉积过程的GR或AC曲线开展研究,且过去多注重测井曲线形态及幅值特征的应用。近年来学者们倡导了将测井数据小波变换和Hilbert-Huang变换分别应用于层序地层划分的新方法,但并未对这2种方法进行过对比分析。笔者简要介绍了小波变换与Hilbert-Huang变换的原理,通过对所构建的仿真分段信号的处理分析,展示了2种方法在识别信号中的突变点与突变区域方面的有效性,结合实例分析了2种方法在层序地层划分应用中的差异性。结果表明:小波变换更宜于识别测井信号中的多尺度旋回性,适于不同级别层序的划分;而Hilbert-Huang变换更宜于识别测井信号中的突变性,对信号中突变点位置的识别更加准确,适于确定层序的界面。将2种方法相结合,把济阳坳陷惠民凹陷钱斜14井沙河街组划分为3个长周期基准面旋回、7个中周期基准面旋回和23个短周期旋回,与前人传统方法研究结果具有一致性,展现了组合方法的优越性。

Abstract：There are many case studies concerning the application of well logs to the study of sequence stratigraphy. Normally, Gamma Ray（GR） or Acoustic curves （AC） are used for the study because they can sensitively reflect the sedimentary process. Also, in the past, the studies usually focused on the application of the shape and amplitude of log- ging curves. In recent years, Wavelet transform and Hilbert-Huang transform of logging data were introduced to the de- marcation of sequence stratigraphy respectively, which triggered the study interests of many scholars in the area. How- ever, no comparative study of the two methodolgies has been reported up to now. After the brief introduction on the principles of Wavelet transform and Hilbert-Huang transform, this paper presented the validity of the two methods in the recognition of mutational site and region of the simulated segmented signal constructed. In addition, the difference between the applications of Wavelet transform and Hilbert-Huang transform to the demarcation of sequence stratigraphy was analyzed with a case study. The results show that Wavelet transform is more suitable for identifying multi-scale cy- cles in logging signals and good at demarcating different orders of sequences while Hilbert-Huang transform is suitable for identifying mutagenicity in logging signals and determining sequence boundaries because of its accuracy in locating the mutational site in well logs. With the combination of the two methods, the Shahejie formation of well Qianxie 14 in Jiyang depression Huimin sag was classified into three long base-level cycles, seven middle base level cycles and 23 short-term cycles. The result is consistent with the geological methods, exhibiting the advantage of the combination methodology. Key words ：wavelet transform ; Hilbert-Huang transform ; demarcation of sequences