LR算法同步挤压广义S变换在隐蔽与复合型河道刻画中的应用

Application of the synchrosqueezing generalized S-transform with theLucy-Richardson algorithm in the characterization of subtle andcomplicated river channels

近年来,在致密油气藏的部分井上钻遇了一些前期未预测出的砂体与泥岩。如何有效识别隐蔽河道并精准钻遇砂岩成为气田增储上产急需解决的突出问题之一。谱分解技术可以从全频地震资料中剥离出反映河道地质特征的特定频段信息,有助于河道精细识别。传统时频分析方法分频结果的分辨率低,难以清晰刻画出河道边界及内幕,为此首次引入LR算法同步挤压广义S变换进行陆相致密砂岩的识别。具体方法是,先利用频谱分析结果及研究区已钻井的录井岩性资料,明确砂泥岩地震响应特征及优势频段,再利用所提出的LR算法同步挤压广义S变换分解出河道特征最明显的单频体进行河道刻画。数值模拟结果表明,同步挤压算法极大地提高了广义S变换时频分辨率,较好地区分了非平稳信号中不同频率的信号分量,更适用于实际地震信号的时频谱分解。四川盆地东南缘涪陵区块实际地震资料应用结果表明,所采用的谱分解方法精确刻画出了研究区“隐蔽”河道展布及复杂河道带内幕,分频结果与相似研究区曲流河道沉积模式匹配,证明了方法的有效性和优越性。

Some wells deployed for tight oil and gas in the Sichuan Basin have penetrated the sandstones and mudstones that were not predicted.To ensure reserves and production increase,the point is how to identify subtle channel sands accurately and design the horizontal wellbore track exactly penetrating target sands.Spectral decomposition is a useful technique,as it could extract the information at specific frequencies for channel reservoir characterization.However,traditional time-frequency analysis methods suffer from low resolution,making it difficult to detect channel boundaries and internal details clearly.Therefore,we introduce the synchrosqueezing generalized S-transform with the Lucy-Richardson algorithm for seismic time-frequency spectral decomposition.We use spectral analysis and lithologic log data to clarify seismic responses and the dominant frequency band of sands,and then use the algorithm in this paper to extract the single-frequency signals with the most obvious characteristics of river channels.Numerical simulation shows that the synchrosqueezing algorithm greatly improves the time-frequency resolution of the generalized S-transform to distinguish among different frequency components in non-stationary signals,and hence it is more suitable for the time-frequency spectral decomposition of actual seismic signals.The practical application shows that our method could accurately predict the distribution of subtle channels and the internal details of complicated channels in the study area.The results of frequency division are consistent with the sedimentation model of meandering channels in similar areas,which confirms the effectiveness and superiority of the method.