近地表结构特征属性平衡迭代规约和层次聚类分析

地震反射信号频宽变窄的主要因素是地层吸收衰减,其中表层吸收衰减尤为严重,是地震信号分辨率降低的主要因素之一。以华北油田冀中探区近10000km2的三维地震工区为研究目标,结合微测井资料处理成果,以雷克子波穿透低、降速带为例,分析了低、降速带厚度、速度以及Q值等属性参数对地震信号吸收衰减的影响,构建了基于上述参数的平均衰减因子,并基于该因子将近地表结构划分为3类,讨论了不同类别近地表条件对地震信号的能量、频谱特征、分辨率和信噪比的影响。提出了基于机器学习的多属性平衡迭代规约和层次聚类分析方法,利用低、降速带多个特征属性参数构成特征向量,对近地表结构特征进行聚类分析,将华北油田冀中探区近地表结构分成9类,为野外采集参数的合理选取和观测系统设计提供了依据,也为后续地震资料处理提供了帮助。

Previous research results showed that absorption is the major factor narrowing the bandwidth of the reflection signal.Near-surface absorption,in particular,is one of the primary causes of low-resolution in seismic data.In this study,a three-dimensional seismic survey area of nearly 10,000km2 in the Jizhong exploration area of the North China oil field is investigated.By using processed uphole data and taking the Ricker wavelet penetrating the weathering zones as an example,the influence of attributes,such as thickness,velocity,and Q-value,of weathering zones on the absorption and attenuation of the seismic signal is analyzed.An average attenuation factor is constructed,which permits to classify the near-surface structures into several classes.The relationships between near-surface classes and energy,spectral characteristics,resolution,and signal-to-noise ratio of seismic signals are discussed.A balanced iterative reduction and clustering using a hierarchical analysis method is proposed.Using several attributes of the weathering zone,the eigenvector is constructed,based on which the clustering analysis of the near-surface structural characteristics is conducted.According to the proposed method,the near-surface structure of the Jizhong exploration area was divided into nine classes.The results of this study can aid in the selection of appropriate field acquisition parameters and in the design of the observation system.Moreover,they can facilitate the subsequent seismic data processing and interpretation.

中国石油天然气股份有限公司科技攻关课题“河套盆地新区新领域勘探潜力与高效勘探关键技术研究”(2019D-0815)和国家科技重大专项“致密气有效储层预测技术”(2016ZX05047002)共同资助。