薄砂体预测的地震沉积学研究方法

立足砂泥岩薄互层地质背景下厚度小于λ/4（λ为地震子波波长）的单砂体地震储层预测，按照精细地层划分→砂体平面形态分析→砂体厚度预测的研究流程，提出地震沉积学研究思路。除曾洪流等倡导的地震沉积学研究规范涉及的井-震高精度层序格架、地震岩性分析（地震子波相位调整）、地层切片制作等核心研究内容之外，认为以细分岩性为基础的压实校正及古地貌恢复、邻层干涉压制、井-震联动的地层切片浏览、非线性地层切片等技术可准确预测薄砂体平面形态。在精细地层划分方面，地震同相轴的等时性分析技术有助于优选与地质等时界面吻合的同相轴作为层序界面，而基于精细合成记录制作和时深转换的井-震精细对比，可以实现高级层序界面的识别和追踪解释。薄砂层厚度预测方面，除常用的振幅-厚度分析技术和峰值频率技术之外，认为振幅-频率融合和遗传化神经网络技术等综合地震属性预测方法同样可以实现对薄砂体厚度的半定量或定量预测。该方法对准确刻画薄互层中的单砂体具有指导意义。

Thin bed prediction from interbeded background: Revised seismic sedimentological method

Predicting a thin bed (< λ/4 in thickness, where λ is the length of waveform)from the interbedded background is a challenging work for the seismic interpreters. According to the working order from strata division through plane view to thickness prediction, a workflow was proposed by using seismic sedimentological method. Besides the high-order sequence dividing, seismic lithology analyzing (90° degree phasing)and the strata slicing proposed by Zeng Hongliu in their workflow, we emphasized the importance of the following techniques in promoting the accuracy of the thin bed plan-view prediction:(1)palaeogeomorphology recovering based on compaction correction with concern of different lithology; (2)interference suppressing of the neighboring beds; (3)browsing of strata slices linked with well-logs or drilling column; (4)non-linear strata slicing. When coming to the high-order sequence recognition, the isochronism analyzing of the seismic events and the well-seismic matching were recommended. The isochronism analyzing could help us to find the seismic reflections in accordance with the geological surfaces, and the well-seismic matching is useful for high-order sequence recognition. When concerning the thin bed thickness prediction, we firstly introduced two commonly used techniques, amplitude tunning and peak frequency, then proposed amplitude-frequency blending and genetic neural network as two new valuable techniques.