基于敏感地震属性波形分类的流体预测研究

系统对比了针对储层流体的现有各种识别因子识别含气和含水砂岩的能力，结果显示，高灵敏度识别因子 有较强的综合识别能力。以S 盆地L 研究区为例，在主要含气不等厚层段内对高灵敏度流体因子在神经网络模式和 PCA 神经网络模式两种情况下基于形态学进行分类，效果对比明显，结合井解释资料对研究区进行了初步的流体预 测。将波形分类法的应用范围从主要是进行地震相-沉积相分析、储层预测等扩展到对流体因子在目标层段进行分析 处理从而进行层段内流体预测，对仅在剖面和切片上利用流体因子来进行流体识别和预测是一种新的有价值的补充。 首次探讨了将波形分类技术应用于流体预测，认为该方法与其他解释成果结合对于有效降低勘探风险、提高钻井成功 率具有一定的指导意义。

A Study on Fluid Prediction Based on the Classification of Sensitive#br# Seismic Attributes

The ability to identify gas-bearing and water-bearing sands of nine fluid identification factors has been compared， and the results show that the High-Sensitivity-Fluid-Identification-Factor has a strong ability to identify. The neural network and the Principal-Component-Analysis-neural-network are applied to high-quality 3-D data of HSFIF to perform the waveform analysis during the gas bearing interval in study area L of basin-S and good mapping effect has been achieved. The facies maps were analyzed and compared with the logging′interpretation. It proves the application of the PCA-neural network method can greatly reduce the difficulty of seismic facies interpretation of the map. In this paper，the application range of waveform classification is extended from seismic-sedimentary-facies analysis and reservoir prediction to the analysis and processing of the fluid factors in the target layers，thereby we can predict the fluid in layers. It is a new kind of valuable complement to fluid identification and prediction by using fluid factor only in profile and slice. It is the first time that the waveform classification techniques has been applied to fluid prediction. We believe that the method along with the results of other explanations has a guiding significance to reduce exploration risk and enhance drilling success rate.