基于遗传神经网络模型的夹层识别方法研究

Study on interlayer recognition method based on genetic neural network model

由于夹层的存在影响剩余油的分布,认清夹层分布对制定开发计划至关重要,故提出了一种夹层识别的新方法。分析了夹层形成原因及其测井曲线特征,从沉积上来说夹层是沉积间歇面,用砂泥比和泥砂比差值来反映沉积环境。采取滤波构造新的曲线,通过新曲线确定每个小层的测井曲线砂岩值和泥岩值;通过各个突变点的测井曲线相对变化值来反映测井曲线特征的变化,应用遗传神经网络对每个突变点的沉积环境以及测井曲线相对变化值进行分类,从而判定夹层以及夹层的类型。

Interlayer affects distribution of residual oil,understanding interlayer distribution is crucial to make development plan,so a new method of interlayer identified was proposed. Interlayer causes and characteristics of logging curves were analyzed. From the deposition,interlayer is a deposition intermittent surface. Deposition environment is reflected by the difference between shale ratio of sand and sand ratio of shale. Adopting filter construct new curve. The log value of sand and shale for each small layer is determined by the new curve. The relative change in the value of logs of various mutation points use to reflect the log feature change. By applying genetic neural network,each mutation point was classified by sedimentary environment and logs relative change,thereby interlayer and the type of interlayer was determined.