神经网络和分形几何方法在识别测井沉积微相中的应用

本文综合应用人工神经网络和分形几何等最新的模式识别数学方法，遵循地质家的思维方式，进行了测井资料沉积微相解释方法研究；并在工作站上开发了人机联作解释软件。将工作站系统在运算速度、人机交互及图件制作方面的优势与解释人员的经验和推理判断能力有机地结合起来，为测井资料沉积微相解释提供了新的技术手段和良好的计算机辅助工具。将其应用于辽河油田长北地区，对１８口井的沙海组上段～阜新组下段进行了沉积微相连续解释，取得了良好效果。

Application of the Neural Network and Fractal Geometry in Sedimentary Environment Study from Well Logs

Making use of new mathematical methods such as artificial neural network, fractal geometry andpattern identification, following a geologist' s mode of thinking, the research to interpret microfacies fromthe log data was carried out A software of human-computer interactive interpretation was developed un-der workstation. Combined the superiority of workstation in calculating speed, human-computer interac-tive ability and drawing with the interpreter' s experience and judgement, it provided new technical meansand admirable computer assistant tools for the interpretation of microfacies from the log data. The princi-pal ideas were presented as follows.