随钻地质导向智能决策的实现与应用

地质导向决策是随钻地质导向技术的核心,为避免人为因素造成决策结果差别并最终影响施工效果,基于典型的在储层中部钻进、沿储层顶部钻进及追踪储层甜点等3种地质导向场景类型,收集并整理了60口井的3 000条决策样本数据进行神经网络训练,搭建智能决策平台实现随钻地质导向智能决策。训练结果表明,智能决策平台生成的导向决策准确率达到89.7%,满足地质导向实时决策分析的施工要求,并可以对已钻井的地质导向施工质量进行定量分析。该研究提高了地质导向实时决策分析的准确性以及决策效率,可优化井眼轨迹并降低钻井风险,为实现随钻地质导向技术智能化转型和加强地质工程一体化开发提供参考。

contents

Geosteering decision making is the core of geosteering while drilling. In order to prevent artificial factors from leading to deviated decisions and eventually impacting the construction results, this paper collected and sorted out 3 000 decision-making sampled data of 60 wells based on three typical geosteering scenarios (drilling in the middle of the reservoir, drilling along the top of the reservoir and tracing the sweet spot) for neural network training. Then, the intelligent decision making platform was built up for the realization of intelligent geosteering decision making while drilling. The training results show that the steering decision made by the intelligent decision making platform has the accuracy rate of 89.7%, which satisfies the construction requirements of real-time geosteering decision making analysis, and it can be used to quantitatively analyze the geosteering construction quality of drilled wells. The research results improve the accuracy of real-time geosteering decision making analysis and the decision making efficiency, can optimize hole trajectories and reduce drilling risks, and provide the reference for the intelligentization of geosteering while drilling and the enhancement of geology-engineering integration development.