钻井工况智能识别与时效分析技术  

Intelligent identification and time-efficiency analysis of drilling operation conditions

作　　者：胡志强 杨进[3] 王磊[1,2,3] 侯绪田[1,2] 张桢翔 姜萌磊[3] HU Zhiqiang;YANG Jin;WANG Lei;HOU Xutian;ZHANG Zhenxiang;JIANG Menglei(SINOPEC Research Institute of Petroleum Engineering,Beiing 102206,China;State Key Laboratory of Shale Oil and Gas Enrichment Mechanisms and Efective Development,Bejing 102206,China;College of Safety and Ocean Engineering,China University of Petroleum(Beijing),Beijing 102249,China)

机构地区：[1]中国石油化工股份有限公司石油工程技术研究,北京102206 [2]页岩油气富集机理与有效开发国家重点试验室 [3]中国石油大学(北京)安全与海洋工程学院

出　　处：《石油钻采工艺》2022年第2期241-246,共6页Oil Drilling & Production Technology

基　　金：国家自然科学基金联合基金“高温高压油气安全高效钻完井工程基础理论与方法”(编号:U19B6003-05);中国博士后科学基金“高温高压气井油套环空带压泄露点关键参数地面诊断技术研究”(编号:2020M670584);中国石化科技攻关项目“尼日利亚、喀麦隆复杂地层钻井关键技术研究”(编号:PE19009-1)。

摘　　要：目前钻井作业工况识别和钻井时效分析主要依赖于现场仪器的传输效率和工程作业人员的经验诊断,存在无法处理大量实时施工数据、决策反馈机制慢、预测精度低等问题。为高效利用综合录井数据辅助工程人员进行优化决策,根据钻井过程记录的大量录井数据,开发了基于WITS标准和WITSML标准的数据传输模块,创建了将阈值法和神经网络法相结合的融合算法模型,建立包含井场信息数据和单井工况识别结果的历史数据表,编制软件对录井历史数据进行时效分析。研究结果显示,案例井的钻井工况智能识别与实际工况基本符合,预测精度大于90%,钻井时效统计误差小于1%,应用效果良好。该研究有效地提高了钻井工况识别和钻井时效分析的效率,对现场具有指导作用。Currently,the operation condition identification and time-efficiency analysis for drilling are mostly dependent on the data transfer efficiency of on-site devices and empirical diagnosis of engineering operators,which suffer from the incapability of handling massive real-time operation data,slow decision making-feedback mechanism,and low prediction accuracy.To efficiently assist the optimization of decision-making of engineering personnel using mud logging data,the data transfer module based on the WITS and WITSML standards was developed,the algorithm integrating the threshold method and neural network method was constructed,the historic data sheet including wellsite information and operation condition identification results of an individual well was tabulated,and the time-efficiency analysis software based on mud logging historic data was programmed.The research showed that the intelligent drilling operation condition identification results of the case-study well are consistent with the actual operation conditions,with the prediction accuracy over 90% and calculation error less than 1% for the drilling time-efficiency,and the application performance is highly satisfactory.This research effectively improves the efficiency of identifying drilling operation conditions and analyzing drilling time-efficiency,and provides guidance for drilling practice.

关 键 词：钻井工况 时效分析 人工智能 神经网络 录井数据 数据传输 

分 类 号：TE21[石油与天然气工程—油气井工程]