水平井钻井提速-减阻-清屑多目标协同优化方法  

Drag Reduction and Hole Cleaning in Horizontal Well Drilling

作　　者：丁建新 李雪松 宋先知[2,3] 张诚恺 马宝东 刘子豪[2,3] 祝兆鹏 Ding Jianxin;Li Xuesong;Song Xianzhi;Zhang Chengkai;Ma Baodong;Liu Zihao;Zhu Zhaopeng(Kunlun Digital Intelligence Technology Co.,Ltd.;National Key Laboratory of Petroleum Resources and Engineering,China University of Petroleum(Beijing);College of Petroleum Engineering,China University of Petroleum(Beijing))

机构地区：[1]昆仑数智科技有限责任公司 [2]中国石油大学(北京)油气资源与工程全国重点实验室 [3]中国石油大学(北京)石油工程学院

出　　处：《石油机械》2023年第11期1-10,共10页China Petroleum Machinery

基　　金：国家重点研发计划项目“变革性技术关键科学问题”(2019YFA0708300);中国石油天然气集团有限公司-中国石油大学(北京)战略合作科技专项“物探、测井、钻完井人工智能理论与应用场景关键技术研究”(ZLZX2020-03);国家杰出青年科学基金项目“油气井流体力学与工程”(No.52125401);中国石油大学(北京)科学基金项目“深井气侵自动压井井底压力智能预测和控制方法”(2462022SZBH002)。

摘　　要：水平井钻井过程中,钻进破岩、管柱延伸和井筒流动等子系统相互耦合、相互制约,现有的仅考虑单一子系统的钻井优化方法难以兼顾效率和风险。针对此问题,提出了综合考虑破岩提速、管柱减阻和井筒清洁多过程耦合的水平井钻井提速-减阻-清屑多目标协同优化方法,建立了基于随机森林的机械钻速智能预测模型、基于刚杆模型的摩阻系数智能反演模型和基于两层稳态模型的水平井岩屑运移模型,并以钻压、转速和排量为决策变量,以提高机械钻速、降低机械比能为目标,以不发生管柱风险(摩阻系数过大)和井眼清洁风险(岩屑床高度过高)为约束条件,采用带精英策略的非支配排序遗传算法和逼近理想解排序算法进行多目标优化与决策,给出钻压、转速和排量的最优参数组合。将该方法应用于某水平井7600~7650 m水平段进行实例分析。分析结果表明,优化后机械钻速提高了32%,机械比能降低了19%,且摩阻系数和岩屑床高度均在安全范围内。提出的水平井钻井提速-减阻-清屑多目标协同优化方法能够在优化钻井效率的同时降低卡钻和井眼清洁不足等风险。该方法可为水平井安全高效钻进提供保障。During horizontal well drilling,subsystems such as rock breaking,string extension and wellbore flow are coupled and constrained each other.Existing optimization methods only consider a single subsystem,making it difficult to balance efficiency and risk.To solve this problem,first,a multi-objective collaborative optimization method for rate of penetration(ROP)improvement,drag reduction and hole cleaning in horizontal well drilling that has comprehensively considered the coupling of multiple processes such as rock breaking speed increase,drag reduction of pipe string and wellbore cleaning was proposed;second,an intelligent prediction model for ROP based on random forest,an intelligent inversion model for friction coefficient based on rigid rod model,and a horizontal well cuttings transport model based on two-layer steady-state model were built;third,taking weight on bit(WOB),rotation speed and displacement as decision variables,the improvement of ROP and the reduction of mechanical specific energy as the objectives,and avoiding the risk of pipe string(too high friction coefficient)and wellbore cleaning(too high cuttings bed height)as the constraint conditions,a non-dominant sorting genetic algorithm with elite strategy and an approximating ideal solution sorting algorithm were used to conduct multi-objective optimization and decision-making,and give the optimal parameter combination of WOB,rotation speed and displacement;finally,this method was applied to the case analysis of 7600 to 7650 m horizontal section in a horizontal well.The results show that after optimization,the ROP is improved by 32%,the mechanical specific energy is reduced by 19%,and the friction coefficient and cuttings bed height are within the safe range;and the proposed multi-objective collaborative optimization method for ROP improvement,drag reduction and hole cleaning in horizontal well drilling can reduce risks such as sticking and insufficient wellbore cleaning while optimizing drilling efficiency.This method provides guarantee for safe and

关 键 词：钻井优化 提速-减阻-清屑 多过程协同 多目标优化 机器学习 

分 类 号：TE319[石油与天然气工程—油气田开发工程]