摘要
在硬地层钻进过程中时常发生钻头与地层不匹配导致钻进速度缓慢、增加对钻头的磨损,选择与地层相合适的钻头是降低钻井成本、提高钻进速度的关键所在。利用主成分分析(PCA)将邻井测井参数(自然伽玛、井径、自然电位等)作为影响因子进行标准化处理、主成分降维分析等技术手段,建立了硬地层岩石可钻性级值的数学模型。结果表明,声波时差、地层密度、自然伽玛和电阻率与岩石可钻性级值关联度最为密切,经验表明采用多元对数线性方式回归所得到的硬地层岩石可钻性级值预测值与实验值精准度达到95%以上;采用该方法筛选出来的钻头在风城组地层实际钻进中,在保证机械钻速的同时,钻头进尺长度由原来的30 m提升至150 m,减少了钻头更换等非作业时间。
In the process of drilling through a hard formation,use of a bit unsuitable for the formation often leads to slow drilling speed and increased bit wear.Selecting a bit suitable for a formation is the key to reduce drilling cost and increase the rate of penetration.By applying such technical measures as principal component analysis(PCA),which takes the logging parameters of adjacent wells(such as natural gamma,well diameter and natural potential)as influencing factors for standard processing,and principal component reduced-dimension analysis,a mathematical model of drillability level was established for hard formation rocks.The results show that the acoustic time difference,formation density,natural gamma and resistivity are most closely related to the drillability level of rocks.It has been validated that the predicted drillability level value obtained using multivariate log-linear regression method is more than 95%in agreement with the testing result.The drill bit selected using this method was applied to the drilling operation in Fengcheng Formation.With it,the drilling footage per bit has been increased from 30 m to 150 m while maintaining the rate of penetration,thus reducing the non-operation time such as bit replacement.
作者
孔祥伟
陈昊
叶佳杰
李亚东
甘一风
KONG Xiangwei;CHEN Hao;YE Jiajie;LI Yadong;GAN Yifeng(School of Petroleum Engineering,Yangtze University,Wuhan 430100,Hubei,China;Research Institute of Engineering Tech-nology,PetroChina Xinjiang Oilfield Company,Karamay 834000,Xinjiang,China)
出处
《新疆石油天然气》
CAS
2022年第3期6-11,共6页
Xinjiang Oil & Gas
基金
国家自然科学基金项目“交变温度、压力及其耦合作用下‘套管-水泥环-地层’完整性研究”(51904261)。
关键词
可钻性级值
主成分分析
硬地层
多元线性回归
钻头优选
drillability level
principal component analysis(PCA)
hard formation
multiple linear regression
bit selection
