聚晶金刚石
钻石
机制(生物学)
过程(计算)
回归
高斯过程
融合
高斯分布
克里金
计算机科学
微晶
材料科学
人工智能
数据挖掘
算法
统计物理学
数学
统计
机器学习
冶金
物理
化学
计算化学
哲学
语言学
量子力学
操作系统
作者
Zhi Yan,Honghai Fan,Xianzhi Song,Hongbao Zhang,Zhaopeng Zhu,Yuhan Liu,Haoyu Diao,Yuguang Ye
出处
期刊:Spe Journal
[Society of Petroleum Engineers]
日期:2024-10-01
卷期号:: 1-18
摘要
Summary Subterranean oil and gas reserves are abundant, offering significant potential for exploration and development. However, oil and gas drilling often suffers from low efficiency due to the dense rock layers encountered. A major cause of this inefficiency is the rapid wear of bits, which significantly reduces their performance. This not only increases the time spent on inefficient drilling but also leads to frequent bit changes, adding to nonproductive time. Therefore, this study focuses on the wear prediction of the widely used polycrystalline diamond compact (PDC) bit in oil and gas drilling operations. In this study, we focused on exploring and validating the smooth wear failure mode of PDC bits in sandstone and mudstone formations. Based on this pattern, we modified the traditional wear mechanism model to suit a data-driven approach and integrated the nonparametric intelligent algorithm, Gaussian process regression (GPR), which performs well with small sample data, for characterizing bit performance. Finally, we applied an adaptive differential evolution (ADE) algorithm to extract the cumulative wear characteristics curve that leads to the degradation of bit performance. This method has been applied to multiple wells in the southwestern China block and the South China Sea block, achieving more than 90% accuracy in model predictions with small sample data. Furthermore, when this method is incorporated into an engineering parameter optimization model, it further unlocks the potential for bit penetration. In practical field applications, it not only enhances the bit footage but also significantly improves overall time efficiency by 11% and 59%, respectively. The application of this method can assist field engineers in identifying inefficient states in oil and gas drilling operations, thereby reducing nonproductive operation time and guiding engineering parameters to enhance drilling efficiency.
科研通智能强力驱动
Strongly Powered by AbleSci AI