
浏览全部资源
扫码关注微信
1.中国地质大学(北京)海洋学院
2.中国地质科学院勘探技术研究所
3.中国地质学会自动化智能化 钻探装备创新基地
4.广州海洋地质调查局
Received:28 August 2025,
Revised:2026-05-04,
Accepted:23 November 2025,
Published:25 October 2026
移动端阅览
秦如雷,谢文卫,高洁云,等.机理与数据驱动的深海钻井液举升泵布放位置评估[J].石油机械,2026,54(10):47-54.
Qin Rulei,Xie Wenwei,Gao Jieyun,et al.Mechanism and data-driven evaluation of deep-sea drilling fluid lift pump deployment[J].China Petroleum Machinery,2026,54(10):47-54.
秦如雷,谢文卫,高洁云,等.机理与数据驱动的深海钻井液举升泵布放位置评估[J].石油机械,2026,54(10):47-54. DOI: 10.12473/CPM.202508091.
Qin Rulei,Xie Wenwei,Gao Jieyun,et al.Mechanism and data-driven evaluation of deep-sea drilling fluid lift pump deployment[J].China Petroleum Machinery,2026,54(10):47-54. DOI: 10.12473/CPM.202508091.
针对深海无隔水管钻井液循环系统中举升泵安放深度不当易诱发汽蚀的工程难题,传统物理模型优化方法在解决此问题时计算成本高昂、难以开展大范围工况寻优。为此,建立了钻井液循环流体动力学物理模型,通过参数化扫描生成仿真数据,抽取代表性离散样本构建训练集与验证集;在此基础上训练极致梯度提升(XGBoost)、随机森林(RF)和梯度提升(GB)3种树集成回归模型以预测泵入口压力,并采用综合评价方法优选代理模型;进一步提出工况适应性得分指标,在4 000组预设工况下对300~1 300 m多种泵挂深度方案进行比较评估。研究结果表明:XGBoost代理模型预测精度超过99%,特征重要性排序与流体力学机理一致;在设定工程边界条件下,1 100 m是举升泵的最优安放深度,该方案在确保100%安全性的前提下,相比更深的方案具有更好的工程经济性,且系统具有较大的安全操作窗口,对工况波动有较强的鲁棒性。所得结果可为深海钻井液举升系统及相关装备设计与优化提供决策支持。
Improper deployment depth of the lift pump in deep-sea riserless mud recovery (RMR) systems may induce pump cavitation. To address this engineering challenge, traditional physics-based optimization methods are computational costly and cannot realize large-scale working condition optimization. In this regard, a fluid-dynamics physical model for mud recovery was established, simulation data were generated through parametric sweeping, and representative discrete samples were extracted to construct the training set and validation set. Three tree-ensemble regression models, i.e. eXtreme gradient boosting (XGBoost), random forest (RF), and gradient boosting (GB), were trained to predict pump inlet pressure, and a comprehensive evaluation method was used to select the optimal surrogate model. An index of working-condition adaptability score was further proposed to evaluate the schemes of pump deployment depths from 300 m to 1 300 m under 4 000 predefined operating conditions. The results show that the XGBoost surrogate model achieves a prediction accuracy above 99% and the feature-importance ranking consistent with the fluid mechanics mechanisms. Under the specified engineering boundary conditions, 1 100 m is identified as the optimal deployment depth for the lift pump. This scheme ensures 100% safety while offering better engineering economy compared with deeper schemes, and the system exhibits a large safe operating window with strong robustness against working-condition fluctuations. The findings provide decision-making support for the design and optimization of deep-sea drilling fluid lift systems and related equipment.
Mao Gezhen , Jie Zhang , Xie Jiaohao , et al . Transient temperature field simulation and sensitivity factors analysis of riserless mud recovery system [J]. Petroleum Science and Technology , 2025 , 43 ( 8 ): 807 - 830 .
Wang Jintang , Sun Jinsheng , Xie Wenwei , et al . Simulation and analysis of multiphase flow in a novel deepwater closed-cycle riserless drilling method with a subsea pump+gas combined lift [J]. Frontiers in Physics , 2022 , 10 : 946516 .
Hammitt F G . Cavitation and multiphase flow phenomena [M]. [S.l.:s.n.] , 1980 .
Duan Aoqiang , Lin Zhe , Chen Desheng , et al . A review on the hydraulic performance and erosion wear characteristic of the centrifugal slurry pump [J]. Particuology , 2024 , 95 : 370 - 392 .
凌勇坚 . 消除离心式水泵气蚀的措施 [J]. 中国设备工程 , 2005 ( 7 ): 25 - 25 .
Ling Yongjian . Measures to eliminate cavitation of centrifugal water pumps [J]. China Plant Engineering , 2005 ,( 7 ): 25 - 25 .
韩雷 , 邬霁 , 刘永茜 , 等 . 离心泵安装高度的确定及气蚀的控制措施 [J]. 天然气与石油 , 2010 , 28 ( 1 ): 4 - 5,32 .
Han Lei , Wu Ji , Liu Yongqian , et al . Determination of centrifugal pump installation height and measures for gas erosion control [J]. Natural Gas and Oil , 2010 , 28 ( 1 ): 4 - 5, 32 .
杨建一 , 袁中凯 . 离心泵安装高度计算中出现的问题分析 [J]. 流体机械 , 2004 , 32 ( 4 ): 20 - 22 .
Yang Jianyi , Yuan Zhongkai . Analysis on problem of installing height calculating for centrifugal pump [J]. Fluid Machinery , 2004 , 32 ( 4 ): 20 - 22 .
李昊 , 孙宝江 , 高永海 , 等 . 深水控制泥浆帽钻井水力参数设计与计算 [J]. 石油钻探技术 , 2013 , 41 ( 3 ): 13 - 18 .
Li Hao , Sun Baojiang , Gao Yonghai , et al . Design and calculation of hydraulic parameters for controlling mud cap in deepwater drilling [J]. Petroleum Drilling Techniques , 2013 , 41 ( 3 ): 13 - 18 .
Tu Yongsha , Zhao Xueying , Lu Lifeng , et al . Flow characteristics and pressure pulsation analysis of cavitation induced in a double-volute centrifugal pump [J]. Water , 2025 , 17 ( 3 ): 445 .
张斌 , 程国赞 , 洪昊岑 , 等 . 基于计算流体仿真的双作用叶片泵气蚀机理分析 [J]. 吉林大学学报(工学版) , 2021 , 51 ( 3 ): 831 - 839 .
Zhang Bin , Cheng Guozan , Hong Haocen , et al . Cavitation mechanism of double acting vane pump based on computational fluid dynamics simulation method [J]. Journal of Jilin University (Engineering and Technology Edition) , 2021 , 51 ( 3 ): 831 - 839 .
马希青 , 陈明辉 , 孙旋 . 叶片泵叶轮流场仿真与参数优化分析 [J]. 机械设计与制造 , 2017 ( 4 ): 55 - 58 .
Ma Xiqing , Chen Minghui , Sun Xuan . Analysis of a vane pump impeller of flow field simulation and parametric optimization [J]. Machinery Design & Manufacture , 2017 ( 4 ): 55 - 58 .
Lu Jiaxing , Li Jiarui , Zhang Chuan , et al . Research on the cavitation characteristics of pump turbines based on mode decomposition [J]. Processes , 2025 , 13 ( 3 ): 1 - 11 .
Silva P A S F , Shinomiya L D , de Oliveira T F , et al . Analysis of cavitation for the optimized design of hydrokinetic turbines using BEM [J]. Applied Energy , 2017 , 185 : 1281 - 1291 .
Wei Changxu , He Chunlei , Tan Helong , et al . A predictive model for tool wear behavior during ultra-precision lapping [J]. The International Journal of Advanced Manufacturing Technology , 2025 , 136 ( 2 ): 875 - 895 .
张琴 , 冯定 , 张慢来 , 等 . 空穴影响离心泵工作特性的CFD分析 [J]. 石油机械 , 2015 , 43 ( 6 ): 61 - 65 .
Zhang Qin , Feng Ding , Zhang Manlai , et al . CFD analysis on the effect of cavity on the performance of centrifugal pump [J]. China Petroleum Machinery , 2015 , 43 ( 6 ): 61 - 65 .
张伟伟 , 王旭 , 寇家庆 . 面向流体力学的多范式融合研究展望 [J]. 力学进展 , 2023 , 53 ( 2 ): 433 - 467 .
Zhang Weiwei , Wang Xu , Kou Jiaqing . Prospects of multi-paradigm fusion methods for fluid mechanics research [J]. Advances in Mechanics , 2023 , 53 ( 2 ): 433 - 467 .
宋洪庆 , 都书一 , 王九龙 , 等 . 数智流体力学的发展及油气渗流领域应用 [J]. 力学学报 , 2023 , 55 ( 3 ): 765 - 791 .
Song Hongqing , Du Shuyi , Wang Jiulong , et al . Development of digital intelligence fluid dynamics and applications in the oil & gas seepage fields [J]. Chinese Journal of Theoretical and Applied Mechanics , 2023 , 55 ( 3 ): 765 - 791 .
Ullrich K , von Elling M , Gutzeit K , et al . AI-based optimisation of total machining performance: a review [J]. CIRP Journal of Manufacturing Science and Technology , 2024 , 50 : 40 - 54 .
曾红 , 王全玉 , 张志华 . 离心泵遗传算法多目标优化与压力脉动分析 [J]. 制造业自动化 , 2021 , 43 ( 1 ): 118 - 122 .
Zeng Hong , Wang Quanyu , Zhang Zhihua . Multi-objective optimization and pressure pulsation analysis of genetic algorithm for centrifugal pump [J]. Manufacturing Automation , 2021 , 43 ( 1 ): 118 - 122 .
周逸龙 . 基于改进核极限学习机的离心泵空化状态识别研究 [D]. 镇江 : 江苏科技大学 , 2022 .
Zhou Yilong . Research on cavitation state identification of centrifugal pumps based on improved kernel limit learning machine [D]. Zhenjiang : Jiangsu University of Science and Technology , 2022 .
Wang Yuqin , Shao Jiale , Yang Fan , et al . Optimization design of centrifugal pump cavitation performance based on the improved BP neural network algorithm [J]. Measurement , 2025 , 245 : 116553 .
0
Views
6
下载量
0
CSCD
Publicity Resources
Related Articles
Related Author
Related Institution
京公网安备11010802024621