基于深度学习的电潜泵井虚拟流量计量系统研究  

Deep Learning-Based Virtual Flow Metering System for Electric Submersible Pump Wells

作　　者：许海丰 武同山 孙国宝 李志辉 陈冰 

机构地区：[1]昆仑数智科技有限责任公司,北京100007 [2]中国石油工程建设有限公司西非公司,北京100085 [3]北京兴油工程项目管理有限公司,北京100083

出　　处：《自动化博览》2024年第4期70-73,共4页Automation Panorama1

摘　　要：本文针对国内、外采油计量系统地面建设投资和维护费用高的问题,建立了以电潜泵井为研究对象的趋势分析、工况诊断预警、产液量计量等业务模型和算法,并研发形成了一套虚拟流量计量系统。该系统具备实时数据采集监控、产量计量等功能,为优化油田地面流程工艺、生产降本增效和提升精细化管理水平提供了技术支撑。Addressing the high investment and maintenance costs of domestic and international oil production metering systems,this study takes electric submersible oil pump wells as the research object,develops business models and algorithms for trend analysis,diagnosis,early warning,and production metering,and develops a set of virtual flow metering system.Equipped with functions of real-time data acquisition and monitoring,production metering,the system provides technical support for the oilfield to optimize the ground process technology,reduce cost and increase efficiency,and improve the level of finemanagement.

关 键 词：虚拟流量计量 深度学习 电潜泵 

分 类 号：TP216[自动化与计算机技术—检测技术与自动化装置]