基于特征标记的油田无人值守站运行异常辨识研究  

Research on abnormal operating identification of oilfield unattended stations based on characteristic markers

作　　者：赵敏[1] 王云辉[1] 王华[1] 付兵 金峰[1] ZHAO Min;WANG Yunhui;WANG Hua;FU Bing;JIN Feng(No.5 Oil Production Plant of Changqing Oilfield)

机构地区：[1]长庆油田第五采油厂

出　　处：《石油石化节能与计量》2024年第4期25-29,共5页Energy Conservation and Measurement in Petroleum & Petrochemical Industry

摘　　要：油田无人值守站的运行异常受多种因素综合影响,产生的数据可能存在噪声或错误,导致油田无人值守站运行异常辨识难度上升,所以研究一种新的基于特征标记的油田无人值守站运行异常辨识方法。按照既定时长采集无人值守站运行数据,对采集到的数据进行分析和降噪处理。以数据处理结果为基础,结合均值、方差和均方根值提取异常数据特征,并对异常状态特征进行标记处理,结合油田无人值守站运行异常辨识模型得到相关的辨识结果。实验结果表明,该设计的基于特征标记的油田无人值守站运行异常辨识方法的辨识准确率为93.77%,辨识效果好,可以在相关领域实现广泛应用。The abnormal operation of oilfield unattended stations is influenced by multiple factors,and the generated data may contain noise or errors,which increases the difficulty of identifying abnormal operation of oilfield unattended stations.Therefore,a new method based on characteristic markers for identifying abnormal operation of oilfield unattended stations is studied.The operating data of unmanned station is collected according to the specified time,and it is analyzed and reduced noise in time.Based on the data processing results,abnormal data features are extracted by combining mean,variance,and root mean square values,and abnormal state features are labeled and processed.Relevant identification results are obtained by combining with the abnormal identification model of oilfield unattended stations.The experimental results show that the identification accuracy of characteristic markers based anomaly identification method for oilfield unattended stations designed is 93.77%,and the identification effect is good,which can be widely applied in related fields.

关 键 词：特征标记 无人值守站 运行异常辨识 数据处理 异常辨识模型 

分 类 号：TE48[石油与天然气工程—油气田开发工程]