基于聚类算法的海洋石油固井设备红外监测大数据预警模型  

Big data early warning model for infrared monitoring of offshore oil cementing equipment based on clustering algorithm

作　　者：张冉[1] 侯林[1] 陈峰 雷彪[1] 邵振友 蔡宝 季威 ZHANG Ran;HOU Lin;CHEN Feng;LEI Biao;SHAO Zhen-you;CAI Bao;JI Wei(China Oilfield Services Ltd.,Hebei 065201,China)

机构地区：[1]中海油田服务股份有限公司,河北廊坊065201

出　　处：《激光杂志》2022年第11期190-193,共4页Laser Journal

基　　金：中海油服固井设备状态监测及预警系统开发(No.G2117A-0521G271)。

摘　　要：传统的海洋石油固井设备预警准确率较低,大数据聚类存在误差。为了解决这一问题,提出基于聚类算法的海洋石油固井设备红外监测大数据预警模型。构建海洋石油固井设备红外监测大数据参数采集模型,对采集的海洋石油固井设备红外监测大数据进行特征分组,建立并行模糊特征组合控制模型,采用模糊C均值聚类的方法,实现对海洋石油固井设备红外监测大数据聚类分析,通过并行特征分布式重组,根据差分进化方法和遍历寻优的方法,实现海洋石油固井设备红外监测大数据的聚类中心寻优,根据聚类特征的差异性水平,实现对海洋石油固井设备红外监测大数据预警。结果表明,本方法大数据聚类时间较短仅为24 s,特征数据收敛误差较小,为0.016 3,预警准确率水平较高,异常数据特征分析准确率为90.5%,具有实际应用价值。The accuracy of early warning of traditional offshore oil cementing equipment is low, and there are errors in big data clustering. To solve this problem, a big data early warning model for infrared monitoring of offshore oil cementing equipment based on clustering algorithm is proposed. A parameter acquisition model for infrared monitoring big data of offshore oil cementing equipment is constructed, and the collected big data are grouped by features, and a parallel fuzzy feature combination control model is established. Fuzzy C-means clustering method is adopted to realize cluster analysis of the big data of infrared monitoring of offshore oil cementing equipment. According to the difference level of cluster characteristics, the infrared monitoring big data early warning of offshore oil cementing equipment is realized. The results show that the big data clustering time of this method is short, only 24 s, the convergence error of characteristic data is small, which is 0.016 3, the accuracy of early warning is high, and the accuracy of abnormal data feature analysis is 90.5%.

关 键 词：聚类算法 海洋石油 固井设备 红外监测 大数据预警 

分 类 号：TN209[电子电信—物理电子学]