摘要
为提高石油钻采装备外部故障检测能力,提出基于随机森林算法的监测方法。构建大数据采集模型,以石油钻采装备的异常振动数据为研究对象,进行故障特征提取和信息融合,构建故障工况下的信息融合和特征聚类模型,通过模糊C均值聚类进行故障特征的量化分解和分类识别,在随机森林学习算法下实现对故障检测和诊断的自适应寻优。仿真结果表明,采用该方法进行故障检测,可有效提高故障的自动监测能力,且准确性较高,实时性较好。
In order to improve the ability of oil drilling equipment external fault detection,a monitoring method based on random forest algorithm is proposed.Based on the abnormal vibration data of oil drilling and production equipment,the big data acquisition model is constructed,and the fault feature extraction and information fusion are carried out.The information fusion and feature clustering model under fault conditions are constructed.The fault features are quantitatively decomposed and classified by fuzzy c-means clustering,and the adaptive optimization of fault detection and diagnosis is realized under the random forest learning algorithm.The simulation results show that the method can effectively improve the ability of automatic fault monitoring,with high accuracy and good real-time performance.
作者
雷彪
陈江
侯林
LEI Biao;CHEN Jiang;HOU Lin(COSL Oilfield Chemicals Division Mexico Operation Company,Ciudad Del Carmen 24129 Mexico;COSL Oilfield Chemicals Division,Yanjiao 065200 China)
出处
《自动化技术与应用》
2021年第7期125-128,155,共5页
Techniques of Automation and Applications
关键词
随机森林算法
石油钻采装备
外部故障
自动监测
random forest algorithm
oil drilling and production equipment
external fault
automatic monitoring
作者简介
雷彪(1984-),男,本科,中级设备工程师,研究方向:海洋石油装备研究及管理。
