基于支持向量机的钻柱黏滑振动等级评估方法  

Research on Stick-slip Vibration Level Estimation of Drillstring Based on SVM

作　　者：陈冲 张仕民[1] 彭鹤[1] 崔灿 刘杨[2] Chen Chong;Zhang Shimin;Peng He;Cui Can;Liu Yang(College of Mechanical and Transportation Engineering, China University of Petroleum (Beijing;PetroChina Beijing Oil and Gas Control Center)

机构地区：[1]中国石油大学(北京)机械与储运工程学院 [2]中国石油天然气股份有限公司北京油气调控中心

出　　处：《石油机械》2019年第1期20-26,共7页China Petroleum Machinery

摘　　要：为评估钻柱黏滑振动的严重程度,提出了一种基于因子分析(FA)与支持向量机(SVM)的黏滑振动风险评估方法。对仿真得到的扭矩数据进行时域与频域分析,提取信号的特征值,然后应用因子分析方法减少高维特征的冗余信息,获取特征向量,最后通过SVM对降维处理后的数据进行黏滑振动等级分类。研究结果表明:基于井口扭矩信号的SVM黏滑振动等级预测方法的整体预测精度超过80%,能够较准确地对黏滑振动强度等级进行预测。因此该方法是一个有效的黏滑振动等级分类方法,应用该方法能够有效地对钻柱黏滑振动等级进行识别,有助于司钻根据钻柱黏滑振动剧烈程度实时调整钻井参数,减轻黏滑振动产生的危害,提高钻井作业效率和安全性。To evaluate the severity of stick-slip vibration of drill string, a risk assessment method based on factor analysis (FA) and support vector machine (SVM) is proposed.By performing time domain and frequency domain analysis on the simulated torque data, the signal feature values are extracted.The factor analysis method is used to reduce the redundant information of high-dimensional features to obtain the feature vector.The SVM is used to classify stick-slip vibration levels based on the dimensionality-reduced data.The results show that the overall prediction accuracy of the SVM stick-slip vibration level prediction method based on the wellhead torque signal is more than 8 0 %, which can accurately predict the stick-slip vibration level, presenting an effective method for classifying the stick-slip vibration level.The method can effectively identify the stick-slip vibration level of the drill string, which could help the driller to adjust the drilling parameters in real time according to the severity of the stick-sliding vibration of the drill string, so as to minimize stick-slip vibration hazard and improve drilling efficiency and safety.

关 键 词：黏滑振动 风险评估 时域频域分析 因子分析 特征向量 支持向量机 

分 类 号：TE921[石油与天然气工程—石油机械设备]