利用对支持向量机识别石油钻机轴承故障  

Fault Diagnosis of Petroleum Drilling Rig Bearing Based on Twin Support Vector Machine

作　　者：杜明贤[1,2] 张革民[1,2] 申中杰 曹立天[1,2] 王富贵[1,2] 段武[1,2] 

机构地区：[1]宝鸡石油机械有限责任公司研究院,陕西宝鸡721002 [2]国家油气钻井装备工程技术研究中心,陕西宝鸡721002 [3]中煤科工集团西安研究院有限公司,西安710077

出　　处：《机械工程师》2015年第2期131-133,共3页Mechanical Engineer

摘　　要：轴承是石油钻机中易发生故障的旋转支撑零部件。在石油钻机轴承运行状态的监测数据中,发生故障的样本所占的比例很小。如何从监测数据中准确地识别出少量的故障样本成为一个难题。对支持向量机(Twin Support Vectorm achine,TWSVM)能很好地解决这一难题,它通过构造两个非平行的超平面来完成,其中每个超平面都要尽可能地接近自身对应的样本而远离别的样本。文中运用石油钻机轴承故障诊断实验检验TW SV M的效果,结果表明TWSVM的识别精度高,运行速度快,为以后做类似的石油钻机轴承故障检测提供了一种新的分析方法。Bearing is the support components of petroleum drilling rig and prone to be failure. In the monitoring data of the bearing state, only small proportions of signal are available for the fault samples analysis. To diagnose the few f^uh samples from monitoring data accurately becomes a challenge in the unbalance sample set. Twin Support Vector machine （TWSYM） can deah with the prol）lem well by constructing two non-parallel hyperplanes. Each hyperplane is closer to the corresponding samples and is as far as possible from the other samples. The bearing failure diagnosis experiment is used to examine the effects of TWSVM, and the results show that TWSVM has the high accuracy and running speed and a new method is given for doing similar fault diagnosis of petruleum drilling rig bearing.

关 键 词：轴承 故障识别 对支持向量机 超平面 

分 类 号：TV538.3[水利工程—水利水电工程]