Abstract: It is very difficult to gain actual information about fault diagnosis of bearings in reciprocating pump used for oil drilling at the drilling site owing to abominable working conditions and enormous background noise.The demodulated resonance technique combined with spectrum analysis and fuzzy-closest-principle were developed and applied to the fault diagnosis of bearings in drilling pump.The diagnosis result of bearings in reciprocating pump at drilling site showed the probability of this method.

Key words: drilling pump, bearing, fault diagnosis, demodulated resonance technique, spectrum analysis, fuzzy diagnosis

摘要： 在钻井生产现场,由于工作条件恶劣及振源多,背景噪声强烈,因而用理论计算所得到的故障特征频率与实测信号经一般频谱分析后得到的响应信息缺乏一一对应关系.为了解决钻井泵轴承故障特征难以提取的问题,采用共振解调技术来提取深埋于正常振动信息中轴承的故障信息,在此基础上进行频谱分析.但由于实际轴承的几何尺寸有误差,再加上现场各种因素的干扰,使这样得到的频谱图上的频率值与计算所得故障特征频率值仍有出入.进一步采用模糊"择近原则"来识别各种故障特征频率,从而较好地实现了钻井用往复泵轴承的故障诊断.

关键词: 钻井泵, 轴承, 故障诊断, 共振解调技术, 频谱分析, 模糊识别