往复式泥浆泵故障征兆提取的多参量分析  

Multi-parameter Analysis of Reciprocating Mud Pump for Failure Feature Extraction

作　　者：刘岩[1] 李西兵[1] 王金东[2] 郭建华[1] 

机构地区：[1]齐齐哈尔大学,黑龙江齐齐哈尔161006 [2]东北石油大学,黑龙江大庆163318

出　　处：《机床与液压》2013年第23期154-156,共3页Machine Tool & Hydraulics

基　　金：国家自然科学基金项目(51075218;51175273)

摘　　要：在采用G_P算法识别往复机械故障征兆过程中,嵌入维数、降噪方法、延迟时间等参数对算法的影响至关重要。通过仿真分析单变量算法的弊端,提出基于经验模态分解(EMD)与伪相图技术结合的多变量融合算法,采用归一化的小波降噪技术,将此法应用在矿用泥浆泵故障征兆提取中,诊断效果较好。In the course of reciprocating mechanical fault feature is extracted by using G P algorithm, such parameters like em- bedding dimension, de-noise method, delay-time have crucial influence on algorithm. Through simulation analyzing flaws of univariate time series algorithm, a new fusion algorithm of combining empirical mode decomposition with pseudo-phase portrait was proposed. Using normalize wavelet de-noising method, the method was applied to identify fault feature of reciprocating mud pump for mining and better results were obtained.

关 键 词：泥浆泵 故障诊断 多变量融合算法 关联维数 经验模态分解 伪相图 

分 类 号：TH457[机械工程—机械制造及自动化] TP277[自动化与计算机技术—检测技术与自动化装置]