本文選題：抽油機(jī)系統(tǒng) + 特征參數(shù)�。� 參考：《自動化技術(shù)與應(yīng)用》2016年10期

【摘要】：水基動力無桿抽油機(jī)作為一項新技術(shù),尚未大規(guī)模的投入使用,對其故障的研究還不成熟。為了提高抽油機(jī)系統(tǒng)故障信號的可分性和診斷正確率,針對抽油機(jī)的典型故障,對采集到的故障信號進(jìn)行數(shù)據(jù)歸一化處理以及EEMD降噪。研究了基于時頻域特征參數(shù)提取的方法以及基于流形學(xué)習(xí)的降維故障診斷方法。最后提出采用距離差異度和流形差異度尋找故障的敏感因子。研究結(jié)果表明,此種方法能夠提取抽油機(jī)故障的敏感因子,為后續(xù)的研究奠定了基礎(chǔ)。
[Abstract]:As a new technology, water-based rod-less pumping unit has not been put into use on a large scale. In order to improve the separability and diagnostic accuracy of fault signals in pumping unit system, the collected fault signals are normalized by data normalization and EEMD noise reduction is carried out in view of typical faults of pumping units. The method of feature parameter extraction in time and frequency domain and the method of dimensionality reduction fault diagnosis based on manifold learning are studied. Finally, the sensitivity factors for fault detection by distance difference and manifold difference are proposed. The results show that this method can extract the sensitive factors of pumping unit fault, and lay a foundation for further research.
【作者單位】： 北京信息科技大學(xué)機(jī)電工程學(xué)院;
【基金】：北京市教委科研計劃重點項目(編號KZ201311232036)
【分類號】：TE933.1
，

本文編號：1921496