抽油机故障音频智能诊断技术应用研究

Application of audio intelligent diagnosis technology in pumping unit faults

针对人工巡检的局限性,提出了一种基于大数据分析的抽油机故障音频智能诊断方法。首先利用抽油机音频智能采集器,采集抽油机的音频数字信号,而后通过音频特征值提取、数据降维与可视化,建立抽油机故障特征音频库;最后将监测抽油机音频进行数据可视化和自动分类分析,与特征音频库对比分析,对抽油机进行故障分类和故障报警。抽油机故障音频智能诊断技术在江苏油田现场应用112井次,发现故障58井次,经现场核实53井次诊断正确,故障诊断准确率达91.4%。应用表明:基于大数据分析的抽油机故障音频智能诊断方法能够准确、有效地识别抽油机故障类型,具有良好的应用前景。

Aiming at the limitations of manual inspection,an intelligent audio diagnosis method for pumping unit faults based on big data analysis is proposed.Firstly,the audio digital signals of the pumping unit are collected by using the audio intelligent collectors.And then the audio feature database of pumping unit faults is established through audio feature extraction,data dimensionality reduction,and visualization.Finally,after the data visualization and automatic classification analysis,the comparative analysis of monitored pumping units’audios and characteristic audios library,and the pumping unit’s fault classification and fault alarm are carried out.The audio intelligent faults diagnosis technology for pumping unit faults has been applied to 112 wells in Jiangsu Oilfield,and 58 wells were found to be faulty.Among them,the diagnosis of 53 wells is accurate after on-site verification,and the accuracy rate of fault diagnosis is 91.4%.The application shows that the audio intelligent diagnosis method of pumping unit faults based on big data analysis can accurately and effectively identify the fault types of pumping units and has a good application prospect.