基于GWO-FCM的输油泵故障诊断模型自学习框架  

A self⁃learning framework for a fault diagnosis model of oil pump based on grey wolf optimization−fuzzy C⁃means clustering (GWO−FCM)

作　　者：郭俊霞[1] 谢自力 毛申申 魏聪聪 邢健[3] GUO JunXia;XIE ZiLi;MAO ShenShen;WEI CongCong;XING Jian(College of Information Science and Technology,Beijing University of Chemical Technology,Beijing 100029;Eastern Crude Oil Storage&Transportation Company Limited Pipe China,Xuzhou 221008;College of Mechanical and Electrical Engineering,Beijing University of Chemical Technology,Beijing 100029,China)

机构地区：[1]北京化工大学信息科学与技术学院,北京100029 [2]国家管网集团东部原油储运有限公司,徐州221008 [3]北京化工大学机电工程学院,北京100029

出　　处：《北京化工大学学报（自然科学版）》2024年第6期79-86,共8页Journal of Beijing University of Chemical Technology(Natural Science Edition)

摘　　要：随着输油泵场站无人化建设的发展,企业对输油泵故障诊断技术的要求也越来越高。目前,被广泛使用的利用机器学习算法进行输油泵故障诊断的方法都只能针对模型训练集中已包含的几类故障进行诊断,在企业的实际使用中,仍会出现其他不包含在训练集中的故障而不能被正确自动识别、诊断。针对上述问题,设计了一种输油泵故障诊断模型自学习框架,通过信号处理技术结合深度学习提取深层故障特征,提高工业现场数据的可分性;通过模糊C均值聚类结合相似度度量判别已知故障和未知故障,对出现的未知故障模式进行识别和记录;利用频繁出现的未知故障数据重训练模型,在原有诊断功能的基础上提高对未知故障的识别、诊断及学习能力。为验证方法的有效性,使用工业现场采集的输油泵数据进行实验,结果表明,现有诊断方法所提出的输油泵故障诊断模型自学习框架能够实现对未知故障的准确识别。Oil pumps are critical pieces of equipment in the petroleum industry,and their reliability and stability are crucial for oil transportation.With the construction of intelligent and unmanned stations,the demand for oil pump fault diagnosis technology is increasing.Currently,using machine learning for oil pump fault diagnosis has achieved some positive results.However,existing fault diagnosis methods can only diagnose the types of faults included in the model training set.In actual usage,other faults not included in the training set may occur and cannot be correctly identified and diagnosed automatically.To address this issue,this paper proposes a self-learning framework for the fault diagnosis model of oil pumps.By combining signal processing techniques with deep learning methods to extract deep fault features,the separability of industrial field data is improved.By using unsupervised clustering and similarity measurement methods to differentiate known and unknown faults,the framework can identify and record unknown fault patterns that occur and retrain the model using frequently occurring unknown fault data.By adding the self-learning mechanism to the existing fault diagnosis model,the framework can increase its ability to recognize,diagnose,and learn from unknown faults while maintaining the original diagnostic function.To validate the effectiveness of the proposed method,experimental oil pump data collected from an industrial oilfield was used.The results show that the self-learning framework for the fault diagnosis model of oil pumps proposed in this paper can accurately identify unknown faults.

关 键 词：输油泵 故障诊断 自学习 模糊C均值聚类 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]