基于SOM-BP级联神经网络的电驱离心泵健康状态识别方法  

Health status recognition method of electrically driven centrifugal pump based on cascade-structured SOM-BP neural networks

作　　者：江新星 吴杰 薛一冰 彭世亮 JIANG Xinxing;WU Jie;XUE Yibing;PENG Shiliang(Compressor Maintenance and Repair Company,PipeChina Pipeline Technology Development Co.,Ltd.;Hunan Branch of PipeChina;College of Mechanical and Transportation Engineering,China University of Petroleum(Beijing))

机构地区：[1]国家管网集团储运技术发展有限公司压缩机组维检修分公司 [2]国家管网集团湖南分公司 [3]中国石油大学(北京)机械与储运工程学院

出　　处：《油气储运》2025年第3期350-359,共10页Oil & Gas Storage and Transportation

基　　金：国家自然科学基金青年项目“智能化设备健康状态评估与故障诊断方法研究”,52401412。

摘　　要：【目的】离心泵预测性维护是提升设备可靠性与运行效率的核心技术之一,在该过程中,对离心泵设备的健康状态识别是关键环节。然而,传统的健康状态识别方法多依赖于机器学习技术,高度依赖足量标记数据,难以直观清晰地表征监测数据与健康状态之间的对应关系,使其在实际复杂工况中的应用效果受限,亟需开发更精准、高效且适应性更强的健康状态识别方法。【方法】提出一种基于自组织映射(Self-Organization Map, SOM)神经网络与BP(Back Propagation)神经网络级联的电驱离心泵健康状态识别方法。首先采用SOM神经网络方法对离心泵全生命周期振动数据进行预处理,提取时域、频域及时频域的多种统计特征与熵特征,从而全面表征设备的运行状态;其次,采用主成分分析法(Principal Component Analysis, PCA)对已提取的轴承振动信号特征进行降维与融合,有效减少冗余信息和计算复杂度,优化输入参数的模式,提升建模效率;最后,综合SOM神经网络与BP神经网络的优点,建立了基于SOM-BP级联神经网络的电驱离心泵健康状态识别模型。【结果】以某电驱离心泵的健康状态监测数据集为算例,对比了SOM-BP模型与常见的机器学习方法(随机森林模型、XG-boost模型)识别电驱离心泵健康状态的准确率,以R^(2)、MSE、RMSE为模型评价指标,结果表明:基于SOM-BP级联神经网络模型的R^(2)值、MSE值、RMSE值分别为0.901、0.8×10^(-6)m^(2)/s^(4)、9.12×10^(-4)m/s^(2),显著优于传统的机器学习方法,展现出良好的鲁棒性与适应性。【结论】基于SOM-BP级联神经网络计算方法不仅提升了离心泵健康状态识别的精度,还可为离心泵故障诊断与剩余寿命预测提供数据支撑,同时为其他旋转机械的健康状态管理与诊断提供了新思路。[Objective]Predictive maintenance is recognized as one of the core approaches to improving the reliability and operational efficiency of centrifugal pumps.A crucial component of implementing this strategy lies in the health status recognition of these pumps.However,traditional methods for health status recognition primarily depend on machine learning techniques that require a substantial amount of labeled data.This dependency often leads to their limitations in effectively and clearly representing relationships between monitoring data and health status,thus constraining their application effect in real-world conditions,which are typically complex.Therefore,it is urgently needed to develop health status recognition methods that are more accurate,efficient,and adaptable.[Methods]This paper presents a health status recognition method for electrically driven centrifugal pumps,utilizing a cascade configuration of the Self-Organizing Map(SOM)neural network and the Back Propagation(BP)neural network.First,the SOM neural network was employed to preprocess the life-cycle vibration data of centrifugal pumps,focusing on extracting various statistical characteristics and entropy features in the time,frequency,and time-frequency domains, to fully characterize their operating states. Next, the extracted bearing vibration signal features were subjected todimensionality reduction and integration using Principal Component Analysis (PCA). This facilitates the reduction of redundantinformation and computational complexity, the optimization of the pattern of input parameters, and the enhancement of modeling efficiency.Finally, a health status recognition model was established for electrically driven centrifugal pumps based on cascaded SOM-BPneural networks, leveraging the advantages of both neural networks. [Results] Based on the health status monitoring dataset of anelectrically driven centrifugal pump, a comparison was conducted between the SOM-BP model and common machine learning techniques,such as random forest and XGBoost model

关 键 词：自组织映射 BP神经网络 电驱离心泵 健康状态识别 

分 类 号：TE832[石油与天然气工程—油气储运工程]