基于深度学习的风电机组轴承诊断系统  

Deep Learning-based Bearing Diagnosis System for Wind Turbines

作　　者：骆东松[1] 马立东 LUO Dongsong;MA Lidong(College of Electrical and Information Engineering,Lanzhou University of Technology,Lanzhou 730050,China)

机构地区：[1]兰州理工大学电气工程与信息工程学院,甘肃兰州730050

出　　处：《机械与电子》2024年第7期69-75,共7页Machinery & Electronics

摘　　要：风电机组的轴承作为其关键部件,正常工作时常出现因磨损和裂纹而失效的状况,导致所采集到的振动数据含有其他干扰信号,而传统的诊断方法故障检测误差较大,导致诊断结果准确性较差。针对这一问题,提出了一个基于DRSN-CW模型的风电机组轴承诊断系统。该系统结合了软、硬件设计,算法理论和仿真结果分析,旨在提高轴承故障检测的准确性和效率。通过合适的传感器和数据采集设备进行硬件设计,构建了包括数据预处理、特征提取、模型训练和推理等模块的软件设计,实现了自动化的轴承故障诊断过程。选择DRSN-CW作为基础模型,结合软阈值化、残差网络和注意力机制,有效学习轴承信号中的重要特征。使用凯斯西储大学轴承振动数据进行仿真实验,并对多种不同的故障诊断算法进行了分析。实验结果表明,DRSN-CW模型的准确率优于其他方法。The bearings of wind turbines,as their key components,often fail due to wear and cracks under normal operation,resulting in the collected vibration data containing other interfering signals,and the traditional diagnostic methods have large error in fault detection,leading to poor accuracy of the diagnostic results.To address this problem,a wind turbine bearing diagnosis system based on the DRSN-CW model is proposed.The system combines software and hardware design,algorithm theory and simulation results analysis,aiming to improve the accuracy and efficiency of bearing fault detection.The hardware design is carried out by suitable sensors and data acquisition devices,and the software design including modules of data preprocessing,feature extraction,model training and inference is constructed to realize the automated bearing fault diagnosis process.DRSN-CW is selected as the base model,which combines soft valorization,residual network and attention mechanism to effectively learn important features in bearing signals.Simulation experiments are conducted using Case Western Reserve University bearing vibration data,and several different fault diagnosis algorithms are analyzed.Experimental results show that the accuracy of the DRSN-CW model outperforms other methods.

关 键 词：风电机组 轴承 故障诊断 收缩网络 残差网络 

分 类 号：TM315[电气工程—电机] TP181[自动化与计算机技术—控制理论与控制工程]