电机轴承作为旋转机械中的关键部件，其工作状态影响整个机械设备的运转乃至人员安全，因此对电机轴承进行故障诊断具有非常重要的现实意义。随着智能制造技术的不断完善和发展，故障诊断将会为高效、安全和稳定的工业生产提供可靠保障。本文主要从电机轴承振动信号的处理、故障特征的提取与降维和故障诊断模型的建立与改进三个方面入手，对电机轴承故障诊断进行深入研究，以达到提高故障诊断准确率及减少诊断时间的目标。

       首先，本文在了解电机轴承基本结构和失效形式的基础上，对其故障特征频率模型进行分析。针对小波阈值函数在去噪方面的不足，提出一种改进的阈值函数以消除噪声影响。同时引入变分模态分解方法(Variational Mode Decomposition, VMD)对振动信号进行分解，根据提出的综合评价指标筛选出有效分量进行重构，对其进行包络谱分析，得到每一种故障的特征频率。其次，对重构后的信号进行多域指标求解并构建故障特征集。将利用类内、类间距离筛选出的特征指标与使用平均影响值(Mean Impact Value, MIV)筛选出来的特征指标进行共性特征提取，得到一个维数更低的故障特征集。最后，针对传统优化算法收敛速度慢和识别率低的问题，引入Tent混沌映射、指南针算子和社会因子，并借鉴萤火虫算法的吸引跟随策略对麻雀搜索算法(Sparrow Search Algorithm, SSA)进行改进。在此基础上利用改进后的麻雀搜索算法(Improved Sparrow Search Algorithm, ISSA)对支持向量机(Support Vector Machine, SVM)的惩罚参数C和核参数g进行寻优，建立ISSA-SVM故障诊断模型。综上分析，本文建立一种基于改进小波阈值去噪(Improved Wavelet Threshold Denoising, IWTD)、自适应变分模态分解和改进麻雀搜索算法优化支持向量机相结合的电机轴承故障诊断模型。

       在上述工作的基础上，利用在实验平台上所采集的轴承振动信号对本文所提故障诊断模型进行验证。通过对比本文构建的五种故障特征集，验证了本文所提降维方法的有效性。将SVM、SSA-SVM、WOA-SVM和ISSA-SVM模型进行对比，结果表明ISSA具有较强的优化能力。通过BPNN、RF、ELM和ISSA-SVM模型的对比分析，结果说明本文模型在诊断准确率和诊断时间等方面具有一定的优越性。继而，借助MATLAB 软件的图形用户界面(Graphical User Interfaces, GUI)工具开发出一套电机轴承故障诊断系统。通过本文所建立的电机轴承故障诊断模型，能够有效提高诊断准确率，减少诊断时间，进而确保系统的正常运行和稳定。

  As a key part of rotating machinery, the working state of motor bearing affects the operation of the whole mechanical equipment and even the safety of personnel, so the fault diagnosis of the motor bearing has very important practical significance. With the continuous improvement and development of intelligent manufacturing technology, fault diagnosis will provide a reliable guarantee for efficient, safe and stable industrial production. In this paper, the fault diagnosis of motor bearing is lucubrated from three aspects: the processing of motor bearing vibration signals, the extraction and dimension reduction of fault feature, the establishment and improvement of fault diagnosis model, so as to improve the accuracy of fault diagnosis and reduce the diagnosis time.

  Firstly, the bearing fault characteristic frequency model is analyzed based on the understanding of motor bearing basic structure and failure form. Aiming at the shortcomings of wavelet threshold function in denoising, an improved threshold function is proposed to eliminate the noise effect. At the same time, the variational mode decomposition (VMD) method is introduced to decompose the vibration signal. According to the proposed comprehensive evaluation index, the effective components are selected for reconstruction, and the envelope spectrum analysis is carried out to obtain the characteristic frequency of each fault. Secondly, the reconstructed signal is solved by multi-domain index and the fault feature set is constructed. The feature indexes selected by the intra-class and inter-class distance and the feature indexes selected by the average influence value (MIV) are used to extract the common features, and a fault feature set with lower dimension is obtained. Finally, aiming at the problems of slow convergence speed and low recognition rate of traditional optimization algorithms, sparrow search algorithm (SSA) is improved by introducing Tent chaotic mapping, compass operator and social factor, and referring to the attraction following strategy of firefly algorithm. On this basis, the improved sparrow search algorithm (ISSA) is used to optimize the penalty factor and kernel parameters of support vector machine (SVM), and the ISSA-SVM fault diagnosis model is established. In summary analysis, a motor bearing fault diagnosis model based on improved wavelet threshold denoising (IWTD), adaptive VMD and ISSA-SVM is established.

  On the basis of the above work, the bearing vibration signals collected on the experimental platform are used to verify the fault diagnosis model proposed in this paper. By comparing the five fault feature sets constructed in this paper, the effectiveness of the proposed dimension reduction method is verified. The comparison of SVM, SSA-SVM, WOA-SVM and ISSA-SVM models, results show that ISSA has strong optimization ability. Through the comparison and analysis of BPNN, RF, ELM and ISSA-SVM models, results show that the proposed model has certain advantages in diagnosis accuracy and diagnosis time. Then, with the help of graphical user interfaces (GUI) tool of MATLAB software, a set of motor bearing fault diagnosis system is developed. The motor bearing fault diagnosis model constructed in this paper can effectively improve the diagnosis accuracy, reduce the diagnosis time, and ensure the normal operation and stability of the system.

[1] Zhou J M, Yang X T, Liu L L, et al. Fuzzy broad learning system combined with feature-engineering-based fault diagnosis for bearings[J]. Machines, 2022, 10(12): 1229.

[2] He Z Y, Shao H D, Ding Z Y, et al. Modified deep autoencoder driven by multisource parameters for fault transfer prognosis of aeroengine[J]. IEEE Transactions on Industrial Electronics, 2021, 69(1): 845-855.

[3] Xu S J, Yuan R, Lv Y, et al. A novel fault diagnosis approach of rolling bearing using intrinsic feature extraction and CBAM-enhanced InceptionNet[J]. Measurement Science and Technology, 2023, 34(10): 5111.

[4] Xu Y W, Cai W W, Xie T C. Fault diagnosis of subway traction motor bearing based on information fusion under variable working conditions[J]. Shock and Vibration, 2021: 1-12.

[5] Scott S, Ali Z. Fabrication methods for microfluidic devices: An overview[J]. Micromachines-Basel, 2021, 12: 319.

[6] Li N P, Lei Y G, Lin J, et al. An improved exponential model for predicting remaining useful life of rolling element bearings[J]. IEEE Transactions on Industrial Electronics, 2015, 62(12): 7762-7773.

[7] Sun W J, Zhao R, Yan R Q, et al. Convolutional discriminative feature learning for induction motor fault diagnosis[J]. IEEE Transactions on Industrial Informatics, 2017, 13(3): 1350-1359.

[8] Jiang J H, Liu Y Q, Xu C Z, et al. Research on motor bearing fault diagnosis based on the adaboost algorithm and the ensemble learning with bayesian optimization in the industrial internet of things[J]. Security and Communication Networks, 2022.

[9] Lee S, Jeong J. SSA-SL transformer for bearing fault diagnosis under noisy factory environments[J]. Electronics, 2022, 11(9): 1504.

[10]邱翔. 基于变分模态分解的电机轴承故障检测方法研究[D]. 大连: 大连海事大学, 2022.

[11]余忠潇. 基于LCD和模糊聚类的滚动轴承故障诊断方法研究[D]. 石家庄: 石家庄铁道大学, 2020.

[12]Mian T, Choudhary A, Fatima S. Vibration and infrared thermography based multiple fault diagnosis of bearing using deep learning[J]. Nondestructive Testing and Evaluation, 2023, 38(2): 275-296.

[13]宋向金, 赵文祥. 交流电机信号特征分析的滚动轴承故障诊断方法综述[J]. 中国电机工程学报, 2022, 42(4): 1582-1596.

[14]Gangsar P, Tiwari R. Signal based condition monitoring techniques for fault detection and diagnosis of induction motors: A state-of-the-art review[J]. Mechanical systems and signal processing, 2020, 144: 106908.

[15]Fan H W, Ma J T, Cao X G, et al. An intelligent diagnosis approach combining resampling and CWGAN-GP of single-to-mixed faults of rolling bearings under unbalanced small samples[J]. International Journal of Pattern Recognition and Artificial Intelligence, 2023, 37(12): 2356017.

[16]Yan G X, Chen J, Bai Y, et al. A survey on fault diagnosis approaches for rolling bearings of railway vehicles[J]. Processes, 2022, 10(4): 724.

[17]Lv D, Wang H, Che C. Multiscale convolutional neural network and decision fusion for rolling bearing fault diagnosis[J]. Industrial Lubrication and Tribology, 2021, 73(3): 516-522.

[18]王进花, 马雪花, 岳亮辉, 等. 少量样本下基于PCA-BNs的多故障诊断[J]. 振动与冲击, 2024, 43(04): 288-296.

[19]王照伟, 郭雯君, 宋向金, 等. 融合TKEO和Goerztel算法的感应电机转子断条故障诊断[J/OL]. 电工技术学报, 2024: 1-12.

[20]何赟泽, 刘松源, 白芸, 等. 基于声发射的Cascode型GaN HEMT器件机械应力波检测与分析[J]. 中国电机工程学报, 2023, 43(2): 750-761.

[21]Kang M, Kim J, Wills L M, et al. Time-varying and multiresolution envelope analysis and discriminative feature analysis for bearing fault diagnosis[J]. IEEE Transactions on Industrial Electronics, 2015, 62(12): 7749-7761.

[22]Liu Z P, Wang X F, Zhang L. Fault diagnosis of industrial wind turbine blade bearing using acoustic emission analysis[J]. IEEE Transactions on Instrumentation and Measurement, 2020, 69(9): 6630-6639.

[23]周刚, 张舜磊. 关于地铁车辆牵引电机异响的分析[J]. 铁道车辆, 2023, 61(6): 133-138.

[24]郭军. 基于GAN与CNN的煤机设备故障智能诊断方法研究[J]. 煤矿机械, 2023, 44(12): 182-185.

[25]徐敏敏. 滚动轴承波纹度表征与典型内外激励耦合动力学建模方法研究[D]. 重庆: 重庆大学, 2022.

[26]陈剑, 黄志, 徐庭亮, 等. 基于改进二进制粒子群算法优化DBN的轴承故障诊断[J]. 组合机床与自动化加工技术, 2024, (1): 168-173.

[27]周智勇. 舰船电机轴承外圈故障诊断方法[J]. 舰船科学技术, 2023, 45(24): 128-131.

[28]宋金波, 王德平, 刘霞. 基于 EMD 瞬时功率谱熵的神经网络滚动轴承故障诊断[J]. 化工自动化及仪表, 2016, 43(8): 793-796.

[29]Xu P, Ghasemloonia A, Sun Q. Automatic band selection algorithm for envelope analysis[J]. Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science, 2019, 233(5): 1641-1654.

[30]严文超, 王伟奇, 黄蓉. 基于RSSD和小波变换的滚动轴承故障诊断[J]. 武汉工程大学学报, 2019, 41(4): 399-404.

[31]Richa S, Ravi M, Shri S. Comparison between FPGA Implementation of discrete wavelet transform, dual tree complex wavelet transform and double density dual tree complex wavelet transform in verilog HDL[J]. International Journal of Trend in Scientific Research and Development, 2018, 2(4): 1153-1156.

[32]邹腾枭, 王林军, 刘洋, 等. 基于WTD和CEEMD的轴承故障特征提取方法[J]. 机床与液压, 2023, 51(11): 194-198.

[33]朱琴跃, 于逸尘, 占岩文, 等. 基于短时傅里叶变换和深度网络的模块化多电平换流器子模块IGBT开路故障诊断[J/OL]. 电工技术学报, 2024: 1-15.

[34]丁小健, 周健, 梁超, 等. 基于小波降噪和共振解调法的异步电机故障诊断方法[J]. 电机与控制应用, 2020, 47(9): 106-110.

[35]李俊卿, 刘静. 结合卷积神经网络和迁移学习的电机轴承故障诊断方法[J/OL]. 华北电力大学学报(自然科学版), 2023, 50(01):76-83+91.

[36]Amarnath M, Krishna I R P. Local fault detection in helical gears via vibration and acousticsignals using EMD based statistical parameter analysis[J]. Measurement, 2014, 58: 154-164.

[37]Luo C X, Jing S X, Yang X, et al. Applying weak signal fault diagnosis method for bearing outer ring based on EMD and spectral kurtosis[J]. Electronic Journal of Geotechnical Engineering, 2016, 21: 9753-9765.

[38]Jiang F, Zhu Z C, Li W, et al. A fusion feature extraction method using EEMD and correlation coefficient analysis for bearing fault diagnosis[J]. Applied Science, 2018, 8(9): 1621.

[39]Liu C L, Tan J P. An enhanced variational mode decomposition based on correntropy and a periodicity-assisted log-cycligram for bearing fault diagnosis[J]. Measurement Science and Technology, 2022, 33(6): 065108.

[40]Wang C G, Li H K, Huang G J, et al. Early fault diagnosis for planetary gearbox based on adaptive parameter optimized VMD and singular kurtosis difference spectrum[J]. IEEE Access, 2019, 7: 31501-31516.

[41]Ding J K, Huang L P, Xiao D M, et al. GMPSO-VMD algorithm and its application to rolling bearing fault feature extraction[J]. Sensors, 2020, 20(7): 1946.

[42]彭成, 李凤娟, 蒋金元. 基于深度学习的故障诊断与预测方法综述[J]. 现代电子技术, 2022, 45(3): 111-120.

[43]王健, 邬娜, 杨建伟, 等. 基于改进DANN和注意力机制的轴箱故障诊断方法[J/OL]. 机床与液压, 2024: 1-9.

[44]邹筱瑜, 孙国庆, 王忠宾, 等. 基于时频融合深度网络的矿用钻机轴承故障诊断[J/OL]. 中国机械工程, 2024: 1-10.

[45]曾敏敏, 张兵, 张晓宁, 等. 决策级融合轴承故障混合智能诊断[J]. 中国测试, 2024, 50(2): 126-135.

[46]谢苗, 孟庆爽, 李博, 等. 融合短时傅里叶变换和卷积神经网络的托辊故障诊断方法[J/OL]. 工程设计学报, 2024: 1-11.

[47]沈凯, 许洁. 结合深度嵌入度量学习和LSTM的装备故障诊断技术研究[J]. 中国机械, 2023, (27): 2-9.

[48]刘雨轩, 王琳, 张鹏镇, 等. 基于深度置信网络的轴承故障识别分析与研究[J]. 沈阳工程学院学报(自然科学版), 2023, 19(4): 84-89.

[49]钱存元, 吴昊, 陈昊然. 基于EMD-BPNN的逆变器IGBT开路故障诊断方法[J/OL]. 电气工程学报, 2023: 1-11.

[50]李超, 郭瑜. 基于SK等指标和SVM的滚动轴承性能退化评估研究[J]. 电子科技, 2020, 33(1): 6-12.

[51]Parmar U, Pandya D H. Experimental investigation of cylindrical bearing fault diagnosis with SVM[J]. Materials Today: Proceedings, 2021, 44(1): 1286-1290.

[52]汤捷中. 基于遗传算法优化BP神经网络的内燃机轴承故障诊断方法[J]. 内燃机与配件, 2024, (3): 19-21.

[53]熊军华, 师刘俊, 康义. 基于灰狼算法优化支持向量机的变压器故障诊断[J]. 信息技术与信息化, 2020(11): 144-147.

[54]路照坭, 朱希安. 机械传动电机轴承故障信号诊断仿真研究[J]. 自动化仪表, 2019, 40(9): 46-51.

[55]李国友, 杨梦琪, 杭丙鹏, 等. 基于模糊粗糙集和鲸鱼优化支持向量机的化工过程故障诊断[J]. 振动与冲击, 2022, 41(2): 177-184.

[56]栾孝驰, 李彦徵, 徐石等. 基于小波包变换与CEEMDAN的滚动轴承故障诊断方法[J/OL]. 航空动力学报: 2022: 1-16.

[57]李可, 熊檬, 宿磊, 等. 基于改进深层极限学习机的故障诊断方法[J]. 振动测试与诊断, 2020, 40(6): 1120-1127+1232.

[58]乔文山, 华晋, 楼荣. 改进蝴蝶算法优化核极限学习机的柴油机故障诊断[J]. 机械设计与研究, 2022, 38(1): 211-214.

[59]赵云, 宿磊, 李可, 等. 基于改进MPE与KELM的滚动轴承故障诊断[J]. 噪声与振动控制, 2022, 42(1): 125-131.

[60]朱明新, 尚凯, 杨兴园, 等. 基于随机森林算法的汽车轴承故障检测[J]. 机床与液压, 2020, 48(1): 179-182.

[61]Zhang W, Peng G L, Li C H, et al. A new deep learning model for fault diagnosis with good anti-noise and domain adaptation ability on raw vibration signals[J]. Sensors, 2017, 17(2): 425.

[62]Wen L, Li X Y, Gao L, et al. A new convolutional neural network-based data-driven fault diagnosis method[J]. IEEE Transactions on Industrial Electronics, 2017, 65(7): 5990-5998.

[63]许爱华, 杜洋, 袁涛. 基于深度学习的电机轴承故障诊断研究[J]. 组合机床与自动化加工技术, 2020, (3): 45-48+54.

[64]Aljemely A H , Xuan J , Al-Azzawi O , et al. Intelligent fault diagnosis of rolling bearings based on lstm with large margin nearest neighbor algorithm[J]. Neural Computing and Applications, 2022, 34(22): 19401-19421.

[65]Tang Z, Bo L, Liu X F, et al. A semi-supervised transferable lstm with feature evaluation for fault diagnosis of rotating machinery[J]. Applied Intelligence, 2022, 52(2): 1703-1717.

[66]Li Y B, Zou W T, Li J. Fault diagnosis of rotating machinery based on combination of wasserstein generative adversarial networks and long short term memory fully convolutional network[J]. Measurement, 2022, 191: 110826.

[67]谢子峰. 基于电气特征分析的电机轴承故障诊断方法研究[D]. 北京: 华北电力大学, 2022.

[68]陈超, 曾昭洋, 罗军, 等. 航空发动机主轴轴承失效模式分析[J]. 润滑与密封, 2020, 45(3): 126-131.

[69]龚俊, 张月义, 陈思戢, 等. 基于SWT与改进卷积神经网络的轴承故障诊断[J]. 现代电子技术, 2024, 47(6): 68-74.

[70]王恒迪, 邓四二, 杨建玺, 等. 基于参数优化变分模态分解的滚动轴承早期故障诊断[J]. 振动与冲击, 2020, 39(23): 38-46.

[71]吴浩原, 熊辛, 闵卫东, 等. 基于多级特征融合和时域扩展的行为识别方法[J]. 计算机工程与应用, 2023, 59(7): 134-142.

[72]赵炳锡, 冀大伟, 袁奇, 等. 采用时域与时频域联合特征空间的转子系统摩擦故障诊断[J]. 西安交通大学学报, 2020, 54(1): 75-84.

[73]王望望, 邓林峰, 赵荣珍, 等. 集成KPCA与t-SNE的滚动轴承故障特征提取方法[J]. 振动工程学报, 2021, 34(2): 431-440.

[74]张天瑞, 李金洋. 基于自适应VMD-KPCA特征提取与SSA-SVM方法的滚动轴承故障诊断[J]. 机械设计, 2022, 39(7): 63-73.

[75]陈铁, 吕长钦, 张欣, 等. 基于KPCA-WPA-SVM的变压器故障诊断模型[J]. 电测与仪表, 2021, 58(4): 158-164.

[76]王昱皓, 武建文, 马速良, 等. 基于核主成分分析-Soft Max的高压断路器机械故障诊断技术研究[J]. 电工技术学报, 2020, 35(S1): 267-276.

[77]回立川, 曹威, 闫康, 等. 基于IGJO-VMD-SVM的电机轴承故障诊断研究[J/OL]. 控制工程, 2024: 1-10.

[78]王力, 刘学朋, 张亦弛. 基于BCL-ASA-BP神经网络的模拟电路板芯片故障诊断[J]. 电子测量技术, 2022, 45(14): 164-171.

[79]Fan H W, Xue C Y, Ma J T, et al. A novel intelligent diagnosis method of rolling bearing and rotor composite faults based on vibration signal-to-image mapping and CNN-SVM[J]. Measurement Science and Technology, 2023, 34(4): 044008.