基于多尺度特征融合生成对抗网络的风电机组齿轮箱故障诊断

肖生杲; 王永

doi:10.19912/j.0254-0096.tynxb.2024-2276

PDF(3124 KB)

太阳能学报 ›› 2026, Vol. 47 ›› Issue (5) : 205-215. DOI: 10.19912/j.0254-0096.tynxb.2024-2276

基于多尺度特征融合生成对抗网络的风电机组齿轮箱故障诊断

肖生杲, 王永

作者信息 +

FAULT DIAGNOSIS FOR WIND TURBINE GEARBOXES BASED ON GENERATIVE ADVERSARIAL NETWORKS MERGING MULTI-SCALE FEATURES

Xiao Shenggao, Wang Yong

Author information +

文章历史 +

摘要

针对风电机组齿轮箱故障样本不充足、质量差导致传统深度学习模型诊断不准确的问题,提出一种基于多尺度特征融合生成对抗网络的风电机组齿轮箱故障诊断方法。首先构建多尺度特征融合模块改进生成对抗网络,用于充分学习数据样本的深层和浅层特征,克服生成对抗网络易出现类内模式崩塌的缺点,提高生成样本的质量和多样性。其次,使用Wasserstein距离和谱归一化方法制定新的损失函数,提高模型对抗训练的稳定性,进一步增强生成样本质量。将所提模型用于有限的齿轮箱故障数据集,对多类别健康状态样本进行诊断分析。实验结果表明,与现有模型相比,新模型能对有限样本进行有效扩充,有效推动了智能诊断模型的训练,故障诊断的准确度显著提高。

Abstract

To address the issue of inaccurate diagnoses by traditional deep learning models due to insufficient and poor-quality samples of wind turbine gearbox failures, the fault diagnosis based on adversarial network merging multi-scale feature （MFGAN） is proposed for fault diagnosis of wind turbine gearboxes. Firstly, the generative adversarial network （GAN） merging multi-scale feature is constructed, which can effectively learn both deep and shallow features from data. This approach overcomes the issue of mode collapse within GANs and improves the quality and diversity of the generated images. Secondly, the Wasserstein distance and spectral normalization techniques are used to build a novel loss function. It enhances the stability of the MFGAN during adversarial training while the quality of the generated data is improved. The proposed MFGAN is applied to diagnose multiple faults of gearboxes using a limited dataset. The experimental results demonstrate that MFGAN can effectively generate more useful data based on the limited dataset. Moreover, the accuracy of fault diagnosis for gearboxes of wind turbines is significantly enhanced comparing with the previous methods.

导出引用

肖生杲, 王永. 基于多尺度特征融合生成对抗网络的风电机组齿轮箱故障诊断[J]. 太阳能学报. 2026, 47(5): 205-215 https://doi.org/10.19912/j.0254-0096.tynxb.2024-2276

Xiao Shenggao, Wang Yong. FAULT DIAGNOSIS FOR WIND TURBINE GEARBOXES BASED ON GENERATIVE ADVERSARIAL NETWORKS MERGING MULTI-SCALE FEATURES[J]. Acta Energiae Solaris Sinica. 2026, 47(5): 205-215 https://doi.org/10.19912/j.0254-0096.tynxb.2024-2276

中图分类号： TK83

参考文献

[1] 魏秀业, 程海吉, 贺妍, 等. 基于特征融合与ResNet的行星齿轮箱故障诊断[J]. 电子测量与仪器学报, 2022, 36(5): 213-222.
WEI X Y, CHENG H J, HE Y, et al.Fault diagnosis of planetary gearboxes based on feature fusion and ResNet[J]. Journal of electronic measurement and instrumentation, 2022, 36(5): 213-222.
[2] 龙寰, 杨婷, 徐劭辉, 等. 基于数据驱动的风电机组状态监测与故障诊断技术综述[J]. 电力系统自动化, 2023, 47(23): 55-69.
LONG H, YANG T, XU S H, et al.Review of data-driven condition monitoring and fault diagnosis technologies for wind turbines[J]. Automation of electric power systems, 2023, 47(23): 55-69.
[3] YANG S P, GU X H, LIU Y Q, et al.A general multi-objective optimized wavelet filter and its applications in fault diagnosis of wheelset bearings[J]. Mechanical systems and signal processing, 2020, 145: 106914.
[4] MA Y L, YU G, LIN T R, et al.A high-resolution time-frequency analysis technique based on bi-directional squeezing and its application in fault diagnosis of rotating machinery[J]. ISA transactions, 2024, 147: 382-402.
[5] WANG J, QIAO W, QU L Y.Wind turbine bearing fault diagnosis based on sparse representation of condition monitoring signals[J]. IEEE transactions on industry applications, 2019, 55(2): 1844-1852.
[6] 曹沂风. 基于支持向量机的风轮不平衡故障诊断方法研究[J]. 太阳能学报, 2024, 45(8): 613-620.
CAO Y F.Study on rotor imbalance fault identification method based on multi-kernel learning support vector machine[J]. Acta energiae solaris sinica, 2024, 45(8): 613-620.
[7] XUE H T, DING D Y, ZHANG Z M, et al.A fuzzy system of operation safety assessment using multimodel linkage and multistage collaboration for in-wheel motor[J]. IEEE transactions on fuzzy systems, 2022, 30(4): 999-1013.
[8] 李俊逸, 尧远, 刘明浩. 基于振动信号最优特征提取算法的风力机齿轮箱SVM故障诊断[J]. 太阳能学报, 2024, 45(7): 626-633.
LI J Y, YAO Y, LIU M H.SVM fault diagnosis of wind turbine's gearbox based on optimal feature extraction algorithm of vibration signal[J]. Acta energiae solaris sinica, 2024, 45(7): 626-633.
[9] MERAINANI B, RAHMOUNE C, BENAZZOUZ D, et al.A novel gearbox fault feature extraction and classification using Hilbert empirical wavelet transform, singular value decomposition, and SOM neural network[J]. Journal of vibration and control, 2018, 24(12): 2512-2531.
[10] HAN T, LIU C, YANG W G, et al.A novel adversarial learning framework in deep convolutional neural network for intelligent diagnosis of mechanical faults[J]. Knowledge-based systems, 2019, 165: 474-487.
[11] KONG Y, CHU F L, QIN Z Y, et al.Sparse learning based classification framework for planetary bearing health diagnostics[J]. Mechanism and machine theory, 2022, 173: 104852.
[12] LI W X, SHANG Z W, QIAN S Q, et al.A novel intelligent fault diagnosis method of rotating machinery based on signal-to-image mapping and deep Gabor convolutional adaptive pooling network[J]. Expert systems with applications, 2022, 205: 117716.
[13] YI H K, JIANG Q C, YAN X F, et al.Imbalanced classification based on minority clustering synthetic minority oversampling technique with wind turbine fault detection application[J]. IEEE transactions on industrial informatics, 2021, 17(9): 5867-5875.
[14] WEI J N, HUANG H S, YAO L G, et al.New imbalanced fault diagnosis framework based on Cluster-MWMOTE and MFO-optimized LS-SVM using limited and complex bearing data[J]. Engineering applications of artificial intelligence, 2020, 96: 103966.
[15] 张伊杰, 刘宝良, 王承民, 等. 基于SMOTETomek过采样方法与领域自适应迁移学习的风电机组故障诊断[J]. 太阳能学报, 2024, 45(10): 635-644.
ZHANG Y J, LIU B L, WANG C M, et al.Fault diagnosis of wind turbines based on SMOTETomek oversampling method and domain adaptive transfer learning[J]. Acta energiae solaris sinica, 2024, 45(10): 635-644.
[16] GOODFELLOW I J, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial networks[EB/OL].2014: arXiv: 1406.2661. https://arxiv.org/abs/1406.2661.
[17] ZHOU F N, YANG S, FUJITA H, et al.Deep learning fault diagnosis method based on global optimization GAN for unbalanced data[J]. Knowledge-based systems, 2020, 187: 104837.
[18] ZHANG L, ZHANG H, CAI G W.The multiclass fault diagnosis of wind turbine bearing based on multisource signal fusion and deep learning generative model[J]. IEEE transactions on instrumentation and measurement, 2022, 71: 3514212.
[19] LIU J, ZHANG C H, JIANG X X.Imbalanced fault diagnosis of rolling bearing using improved MsR-GAN and feature enhancement-driven CapsNet[J]. Mechanical systems and signal processing, 2022, 168: 108664.
[20] WANG R G, CHEN Z Y, ZHANG S H, et al.Dual-attention generative adversarial networks for fault diagnosis under the class-imbalanced conditions[J]. IEEE sensors journal, 2022, 22(2): 1474-1485.
[21] ARJOVSKY M, CHINTALA S, BOTTOU L.Wasserstein generative adversarial networks[C]//Proceedings of the 34th International Conference On Machine Learning. Sydney, Australia, 2017.
[22] GULRAJANI I, AHMED F, ARJOVSKY M, et al.Improved training of Wasserstein GANs[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, California, USA, 2017: 5769-5779.
[23] MIYATO T, KATAOKA T, KOYAMA M, et al. Spectral normalization for generative adversarial networks[EB/OL].2018: arXiv: 1802.05957. https://arxiv.org/abs/1802.05957.
[24] LIU S, CHEN J L, QU C, et al.LOSGAN: latent optimized stable GAN for intelligent fault diagnosis with limited data in rotating machinery[J]. Measurement science and technology, 2021, 32(4): 045101.
[25] SHAO S Y, WANG P, YAN R Q.Generative adversarial networks for data augmentation in machine fault diagnosis[J]. Computers in industry, 2019, 106: 85-93.
[26] LI Z X, ZHENG T S, WANG Y, et al.A novel method for imbalanced fault diagnosis of rotating machinery based on generative adversarial networks[J]. IEEE transactions on instrumentation and measurement, 2021, 70: 3500417.
[27] GUO Q W, LI Y B, SONG Y, et al.Intelligent fault diagnosis method based on full 1-D convolutional generative adversarial network[J]. IEEE transactions on industrial informatics, 2020, 16(3): 2044-2053.
[28] LIU D D, CUI L L, CHENG W D.A review on deep learning in planetary gearbox health state recognition: methods, applications, and dataset publication[J]. Measurement science and technology, 2024, 35(1): 012002.
[29] YANG C, LI H K, CAO S X, et al.CE-FFGAN: a feature fusion generative adversarial network with deep embedded category information for limited sample fault diagnosis of rotating machinery under speed variation[J]. Advanced engineering informatics, 2024, 62: 102605.
[30] THOMAZELLA R, LOPES W N, AGUIAR P R, et al.Digital signal processing for self-vibration monitoring in grinding: a new approach based on the time-frequency analysis of vibration signals[J]. Measurement, 2019, 145: 71-83.
[31] 邵海东, 李伟, 林健, 等. 基于改进ACGAN的齿轮箱多模式数据增强与故障诊断[J]. 交通运输工程学报, 2023, 23(3): 188-197.
SHAO H D, LI W, LIN J, et al.Multi-mode data augmentation and fault diagnosis of gearbox using improved ACGAN[J]. Journal of traffic and transportation engineering, 2023, 23(3): 188-197.
[32] ZHANG Q L, LIAN X W, QIN J Y, et al.Research of turbine rotor fault diagnosis based on improved auxiliary classification generative adversarial network[J]. Measurement, 2025, 248: 116991.
[33] SIVAANPU A, PUNITHAKUMAR K, ZHENG R, et al.Speckle noise reduction for medical ultrasound images using hybrid CNN-transformer network[J]. IEEE access, 2024, 12: 168607-168625.
[34] REN H, LIU W Y, SHAN M C, et al.A novel wind turbine health condition monitoring method based on composite variational mode entropy and weighted distribution adaptation[J]. Renewable energy, 2021, 168: 972-980.