To address the challenges of data imbalance and difficult feature extraction in fault diagnosis of wind turbine pitch systems, a fault diagnosis method was proposed based on an improved conditional generative adversarial network(ICGAN) and a multimodal feature fusion network(MMFNet). Based on the traditional conditional generative adversarial network(CGAN) architecture, an attention layer was introduced to strengthen the model's ability to capture key information from input features. An autoregressive model was adopted to generate random noise, which improved the diversity and authenticity of generated samples. A reward mechanism was embedded to optimize the model training process and enhance the quality of generated samples. Based on this, the MMFNet model was constructed to achieve accurate diagnosis of multiple faults in wind turbine pitch systems. Experimental results demonstrate that the proposed ICGAN-MMFNet model achieves precision and F1-score above 0.96, as well as recall above 0.97 under various sample quantity ratios. The effectiveness and generalization of the proposed fault diagnosis model were verified when dealing with imbalanced datasets.
针对风电机组故障诊断中的数据不平衡问题,提出ICGAN-MMFNet联合诊断框架,结构见图2。首先,对传统 CGAN 进行针对性优化,引入自注意力机制与噪声生成模型(ARM),并融入奖励机制——依据生成样本与真实样本的相似度反馈,动态引导生成器精准学习原始数据的特征分布,实现高质量相似样本的定向生成,解决传统生成模型样本质量低、特征拟合不充分的问题。其次,MMFNet的设计突破传统分类网络逻辑:一方面内置自注意力层以强化关键特征捕捉能力,另一方面结合Focal Loss 构建专属分类目标函数,采用“生成样本特征迁移学习”策略,分类函数仅学习源自原始数据特征的生成样本特征,不直接依赖原始数据参与特征学习,在验证阶段通过原始数据验证模型泛化能力,既规避原始不平衡数据对特征学习的干扰,又保障模型真实诊断性能。
ARM是一种用于优化生成器输入噪声分布的模型,在生成长序列时能够逐步优化噪声,避免误差累积,从而生成更高质量的噪声序列[21]。将其设为生成器损失的负向形式,构建二者强关联的对抗博弈关系;以反向激励约束生成器生成贴合真实分布的样本,强化生成对抗动态平衡、规避训练失衡,同时依托该机制提升生成器性能,保障生成样本的真实度与多样性。同时将 ARM 与自注意力机制、奖励机制结合,动态引导生成器精准学习原始数据特征分布,实现高质量相似样本定向生成,有效解决传统生成模型样本质量低、特征拟合不充分的问题。其数学表达式为
CHENXuefeng, GUOYanjie, XUCaibin, et al. Review of Fault Diagnosis and Health Monitoring for Wind Power Equipment[J]. China Mechanical Engineering, 2020, 31(2): 175-189.
[3]
LANJianyu, CHENNaichao, LIHongchuan, et al. A Review of Fault Diagnosis and Prediction Methods for Wind Turbine Pitch Systems[J]. International Journal of Green Energy, 2024, 21(7): 1613-1640.
[4]
DAOP B. Condition Monitoring and Fault Diagnosis of Wind Turbines Based on Structural Break Detection in SCADA Data[J]. Renewable Energy, 2022, 185: 641-654.
LIKe, HEJianguang, SULei, et al. Fault Diagnosis Method for Rolling Bearing Based on CAE-GAN[J]. Journal of Vibration and Shock, 2023, 42(23): 65-70.
[7]
MISRAS, KUMARS, SAYYADS, et al. Fault Detection in Induction Motor Using Time Domain and Spectral Imaging-based Transfer Learning Approach on Vibration Data[J]. Sensors, 2022, 22(21): 8210.
PANHaiyang, XUHaifeng, ZHENGJinde, et al. Mechanical Fault Diagnosis Method Based on Twin Weighted Imbalanced Matrix Classifier[J]. Journal of Mechanical Engineering, 2024, 60(3): 170-180.
[10]
WANGXianbo, CHENHao, ZHAOJing, et al. Wind Turbine Fault Diagnosis for Class-imbalance and Small-size Data Based on Stacked Capsule Autoencoder[J]. IEEE Transactions on Industrial Informatics, 2024, 20(11): 12694-12704.
[11]
ZENGXiangjun, XIALingqin, FENGChen, et al. An Improved Conditional Generative Adversarial Method for Wind Turbine Gearbox Fault Diagnosis with Imbalance Data[J]. IEEE Sensors Journal, 2024, 24(23): 38727-38739.
[12]
ZHANGYuxian, QIAOLikui, ZHAOMengru. Fault Diagnosis for Wind Turbine Generators Using Normal Behavior Model Based on Multi-task Learning[J]. IEEE Transactions on Automation Science and Engineering, 2024, 21(2): 1258-1270.
[13]
LIUYirong, WUZidong, WANGXiaoli. Research on Fault Diagnosis of Wind Turbine Based on SCADA Data[J]. IEEE Access, 2020, 8: 185557-185569.
[14]
MALDONADO-CORREAJ, TORRES-CABRERAJ, MARTÍN-MARTÍNEZS, et al. Wind Turbine Fault Detection Based on the Transformer Model Using SCADA Data[J]. Engineering Failure Analysis, 2024, 162: 108354.
[15]
HUANGDajian, ZHANGWenan, GUOFanghong, et al. Wavelet Packet Decomposition-based Multiscale CNN for Fault Diagnosis of Wind Turbine Gearbox[J]. IEEE Transactions on Cybernetics, 2023, 53(1): 443-453.
[16]
ZHONGQuan, LIUShuai, LIUChangliang, et al. Fault Diagnosis of Wind Turbine Gearboxes Based on Multisource Signal Fusion[J]. IEEE Transactions on Instrumentation and Measurement, 2025, 74: 3526413.
[17]
SUXiangjing, DENGChao, SHANYanhao, et al. Fault Diagnosis Based on Interpretable Convolutional Temporal-spatial Attention Network for Offshore Wind Turbines[J]. Journal of Modern Power Systems and Clean Energy, 2024, 12(5): 1459-1471.
WEIShurong, YINShijie, YANMengfei, et al. Offshore Wind Turbine Fault Data Enhancement and Diagnosis Based on an Improved Generative Adversarial Network[J]. Power System Protection and Control, 2025, 53(1): 114-124.
[20]
XIONGJundi, CUIShihai, TANGHaihong. A Novel Intelligent Bearing Fault Diagnosis Method Based on Signal Process and Multi-kernel Joint Distribution Adaptation[J]. Scientific Reports, 2023, 13: 4535.
[21]
WANGXing, ZHANGHan, DUZhaohui. Multiscale Noise Reduction Attention Network for Aeroengine Bearing Fault Diagnosis[J]. IEEE Transactions on Instrumentation and Measurement, 2023, 72: 3513810.
[22]
JINZhenzhen, CHENDiao, HEDeqiang, et al. Bearing Fault Diagnosis Based on VMD and Improved CNN[J]. Journal of Failure Analysis and Prevention, 2023, 23(1): 165-175.
[23]
XIANGWeijia, WUYunru, PENGCheng, et al. A Fault Diagnosis Method for Electric Check Valve Based on ResNet-ELM with Adaptive Focal Loss[J]. Electronics, 2024, 13(17): 3426.
QIAOHuihui, ZHAOErxian, HAORujiang,et al. Operating condition Domain Generalization Fault Fiagnosis Method for Bearings Based on a Multiscale Convolutional Neural Network and the Two-stage Attention Mechanism[J]. Journal of Vibration and Shock, 2025, 44(2) : 267-278.
[26]
HASSANIH, MASHHADL M, ROYER-CARENZIM, et al. White Noise and Its Misapplications: Impacts on Time Series Model Adequacy and Forecasting[J]. Forecasting, 2025, 7(1): 8.