计算机科学
涡轮机
残余物
小波
振动
变压器
风力发电
小波变换
模式识别(心理学)
连续小波变换
人工智能
频域
断层(地质)
离散小波变换
特征提取
语音识别
声学
状态监测
稳健性(进化)
故障检测与隔离
方位(导航)
时频分析
控制理论(社会学)
时域
解耦(概率)
频带
特征(语言学)
作者
Yan Zhang,Yunfei Zhang,Xiaoqiang Zhao
标识
DOI:10.1088/1361-6501/aeb4e1
摘要
Wind turbine bearing fault diagnosis under low signal-to-noise ratio (SNR) conditions remains challenging because vibration signals are often affected by rotational-speed fluctuations, load variations, and non-stationary background interference. These factors may weaken fault-related impulsive components, disturb time–frequency energy distributions, and reduce the discriminability of fault features. To address these issues, this paper proposes a learnable wavelet-enhanced localized Vision Transformer (ViT) framework, termed LW-ViT. First, to improve the adaptability of time–frequency representation (TFR) to spectral variations caused by changes in operating conditions, a learnable continuous wavelet transform (LCWT) module is developed to adaptively construct TFRs of vibration signals. Second, to reduce the propagation of noise-dominated responses caused by unrestricted global token interaction, a localized Transformer attention mechanism is introduced to constrain token interactions within neighboring time–frequency regions. Third, to suppress the accumulation of noise-sensitive responses in deeper feature representations, a noise-gated multi-layer perceptron block is incorporated to regulate the contribution of the residual feed-forward pathway. Experiments are conducted on an in-house Lanzhou University of Technology (LUT) wind turbine bearing dataset under different SNR levels, rotational speeds, and load conditions. The CWRU dataset is further used as a public benchmark for supplementary validation of the general applicability of the proposed framework. On the LUT dataset, LW-ViT achieves classification accuracies of 93.3% and 98.4% at −6 dB and −4 dB, respectively. Comparative experiments, ablation studies, and feature visualizations indicate that the proposed framework improves adaptive TFR, localized feature interaction, and fault classification under low-SNR conditions.
科研通智能强力驱动
Strongly Powered by AbleSci AI