With the rapid development and increasing energy production capacity of high-power wind turbines, a corresponding increase in maintenance requirements has been observed. Reducing the failure rate of wind turbines is a critical objective, alongside decreasing affiliated operation and maintenance costs. This review focuses on the status monitoring, fault diagnosis, fault prediction, and status evaluation of wind turbines. The early fault diagnosis of wind turbines is explored with regard to existing condition monitoring technology. Moreover, the current mathematics-based fault diagnosis and smart fault diagnosis technologies are further explored. Through comprehensive investigation, this paper summarizes the research status of wind turbine fault prediction and complete machine status evaluation, conclusively presenting relevant research points and trends in the fault diagnosis, fault prediction, and status assessment of high-power wind turbines.
31 January 2025 in Measurement Science and Technology
Self-supervised learning (SSL) has been widely used for fault diagnosis of rotating machinery. However, real industrial environments often generate a large amount of noise, and existing SSL-based fault diagnosis methods are, firstly, difficult to apply to fault diagnosis under highly noisy working conditions due to the inherent scarcity of labeled samples. Secondly, most of the existing methods add projection head during pre-training and classification head during fine-tuning, such a design makes the number of parameters and computations of the model excessive, which requires a lot of time for training and cannot meet the needs of low-end edge device deployment. To solve the above problems, this paper proposes a multi-scale channel-mixed depthwise separable convolutional neural network-based fault diagnosis method under SSL. Experimental results on bearing datasets show that the proposed method has a significant advantage over other methods in terms of strong noise fault diagnosis accuracy under limited labeled samples, and also the number of parameters and computational cost of the model are minimized. Specifically, with only 10 pre-training epochs and 50 fine-tuning epochs under 5% labeled data, the proposed method can achieve up to 96.35% average accuracy on the PU dataset with strong noise interference.
Pervasive noise and measurement disturbances in industrial environments pose significant challenges to reliable fault diagnosis, driving continuous advancements in noise-robust fault diagnosis. However, the literature currently lacks a unified review synthesizing both the technical evolution and emerging research trajectories in this domain. To bridge this gap, this paper analyzes 1180 relevant studies published between 2000 and 2025. Through a bibliometric framework, we map the overarching research landscape and categorize existing methodologies into three primary paradigms: mechanism-based approaches, data-driven models, and hybrid mechanism-data fusion strategies. In addition, We also investigate robust diagnostic learning under label noise. Recognizing the limitations of current laboratory-focused models, we outline critical directions necessary for real-world deployment. These include bridging the reality gap between simulated and real-world noise, overcoming data scarcity, managing cross-domain distribution shifts, resolving long-tailed class imbalances, and developing lightweight architectures for edge computing. By articulating these specific constraints, this review aims to accelerate the transition of diagnostic systems into complex, noise-intensive industrial applications.
Rotating machinery continuously generates unlabeled monitoring data during long-term operation, and the data distribution dynamically evolves with changing operating conditions. Conventional static offline diagnostic methods lack continuous adaptation capability, making it difficult to simultaneously handle newly emerging fault categories and cross-condition distribution shifts. This paper investigates class-incremental unsupervised domain adaptation for fault diagnosis (CIUDA-FD). In this setting, a diagnostic model is required to incrementally learn emerging fault categories from an unlabeled target-domain data stream while retaining previously acquired diagnostic knowledge. To address this challenge, this paper proposes a reliable shared-class identification and selective adversarial alignment (ReSCI-SAA) framework. Specifically, a tri-criteria shared-class identification (TriSCI) module with confidence fusion is first designed to identify reliable shared classes by jointly considering cumulative prediction probability, intra-domain clustering compactness, and inter-domain similarity. Based on the identified shared classes, a pseudo-label-guided selective adversarial alignment strategy is developed to align reliable class distributions across domains and reduce negative transfer from unreliable target samples. Furthermore, a prototype-accumulated knowledge distillation mechanism is introduced to mitigate catastrophic forgetting during class-incremental adaptation. Experimental results on the public CWRU bearing dataset and a self-built motor fault dataset demonstrate that the proposed method achieves superior average diagnostic performance over the best-performing existing method across multiple transfer tasks.