Operational safety and reliability have become increasingly paramount in large and complex rotating machinery in industrial manufacturing. Recent strides in data-driven fault diagnosis have significantly advanced the fields of machinery health assessment, fault identification, and prediction. This review delves into the existing literature on fault diagnosis technology for rotating machinery, categorizing the various diagnostic approaches. It further examines the evolution of fault diagnosis technology based on dimensionless indices and elucidates the five latest progressions of foundational theories: the design and optimization of dimensionless indices, extraction and features of mutual dimensionless indices, fault diagnosis technology based on dimensionless indices and artificial immune system, fusion technology of dimensionless indices in time and frequency domains, and datasets. The applications of fault diagnosis technology based on dimensionless indices in the petrochemical industry, high-speed railway, shipping and aerospace are discussed. Future directions for research are outlined across six key areas: more excellent dimensionless indices, rapid and efficient detectors, fault analysis for small-sample fault events, theoretical systems of time–frequency domain fusion diagnosis, fault diagnosis with standardized test database, and expert knowledge base. This comprehensive review bridges the research gap by systematically examining the roles and applications of dimensionless indices in the fault diagnosis of rotating machinery.

Accurate assessment of fault severity in rolling bearings under strong noise remains a critical challenge for intelligent predictive maintenance. In this study, the diagnostic task is formulated as joint fault type and severity classification, with emphasis on severity-level discrimination. In practical industrial environments, vibration signals are frequently contaminated by complex noise, including Gaussian noise, pink noise, Laplace (impulsive) noise, and combined interference, which severely mask transient fault impulses and distort time–frequency representations. To tackle this problem, we propose a Dual-Stream Gated Cross-Modal Attention Fusion Network (DS-GCMAF). The framework simultaneously processes one-dimensional (1D) raw vibration sequences and two-dimensional (2D) time–frequency representations. A bidirectional cross-modal multi-head attention mechanism is introduced to facilitate effective information interaction and alignment across heterogeneous feature spaces. Meanwhile, a quality-aware adaptive gating strategy is employed to dynamically regulate the contribution of each modality based on its reliability. In low signal-to-noise ratio (SNR) conditions, the less reliable 2D branch is selectively attenuated, and the more robust 1D stream serves as an anchor for cross-modal calibration. Extensive experiments were carried out on the Paderborn University (PU) dataset and HUSTbearing dataset under four complex noise types. On the PU dataset across varying SNR levels (−8 dB to 8 dB), DS-GCMAF achieves 85.80% accuracy at −8 dB, surpassing the strongest baseline by 2.69 percentage points. On the HUSTbearing dataset under −10 dB, the proposed method attains 74.04% accuracy under the most challenging combined noise condition, significantly outperforming existing single-stream and conventional fusion approaches. These results demonstrate the superior robustness and its effectiveness in joint fault-type and severity classification under highly noisy environments.