Vol. 59 No. 5 (2025)

Open Access

Article

Article ID: 3354

DRSEL: A dual branch feature level ensemble learning framework based on multi sensor data fusion for fault diagnosis of rotating machines

by Yuan Zhuang, Wei Qu, Junjie Ding, Minling Pan, Jiahua Su

Sound & Vibration, Vol.59, No.5, 2025;

This paper proposes a feature-level ensemble learning framework for fault diagnosis of rotating machinery based on multi-sensor data fusion, aiming to address the inherent limitations of conventional single-model or single-sensor approaches in capturing complex and variable fault characteristics. Firstly, an improved ensemble learner is designed by integrating residual networks with Swin Transformer blocks, enabling the extraction of multi-scale, hierarchical, and complementary features from multi-sensor vibration signals. This design allows the model to capture local fine-grained patterns while simultaneously learning long-range dependencies, thus achieving a more comprehensive and discriminative feature representation. Then, the feature vectors produced by multiple base learners are concatenated to perform feature-level fusion, which effectively leverages the complementary information across heterogeneous sensors and significantly enhances diagnostic accuracy, robustness, and stability. Finally, the proposed framework is validated through two real-world industrial case studies involving a train bogie gearbox and an induction motor, covering diverse operating speeds, load conditions, and fault types. Experimental results demonstrate that the method achieves a fault diagnosis accuracy exceeding 96.88%, markedly outperforming traditional single-model approaches and conventional fusion strategies. Moreover, the framework exhibits strong generalization ability under variable working conditions. These findings highlight its practical applicability in industrial scenarios and underline its potential to support the development of intelligent and reliable predictive maintenance systems.

Open Access

Article

Article ID: 3662

Creating a vibrational digital-twin of a Bell UH-1H helicopter tail-rotor blade for use in simulating centrifugal stiffening

by Daniel Winarski, Marc Lamparelli, Keith Landry, Tyson Winarski

Sound & Vibration, Vol.59, No.5, 2025;

Our objective was to study the centrifugal stiffening in the tail-rotor of a Bell UH-1H helicopter under operational rotation. Our method involved creating a digital-twin of the vibrations of a stationary model of this tail-rotor blade using empirical modal analysis, and STAR7 software from Spectral Dynamics to process 45 empirical frequency response functions.The results from the STAR7 modal analysis identified three out-of-plane flapping modes as well as a torsional mode of vibration. We assessed our digital-twin via the use of the Modal Assurance Criterion (MAC) as well as comparisons to analytical Euler-Bernoulli beam theory and ANSYS finite element analysis simulations. Then we augmented this stationary model to an operational rotating velocity of 173.2 rad/sec via use of the Structural Dynamics Modification (SDM) feature of STAR7 to enable quantification of the centrifugal stiffening of the tail-rotor across the blade’s primary flapping and torsional modes. We concluded that the first flapping mode of vibration had the most centrifugal stiffening, and the succeeding modes experienced less stiffening as the modal frequency increased, which was consistent with modal energy being relatively constant across these modes. The digital twin approach demonstrated excellent agreement with analytical and numerical models, indicating its effectiveness for evaluating rotational stiffening in helicopter rotor components. Overall, the methodology demonstrated that empirical digital twin construction, combined with SDM techniques, provided a means of predicting the dynamic behavior of helicopter rotor components under rotation. This approach may serve as a foundation for applying artificial intelligence to rotorcraft diagnostics, structural health monitoring, and predictive maintenance.

Open Access

Article

Article ID: 3658

Intelligent fault detection of zero-sample rolling bearings driven by combined time-frequency analysis and multimodal knowledge

by Haifeng Wei, Huijian Que, Rongxiang Zheng, Jianbin Liao, Guoqiang Li, Jin Yan

Sound & Vibration, Vol.59, No.5, 2025;

To meet the demand for intelligent monitoring of rolling bearings and overcome the limitation of scarce fault samples in the real world, this paper proposes a zero-fault sample-based condition detection method that integrates the time-frequency analysis of rolling bearings with modal knowledge, which mainly includes: 1) The HHT envelope analysis and high-frequency filtering is introduced to reduce the interference of the base-frequency information of rolling bearings and to enhance the frequency of the fault information. 2) A novel zero-fault sample-based loss function is designed by combining the strong temporal sequence of rolling bearing monitoring data with the a priori knowledge of information mutualism to realize the effective training of the data-driven model. 3) An intelligent fault detection algorithm for rolling bearings is established based on a trained data-driven model. The proposed method is validated using a constructed rolling bearing experimental platform. The validation results show that the proposed method is able to build an effective intelligent fault detection model with zero fault samples of rolling bearings, which shows better fault detection performance than other supervised and unsupervised learning-based methods. The proposed anomaly detection method based on zero-fault samples can quickly establish its state detection model for important rolling bearings applied in engineering practice, providing a new perspective for data-driven fault diagnosis methods for rolling bearings.

Open Access

Article

Article ID: 3745

Noise level in a cow milking parlor

by Dimo Dimov, Toncho Penev, Ivaylo Marinov

Sound & Vibration, Vol.59, No.5, 2025;

The study took place in a dairy cattle farm with 500 Holstein-Friesian cows. Animals were reared under conditions of a freestall housing system and milked in a 2 × 8 "Herringbone" type milking parlor. Noise level reporting was performed three times during each milking (at start, in the middle, and at the end of milking) during the morning, midday, and evening milkings, every month within one year. The noise level in the working environment was measured by means of a Lutron SL-4023SD sound meter. The highest average noise values were recorded during the winter season, especially during the midday and evening milking, 75–76 dB, with deviations reaching over 80 dB. The next season, in terms of noise level, was the summer season, with average values of 72–74 dB. A study on noise levels in a “fishbone” type milking parlor found average values corresponding to moderately high noise levels, exceeding 65–70 dB. Such levels may negatively affect the welfare of dairy cows, as maximum values above the permissible limits were also recorded, particularly during the winter season. Therefore, it is recommended to optimize technological processes in order to reduce noise levels during milking as much as possible, which is essential both for the operators (milkers) and for the comfort of the animals.

Open Access

Article

Article ID: 2531

Nonlinear vibration of thin-walled box beams incorporating axial displacement, torsion distortion and secondary torsion

by Minyao Tan, Dequan Guo, Bo Liu, Yu Liu

Sound & Vibration, Vol.59, No.5, 2025;

In this paper, the boundary element method for nonlinear vibration of thin-walled box beams is established, including axial displacement, torsion, distortion, and quadratic torsion. The thin-walled box girders are subjected to conservative dynamic torsional and warping moments that are arbitrarily distributed or concentrated along the length direction. Its control differential equations and boundary conditions are represented by torsion, distortion, and secondary torsional deformation. Its nonlinear terms show strong coupling. The coupling effects of axial displacement, torsion, distortion, and quadratic torsion deformation are fully considered, and the vibration analysis is carried out, respectively, in the state of buckling or post-buckling. The nonlinear differential algebraic equations can be obtained by using the simulation equation method based on the initial boundary value problem, and solved by using an effective time-discrete scheme. In addition, it is proven by two examples that the nonlinear term has a great influence on the torsional moment and the mode of the thin-walled box girder under free vibration. Large torsional rotations increase the torsional stiffness of the thin-walled box beam, ultimately leading to a higher natural frequency. Although the influence of the distortion moment is not as good as that of the torsion moment, it cannot be ignored. The axial inertia term has an obvious influence on the axial stress.

Open Access

Article

Article ID: 3788

Coupling vibration analysis for inspection robot landing on high voltage transmission line

by Xiaodong Zhang, Haiming Shen, Ahmad Bala Alhassan, Haibo Xu

Sound & Vibration, Vol.59, No.5, 2025;

When the high-voltage transmission line inspection robot (HVTIR) with a landing gripper lands during flight, the rigid-flexible coupling vibration between the HVTIR and the transmission line severely undermines landing stability. Therefore, investigating the HVTIR-line coupling vibration characteristics during landing is critical for achieving stable landing control. Ignoring wind and airflow disturbances, this study analyzes the input-output characteristics of the HVTIR-line coupled vibration system and the interaction forces between adjacent components, then derives their motion differential equations. The transmission line is decomposed into a model of multiple unit-length Euler beams hinged by stiffness and damping, and a stiffness-damping coupling model for the gripper-line contact force is established. On this basis, a mass-stiffness-damping model of the HVTIR-line system is constructed to obtain its rigid-flexible coupling vibration characteristics. A vibration test system is built using an acceleration sensor wireless acquisition system, and the vibration characteristic curves of the HVTIR gripper and transmission line are obtained and compared with the theoretical simulation results. The results show that the experimental and simulation data share the same trend, with a maximum error of no more than 13.64%; the interquartile range increment between them ranges from 10.67% to 18.26%, verifying the model’s scientific validity. This study optimizes the transmission line modeling method to ensure calculation accuracy and real-time analysis efficiency, and constructs a novel HVTIR-line rigid-flexible coupling model, which provides a theoretical basis for smooth and rapid HVTIR landing control.

Open Access

Article

Article ID: 3701

Analysis and control of 2f noise in an ultra-high-speed permanent magnet synchronous motor

by Guoping Feng, Ruihong Jing, Jing Yao, Jianjun Zhao, Dongya Li, Wenjian Zhou

Sound & Vibration, Vol.59, No.5, 2025;

Ultra-high-speed permanent magnet synchronous motors (PMSMs) have been widely applied in aerospace, advanced manufacturing, and consumer products due to their small size, high reliability, and high-power density. As the rotational speed increases, losses, cooling requirements, excitation forces, bearing losses, and noise characteristics of the motor are significantly different from those of conventional motors. To minimize iron losses and improve efficiency, ultra-high-speed motors typically employ a low pole count configuration, exemplified by the prevalent 2-pole 3-slot (2P3S) design. In 2-pole machines, low-order electromagnetic excitation forces may cause significant electromagnetic vibration and noise. Reducing electromagnetic noise, especially the two-times mechanical rotational frequency (2f) noise in 2P3S motors, is critical. Therefore, the noise reduction strategies for a 120,000 rpm PMSM are investigated. First, the electromagnetic excitation frequencies and magnitudes of the motor are analyzed. Through changing the pole-slot configuration (such as transitioning from 2P3S to 2P6S), the unbalanced magnetic pull (UMP) in the original 2P3S design is eliminated, significantly reducing the main source of vibration. Then, the impact of the stator assembly gap on electromagnetic excitation is examined using a segmented stator structure. Effective control of this assembly gap further reduces the electromagnetic noise of the motor. Finally, experimental tests demonstrate a significant noise reduction in the 2f frequency band, with a sound pressure level reduction of about 9 dB. The reduction validates the effectiveness of this pole-slot optimization and stator assembly gap control strategies. 

Open Access

Article

Article ID: 3676

Research on low-frequency sound insulation characteristics of metamaterials based on coupling of local resonance and Helmholtz resonators

by Wenmin Zhang, Shangshuai Jia, Xinli Zhao, Lixin Zheng, Cai Zeng

Sound & Vibration, Vol.59, No.5, 2025;

This study addresses the challenge of low-frequency noise control by proposing a novel composite acoustic metamaterial that synergistically couples a multilayer locally resonant acoustic metamaterial (MLRAM) with a dual-cavity Helmholtz resonator. The research systematically investigates the sound insulation performance through an integrated approach combining theoretical analysis, finite element simulation, and impedance tube experimentation. Results demonstrate that while the MLRAM structure generates a significant sound insulation peak of 44.5 dB at 214 Hz, it is accompanied by a pronounced insulation valley of 6.4 dB at 228 Hz. To mitigate this limitation, a dual-cavity Helmholtz resonator was designed, achieving near-perfect absorption (α > 0.99) and corresponding insulation peaks at 236 Hz and 316 Hz. The integrated composite structure effectively elevates the original insulation valley by 3.0 dB and introduces a new insulation peak of 32.0 dB at 311 Hz. Parametric studies reveal that the resonance frequencies can be precisely tuned by adjusting the neck geometry, and increasing the number of cavities broadens the effective bandwidth at the expense of peak amplitude. After optimization, the composite structure achieves a remarkable valley improvement of up to 10.8 dB and an average sound transmission loss of 21.0 dB, significantly enhancing broadband low-frequency sound insulation performance. This work provides a validated strategy for active "valley compensation" in acoustic metamaterials.