Yongsheng Yu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yongsheng Yu has a total of 43 co-authors. Here are the top co-authors by number of shared publications:
Zhuoran Cai
School of Marine and Energy Engineering, Wuhan University of Technology, Wuhan, China
5 shared publications- Linke Zhang
Wuhan University of Technology
4 shared publications Linke Zhang
School of Naval Architecture, Ocean and Energy Power Engineering, Wuhan University of Technology, Wuhan 430000, People’s Republic of China
3 shared publicationsBangling Li
Wuhan University of Technology
2 shared publicationsChunqiu Tang
Wuhan University of Technology
2 shared publicationsFucai Hu
School of Naval Architecture, Ocean and Energy Power Engineering, Wuhan University of Technology, Wuhan, China
2 shared publications
Latest Publications
- 1 October 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (39) , 396111
Acoustic signal enhancement algorithm for transformer based on singular value decomposition and spectral clustering
Transformers are critical components in power transmission and transformation systems, and their stable operation directly impacts the safety and reliability of the power grid. Due to its non-contact nature and ease of deployment, acoustic signal detection method has been widely adopted for power transformer condition monitoring. However, acoustic signals acquired from outdoor transformers are highly susceptible to environmental interference. In particular, the spectral overlap between cooling fan noise and transformer-generated acoustic signals poses a significant challenge for anomalous sound detection. To address this issue, we propose an acoustic signal enhancement method for transformers based on singular value decomposition (SVD) and spectral clustering. First, by leveraging the relative stability of the amplitude and phase across transformer acoustic signal frequency components over short periods, we apply threshold-based filtering to identify noisy frequency points, combining this with spectral subtraction for preliminary signal enhancement. Next, we design an SVD method incorporating elbow detection to attenuate noise while preserving weak yet informative signal components. Finally, to address the critical challenge of separating the intertwined singular values corresponding to fan noise and transformer acoustic signals, we propose a spectral clustering algorithm based on mutual information. By exploiting the amplitude filtering characteristics of SVD, this algorithm achieves an effective separation of fan noise and transformer acoustic signals. Experimental validation was performed on a field-collected dataset of transformer acoustic signals. The results demonstrate that the proposed method exhibits superior overall performance in both noise suppression and effective feature extraction.
- 1 October 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (39) , 396111
Acoustic signal enhancement algorithm for transformer based on singular value decomposition and spectral clustering
Transformers are critical components in power transmission and transformation systems, and their stable operation directly impacts the safety and reliability of the power grid. Due to its non-contact nature and ease of deployment, acoustic signal detection method has been widely adopted for power transformer condition monitoring. However, acoustic signals acquired from outdoor transformers are highly susceptible to environmental interference. In particular, the spectral overlap between cooling fan noise and transformer-generated acoustic signals poses a significant challenge for anomalous sound detection. To address this issue, we propose an acoustic signal enhancement method for transformers based on singular value decomposition (SVD) and spectral clustering. First, by leveraging the relative stability of the amplitude and phase across transformer acoustic signal frequency components over short periods, we apply threshold-based filtering to identify noisy frequency points, combining this with spectral subtraction for preliminary signal enhancement. Next, we design an SVD method incorporating elbow detection to attenuate noise while preserving weak yet informative signal components. Finally, to address the critical challenge of separating the intertwined singular values corresponding to fan noise and transformer acoustic signals, we propose a spectral clustering algorithm based on mutual information. By exploiting the amplitude filtering characteristics of SVD, this algorithm achieves an effective separation of fan noise and transformer acoustic signals. Experimental validation was performed on a field-collected dataset of transformer acoustic signals. The results demonstrate that the proposed method exhibits superior overall performance in both noise suppression and effective feature extraction.
- Open Access
- 23 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 46294-46303
ATF-MUSE: A Lightweight U-Net-Based Model With Adaptive Time-Frequency Transformer for Speech Enhancement
Speech enhancement techniques based on deep learning have demonstrated substantial improvements over traditional signal processing methods, particularly under complex non-stationary noise conditions. In this paper, we propose a novel lightweight speech enhancement model, ATF-MUSE, which is designed to more effectively model the complex interactions between temporal- and frequency-domain features, while dynamically adjusting attention weights to enhance adaptability to varying noise environments. Specifically, we introduce an ATF-Transformer module that alternately models dependencies along the temporal and frequency dimensions, thereby capturing both the global spectral patterns and long-term temporal relationships of speech signals. Moreover, we propose an Adaptive Multi-Head Attention (AMHA) mechanism that dynamically generates feature-dependent attention weights, significantly improving the model’s robustness against non-stationary noise. We evaluate the proposed ATF-MUSE model on the widely used VoiceBank-DEMAND dataset, and experimental results demonstrate that our method outperforms several state-of-the-art approaches across multiple standard evaluation metrics, including PESQ, MOS, and STOI.
- 23 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 46294-46303
Open AccessATF-MUSE: A Lightweight U-Net-Based Model With Adaptive Time-Frequency Transformer for Speech Enhancement
Speech enhancement techniques based on deep learning have demonstrated substantial improvements over traditional signal processing methods, particularly under complex non-stationary noise conditions. In this paper, we propose a novel lightweight speech enhancement model, ATF-MUSE, which is designed to more effectively model the complex interactions between temporal- and frequency-domain features, while dynamically adjusting attention weights to enhance adaptability to varying noise environments. Specifically, we introduce an ATF-Transformer module that alternately models dependencies along the temporal and frequency dimensions, thereby capturing both the global spectral patterns and long-term temporal relationships of speech signals. Moreover, we propose an Adaptive Multi-Head Attention (AMHA) mechanism that dynamically generates feature-dependent attention weights, significantly improving the model’s robustness against non-stationary noise. We evaluate the proposed ATF-MUSE model on the widely used VoiceBank-DEMAND dataset, and experimental results demonstrate that our method outperforms several state-of-the-art approaches across multiple standard evaluation metrics, including PESQ, MOS, and STOI.
- 16 March 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (14)
Multi-view time–frequency dynamic modeling and asymmetric convolutional networks for industrial anomalous sound detection
In recent years, industrial anomalous sound detection (ASD) methods based on self-supervised learning have achieved remarkable progress; however, existing approaches still provide insufficient characterization of the time–frequency dynamics and anisotropic structures of acoustic measurement signals. To address this limitation, this paper proposes an ASD method for non-contact machine-condition monitoring that integrates multi-view time–frequency representations with asymmetric convolutional structures. Specifically, log-Mel spectrograms, raw waveforms, and spectral-centroid-based statistical features are jointly exploited to construct multi-view acoustic representations, thereby simultaneously capturing local time–frequency textures, fine-grained temporal structures, and global frequency evolution. On this basis, a global energy trend encoding (GETE) module and a temporal adaptive attention (TAA) module are designed to explicitly model the energy variation sequences in time–frequency representations and their long-range temporal dependencies. In parallel, a multi-scale temporal context modeling module (MTCMM) is introduced to capture both short-term transient components and long-term periodic patterns. To account for the anisotropy of acoustic features along the time and frequency axes, multi-branch fusion and residual structures based on asymmetric convolutions are further employed to enhance the modeling capability of high-level time–frequency dynamic features. Experimental results on the DCASE 2020 Task 2 dataset show that the proposed method achieves average AUC and pAUC scores of 96.06% and 90.41%, respectively. These results consistently outperform multiple baseline methods, demonstrating the effectiveness of the proposed approach under the Task 2 evaluation protocol and suggesting its potential applicability to industrial acoustic measurements and anomalous sound detection for non-contact machine-condition monitoring.
- 16 March 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (14)
Multi-view time–frequency dynamic modeling and asymmetric convolutional networks for industrial anomalous sound detection
In recent years, industrial anomalous sound detection (ASD) methods based on self-supervised learning have achieved remarkable progress; however, existing approaches still provide insufficient characterization of the time–frequency dynamics and anisotropic structures of acoustic measurement signals. To address this limitation, this paper proposes an ASD method for non-contact machine-condition monitoring that integrates multi-view time–frequency representations with asymmetric convolutional structures. Specifically, log-Mel spectrograms, raw waveforms, and spectral-centroid-based statistical features are jointly exploited to construct multi-view acoustic representations, thereby simultaneously capturing local time–frequency textures, fine-grained temporal structures, and global frequency evolution. On this basis, a global energy trend encoding (GETE) module and a temporal adaptive attention (TAA) module are designed to explicitly model the energy variation sequences in time–frequency representations and their long-range temporal dependencies. In parallel, a multi-scale temporal context modeling module (MTCMM) is introduced to capture both short-term transient components and long-term periodic patterns. To account for the anisotropy of acoustic features along the time and frequency axes, multi-branch fusion and residual structures based on asymmetric convolutions are further employed to enhance the modeling capability of high-level time–frequency dynamic features. Experimental results on the DCASE 2020 Task 2 dataset show that the proposed method achieves average AUC and pAUC scores of 96.06% and 90.41%, respectively. These results consistently outperform multiple baseline methods, demonstrating the effectiveness of the proposed approach under the Task 2 evaluation protocol and suggesting its potential applicability to industrial acoustic measurements and anomalous sound detection for non-contact machine-condition monitoring.
- 27 December 2025
- Published by Springer Nature in Pattern Analysis and Applications
- Vol. 29 (1) , 15
Rectangular kernels for information-dense domains in environmental sound classification
- 27 December 2025
- Published by Springer Nature in Pattern Analysis and Applications
- Vol. 29 (1) , 15
Rectangular kernels for information-dense domains in environmental sound classification
- Open Access
- 21 October 2024
- Published by Springer Nature in Scientific Reports
- Vol. 14 (1) , 1-10
End-to-end underwater acoustic source separation model based on EDBG-GALR
In practical work scenarios, it is common for multiple vessels to be present in close proximity within the same water area. In such cases, the acoustic signals emitted by these vessels often overlap, causing interference and reducing the accuracy of vessel identification. This paper proposes an improved GALR end-to-end underwater acoustic source separation algorithm, with ECA-DE (Efficient Channel Attention-Deep Encoder) and Bi-GRU (Bidirectional Gated Recurrent Unit), which consists of an encoder, separation blocks, and a decoder (EDBG-GALR). Addressing the issue of GALR encoder’s limited expressiveness due to its single-layer one-dimensional convolutional layer, we introduce a new deep encoder, ECA-DE, to enhance the encoder’s capability to process temporal signals and improve the efficiency of the separator’s input. Additionally, we integrate Bi-GRU into the GALR separation block’s local modeling module to enhance the local modeling ability of sequence features, thereby reducing the model’s computational and parameter requirements. Experimental results demonstrate that the proposed EDBG-GALR method can effectively separate mixed multi-target signals into multiple single-target signals, achieving a maximum scale-invariant signal-to-noise ratio (SI-SNR) improvement of 3.32 dB and a signal distortion ratio (SDR) improvement of 3.38 dB over baseline methods. These results highlight the practical applicability of EDBG-GALR in complex underwater environments.
- 21 October 2024
- Published by Springer Nature in Scientific Reports
- Vol. 14 (1) , 1-10
Open AccessEnd-to-end underwater acoustic source separation model based on EDBG-GALR
In practical work scenarios, it is common for multiple vessels to be present in close proximity within the same water area. In such cases, the acoustic signals emitted by these vessels often overlap, causing interference and reducing the accuracy of vessel identification. This paper proposes an improved GALR end-to-end underwater acoustic source separation algorithm, with ECA-DE (Efficient Channel Attention-Deep Encoder) and Bi-GRU (Bidirectional Gated Recurrent Unit), which consists of an encoder, separation blocks, and a decoder (EDBG-GALR). Addressing the issue of GALR encoder’s limited expressiveness due to its single-layer one-dimensional convolutional layer, we introduce a new deep encoder, ECA-DE, to enhance the encoder’s capability to process temporal signals and improve the efficiency of the separator’s input. Additionally, we integrate Bi-GRU into the GALR separation block’s local modeling module to enhance the local modeling ability of sequence features, thereby reducing the model’s computational and parameter requirements. Experimental results demonstrate that the proposed EDBG-GALR method can effectively separate mixed multi-target signals into multiple single-target signals, achieving a maximum scale-invariant signal-to-noise ratio (SI-SNR) improvement of 3.32 dB and a signal distortion ratio (SDR) improvement of 3.38 dB over baseline methods. These results highlight the practical applicability of EDBG-GALR in complex underwater environments.
- 20 October 2024
- Published by Springer Nature
- p. 205-215
Blind Source Separation Based on Neurally Plausible Alternating Optimization-Based Online Dictionary Learning (NOODL)
- 20 October 2024
- Published by Springer Nature
- p. 205-215
Blind Source Separation Based on Neurally Plausible Alternating Optimization-Based Online Dictionary Learning (NOODL)
- 20 October 2024
- Published by Springer Nature
- p. 176-187
Speech Enhancement Method Based on Fusion Attention with Local Recurrence
- 20 October 2024
- Published by Springer Nature
- p. 176-187
Speech Enhancement Method Based on Fusion Attention with Local Recurrence
- 20 October 2024
- Published by Springer Nature
- p. 110-121
Acoustic Fault Identification in Maritime Vessels with Limited Data Using WGAN-Based Approach
- 20 October 2024
- Published by Springer Nature
- p. 110-121
Acoustic Fault Identification in Maritime Vessels with Limited Data Using WGAN-Based Approach
- 20 October 2024
- Published by Springer Nature
- p. 507-518
SRP Sound Source Localization Algorithm Based on BSO and Joint Weighting
- 20 October 2024
- Published by Springer Nature
- p. 507-518
SRP Sound Source Localization Algorithm Based on BSO and Joint Weighting
- 3 September 2024
- Published by Springer Nature
- p. 605-616
Grapevine Leaves Recognition Based on IP-ShuffleNet
- 3 September 2024
- Published by Springer Nature
- p. 605-616
Grapevine Leaves Recognition Based on IP-ShuffleNet