Sheng Wang

Sheng Wang

Auto-generated profile, based on publicly available information

Sheng Wang

School of Mechanical Engineering, Southeast University, Nanjing, China

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 January 2022 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
SWSheng Wang
Sheng Wang
ZLZhitao Luo
Zhitao Luo
PSPeng Shen
Peng Shen
HZHui Zhang
Hui Zhang
ZNZhonghua Ni
Zhonghua Ni
Interpretation of guided wave signals is a central challenge for ultrasonic guided wave-based damage detection and localization technology. Because of the complexity of the guided waves that are scattered from structural damage, existing guided wave-based damage detection methods cannot be used to extract the relationship information hidden in the guided waves for use in damage detection and localization. A graph-in-graph convolutional network is thus proposed for guided wave-based damage detection and localization that constructs spatial–temporal feature representations of the guided wave signals and interconnects them into a global graph to indicate the inherent differences among these signals. By converting the guided wave characteristics into structural and topological information in non-Euclidean space, the proposed method correlates the global graph with the damage location directly and achieves greater damage detection accuracy with fewer training data. Validations are performed using two different experimental datasets, which were collected from aluminum plates and a composite laminate. The results indicate that the proposed method achieves superior performance with high accuracy and stability for even limited and imbalanced datasets acquired with only three transducers.

9 May 2023 in Chinese Journal of Mechanical Engineering
Verified
Journal Article
Open access
PDF file
SZShengkang Zong
Shengkang Zong
SWSheng Wang
Sheng Wang
ZLZhitao Luo
Zhitao Luo
XWXinkai Wu
Xinkai Wu
HZHui Zhang
Hui Zhang
ZNZhonghua Ni
Zhonghua Ni
Ultrasonic guided wave is an attractive monitoring technique for large-scale structures but is vulnerable to changes in environmental and operational conditions (EOC), which are inevitable in the normal inspection of civil and mechanical structures. This paper thus presents a robust guided wave-based method for damage detection and localization under complex environmental conditions by singular value decomposition-based feature extraction and one-dimensional convolutional neural network (1D-CNN). After singular value decomposition-based feature extraction processing, a temporal robust damage index (TRDI) is extracted, and the effect of EOCs is well removed. Hence, even for the signals with a very large temperature-varying range and low signal-to-noise ratios (SNRs), the final damage detection and localization accuracy retain perfect 100%. Verifications are conducted on two different experimental datasets. The first dataset consists of guided wave signals collected from a thin aluminum plate with artificial noises, and the second is a publicly available experimental dataset of guided wave signals acquired on a composite plate with a temperature ranging from 20°C to 60°C. It is demonstrated that the proposed method can detect and localize the damage accurately and rapidly, showing great potential for application in complex and unknown EOC.

26 March 2020 in Journal of Applied Physics
Verified
Journal Article
Closed access
PDF file
ZLZhi-Tao Luo
Zhi-Tao Luo
JWJian Wang
Jian Wang
FMFei-Long Mao
Fei-Long Mao
LSLang Shen
Lang Shen
SWSheng Wang
Sheng Wang
HZHui Zhang
Hui Zhang
The accurate description of the total diffuse-photon-density-wave field inside turbid media, especially in the near-field region, is extremely critical but challenging for many decades. Here, the total diffuse-photon-density-wave field of semi-infinite turbid media was calculated by the third-order simplified spherical harmonics approximation (SP3) and compared with Monte Carlo simulations. To improve the SP3 approximation, the extrapolated Beer–Lambert law model considering the contribution of the coherent-photon-density-wave in the near-field region was proposed and implemented by Levenberg–Marquardt and universal global optimization methods. Last, we demonstrated the superiority of the proposed model over the existing model in fitting the accuracy and applicable source–detector distance range. The high accuracy and simplicity of the proposed model would be extremely helpful for biomedical applications involving photothermal radiometry, and rapidly determining optical properties of media, along with photoacoustic imaging and photodynamic therapy.

9 June 2022 in Journal of Applied Physics
Verified
Journal Article
Closed access
PDF file
ZLZhi-Tao Luo
Zhi-Tao Luo
PSPeng Shen
Peng Shen
HLHao Luo
Hao Luo
SWSheng Wang
Sheng Wang
XWXin-Kai Wu
Xin-Kai Wu
HZHui Zhang
Hui Zhang
Thermal wave radar (TWR) thermography is a high-efficient nondestructive testing technique to increase the signal-to-noise ratio (SNR) and to enhance target detection capability. However, the detection of subsurface defects, especially small-size defects, usually requires a distinctively high SNR and depth resolvability. This paper proposed an orthogonal phase-coded linear frequency modulated (OPCLFM) excitation waveform, which has significantly improved the SNR and depth resolvability of TWR compared to the LFM waveform. The pulse compression quality of the OPCLFM waveform was initially evaluated through a 1D thermal wave analytical model of carbon fiber reinforced polymer (CFRP) laminate. Results show that the OPCLFM waveform combined with the Kaiser window function compresses the largest sidelobe at least by 18.39 dB compared to the LFM waveform. Furthermore, the superior depth resolvability performance of the OPCLFM waveform was also validated by 3D finite element simulation. Finally, the effect of thermal conductivity on the depth resolvability performance of the OPCLFM waveform was evaluated quantitatively by a delaminated CFRP laminate.

25 August 2025 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Open access
SZShengkang Zong
Shengkang Zong
XLXiao Liang
Xiao Liang
ZHZhenyu He
Zhenyu He
SWSheng Wang
Sheng Wang
KZKai Zheng
Kai Zheng
HZHui Zhang
Hui Zhang
We have developed an innovative framework for analyzing online public sentiment, referred to as the Online Public Opinion Situation Assessment (OPOSA) model. This model employs Deep S-parse Autoencoder (DSAE) and Time-Series Modeling Convolutional Neural Networks (TBSMA-CNN) to effectively segment vast populations of online usersąłsuch as those on social media platforms or e-commerce websitesąłinto distinct groups based on shared preferences and evolving sentiment trends. Central to the model's design is a geometry-based feature selection approach that enables the identification of key aspects of public opinion over time. These aspects, which are influenced by a range of temporal and contextual factors, including weakly supervised learning, are captured and iteratively refined to provide a clearer picture of online opinion dynamics. Unlike traditional methods that rely on probabilistic multi-topic distributions, our model uses DSAE to uncover hierarchical representations of opinion data and leverages TBSMA-CNN to track how public sentiment evolves across time. The synergy of DSAE and TBSMA-CNN enhances the model's capacity to represent users' opinions in a hidden feature space, revealing deeper insights into how individual preferences shift over time. To uncover connections between users with similar opinion trajectories, we introduce a graph-based framework. This framework enables the identification of opinion communitiesąłgroups of users bound by shared opinion patterns or trends, such as shifts in stance or collective movements toward particular issues. To ensure that the model's insights are highly personalized, we incorporate a ranking mechanism that recommends content, discussions, or articles that align with the opinion dynamics of each identified community. This helps ensure that recommendations not only resonate with users' current interests but also stay relevant to their changing preferences. We validated the performance of this approach using a large-scale dataset, which included millions of users and their engagement histories. The results highlighted the model's ability to effectively detect and classify distinct opinion communities, offering large-scale, highly personalized content recommendations that are both timely and contextually relevant to each user's evolving preferences and opinions.