Yunfei Ma

Dr. Yunfei Ma

orcid

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

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

Representative Publications

View All Publications
20 May 2022 in Sensors
Verified
Journal Article
Open access
PDF file
HBHuajun Bai
Huajun Bai
LWLiang Wen
Liang Wen
YMYunfei Ma
Yunfei Ma
XJXisheng Jia
Xisheng Jia
It is critical to deploy wireless data transmission technologies remotely, in real-time, to monitor the health state of diesel engines dynamically. The usual approach to data compression is to collect data first, then compress it; however, we cannot ensure the correctness and efficiency of the data. Based on sparse Bayesian optimization block learning, this research provides a method for compression reconstruction and fault diagnostics of diesel engine vibration data. This method’s essential contribution is combining compressive sensing technology with fault diagnosis. To achieve a better diagnosis effect, we can effectively improve the wireless transmission efficiency of the vibration signal. First, the dictionary is dynamically updated by learning the dictionary using singular value decomposition to produce the ideal sparse form. Second, a block sparse Bayesian learning boundary optimization approach is utilized to recover structured non-sparse signals rapidly. A detailed assessment index of the data compression effect is created. Finally, the experimental findings reveal that the approach provided in this study outperforms standard compression methods in terms of compression efficiency and accuracy and its ability to produce the desired fault diagnostic effect, proving the usefulness of the proposed method.

26 June 2020 in IEEE Access
Verified
Journal Article
Open access
YMYunfei Ma
Yunfei Ma
XJXisheng Jia
Xisheng Jia
QHQiwei Hu
Qiwei Hu
HBHuajun Bai
Huajun Bai
CGChiming Guo
Chiming Guo
SWShuangchuan Wang
Shuangchuan Wang
Equipment degradation state recognition and prognosis are considered two significant parts of a prognostics and health management (PHM) system that help to reduce downtime and decrease economic losses. In this paper, a sparse representation (SR) feature is proposed as a new degradation feature, and the hidden semi-Markov model (HSMM) is established. The new method offers three significant advantages over the traditional HSMM. (1) Since the degradation information is incomplete, a Gaussian mixture model (GMM) is used here for degradation data clustering and state division. (2) A new degradation feature based on the wavelet packet transform (WPT) and SR can better extract the structural information of the collected signal and reflect the degradation characteristics. (3) To conduct remaining useful life (RUL) predictions, an improved model is proposed, which adds a control variable that can dynamically adjust the state duration. The effectiveness of the proposed method is demonstrated using 8 groups of bearing data from the Center for Intelligent Maintenance Systems (IMS). The results show that the HSMM with the SR feature achieves the best recognition accuracy, of 85.28%. Moreover, the improved prediction model achieves a prediction accuracy of 86.11% on average for 8 bearings.

9 May 2019 in Electronics
Verified
Journal Article
Open access
PDF file
YMYunfei Ma
Yunfei Ma
XJXisheng Jia
Xisheng Jia
QHQiwei Hu
Qiwei Hu
DXDaoming Xu
Daoming Xu
CGChiming Guo
Chiming Guo
SWShuangchuan Wang
Shuangchuan Wang
Vibration signal transmission plays a fundamental role in equipment prognostics and health management. However, long-term condition monitoring requires signal compression before transmission because of the high sampling frequency. In this paper, an efficient Bayesian compressive sensing algorithm is proposed. The contribution is explicitly decomposed into two components: a multitask scenario and a Laplace prior-based hierarchical model. This combination makes full use of the sparse promotion under Laplace priors and the correlation between sparse blocks to improve the efficiency. Moreover, a K-singular value decomposition (K-SVD) dictionary learning method is used to find the best sparse representation of the signal. Simulation results show that the Laplace prior-based reconstruction performs better than typical algorithms. The comparison between a fixed dictionary and learning dictionary also illustrates the advantage of the K-SVD method. Finally, a fault detection case of a reconstructed signal is analyzed. The effectiveness of the proposed method is validated by simulation and experimental tests.

5 February 2020 in Australian Journal of Mechanical Engineering
Verified
Journal Article
Closed access
YMYunfei Ma
Yunfei Ma
XJXisheng Jia
Xisheng Jia
HBHuajun Bai
Huajun Bai
Gearboxes have developed rapidly, and studies on gearbox degradation assessment are of great significance. In the interest of obtaining an effective gearbox degradation feature, a new methodology based on sparse representation and compressive sensing theory is proposed. The wavelet packet transform is an excellent time-frequency analysis method, and the subbands with the largest energy are selected for training the sparse dictionaries. Furthermore, the overcomplete dictionaries of each degradation state can well reflect the signal structure by calculating the compressive sensing reconstruction residuals. Based on the sparse representation feature, a Euclidean distance technique is performed to calculate the assessment rate for each degradation state. The validity and superiority of the proposed assessment method are validated by crack failure data and broken tooth failure data under three kinds of work loads. A comparison of the sparse representation feature with the features of wavelet packet energy, kurtosis, and Renyi entropy is also performed.

14 December 2020 in Measurement Science and Technology
Verified
Journal Article
Closed access
YMYunfei Ma
Yunfei Ma
XJXisheng Jia
Xisheng Jia
The denoising of vibration signals is useful with regards to various applications such as equipment monitoring. This study proposes a double dictionary learning method with a moving ruler strategy. The moving ruler strategy is used to construct a multi-layer and multi-position dictionary to reduce the gap between adjacent columns. However, the drawbacks of this strategy are high dictionary redundancy and reduced computation efficiency. Considering this issue, a double dictionary model is proposed combining the fixed dictionaries with learning dictionaries. The sparse coding and dictionary updating are improved correspondingly. Simulations demonstrate that the proposed method can outperform the existing denoising approaches. Moreover, an application of gearbox is presented to further validate the proposed method.