A robust indicator based on singular value decomposition for flaw feature detection from noisy ultrasonic signals

Cui, Ximing; Wang, Zhe; Yi-hua, Kang; Pu, Haiming; Deng, Zhiyang

doi:10.1088/1361-6501/aab029

Cited by 6 publications

(4 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Among these weak signal detection methods, SVD has exhibited a very good performance and is widely used to extract the weak features [10][11][12]. The singular value of the SVD reflects the intrinsic characteristics of the data and has good stability and invariance.…”

Section: Introductionmentioning

confidence: 99%

A new selection method of singular values using correlation coefficients for detecting weak exponential damped sinusoidal signals under strong noise

Zhang

2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Weak signal detection has attracted considerable attention in a wide range of research fields, especially the weak signal detection under strong noise is an urgent problem researches are concerned with. In this paper, a new criterion to select singular values using correlation coefficient is proposed for weak exponential damped sinusoidal signal detection, which has a wide variety of signal processing applications. The innovation of our method lies in selecting K most informative singular values rather than the most energetic singular values. The proposed method measures the similarity between component signals and useful signal through autocorrelation function and correlation coefficient, which can preserve more information of the original signal and be more suitable for weak signal detection scenarios under strong noise. The numerical experiments and analysis are performed to verify the efficiency and effectiveness of our method, which indicates the proposed method is superior to the singular value selection methods based on energy or simple difference principle for correlation coefficients. Compared with stochastic resonance methods suitable for weak signal detection under strong background noise, our proposed method also has significant advantages. Thus, the proposed method is beneficial to theoretical analysis and engineering applications.

show abstract

Section: Introductionmentioning

confidence: 99%

A new selection method of singular values using correlation coefficients for detecting weak exponential damped sinusoidal signals under strong noise

Zhang

2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…To overcome the limitation of conventional denoising method, some other methods have been proposed, such as WTD, singular value decomposition (SVD) denoising, and empirical mode decomposition (EMD) denoising. SVD [4]- [8] is a nonlinear and non-stationary signal processing method and provides efficient denoising. But when detecting key information in a strong noise background, it cannot achieve an ideal denoising effect.…”

Section: Introductionmentioning

confidence: 99%

Recognition of Key Information in Non-Stationary Signals Based on Wavelet Threshold Denoising and Back Propagation Neural Network Optimized by Manta Ray Foraging Optimization Algorithm

Jia

Jing

2022

IEEE Access

View full text Add to dashboard Cite

The identification of key information hidden in non-stationary signals is challenging in various fields such as logistics and transportation, biomedicine, and fault diagnosis. To facilitate this identification, we propose a back propagation neural network (BPNN) classification and recognition algorithm based on wavelet threshold denoising (WTD) and manta ray foraging optimization (MRFO) algorithm for the first time.The algorithm first performs WTD on the original signals to obtain denoised signals. Subsequently, the MRFO algorithm is utilized to optimize the initial weights and thresholds of the BPNN. On the base of this, the optimization model is finally obtained to classify and recognize the key information in the non-stationary signals. The comparative experimental results indicate that the proposed WTD-MRFO-BPNN algorithm can be utilized to availably recognize the key information hidden in non-stationary signals. The recognition accuracy reaches 97.25%.

show abstract

“…For different signals, it is necessary to define a basis function in advance, and then select different relevant parameters, which are interfered by human factors greatly. SVD does not have the problems of modal aliasing and the basis function selected, and has proved to have good stability and invariance [10,11]. SVD is an effective signal denoising method, which has been used widely in bearing fault diagnosis [12].…”

Section: Introductionmentioning

confidence: 99%

Rolling bearing performance degradation assessment based on singular value decomposition-sliding window linear regression and improved deep learning network in noisy environment

Dong¹,

Li²,

Zhu³

et al. 2022

Meas. Sci. Technol.

View full text Add to dashboard Cite

It is difficult to evaluate the degradation performance and the degradation state of the rolling bearing in noisy environment. A new method is proposed to solve the problem based on singular value decomposition (SVD)-sliding window linear regression and ResNeXt - multi-attention mechanism's deep neural network (RMADNN). Firstly, the root mean square(RMS) gradient value is calculated on the basis of RMS based on SVD and linear regression of sliding window, which is used as the rolling bearing performance degradation indicator in noisy environment. Secondly, the degradation state of rolling bearing is divided by the RMS gradient. Thirdly, for the part of the deep learning network model, the soft attention mechanism is introduced into the bidirectional long short-term memory network (BiLSTM) to extract more important and deep fault features. At the same time, the ResNeXt layer is added into the convolutional neural network (CNN) to extract more fault features and merge them through multi-scale grouped convolution. Then, the hybrid domain attention mechanism (HDAM) was introduced after the ResNext layer. The HDAM can screen out more important features from the output features of the ResNext in the two dimensions of channel and spatial. Therefore, the improved deep learning network of the ResNeXt - multi-attention mechanism's deep neural network (RMADNN) in this research is established. Finally, the labeled data set is input into the improved model for training, and the Softmax classifier is used to identify the life decline state of the rolling bearing. The result shows that the indicator of RMS gradient proposed has a better characterization, and the RMADNN model can distinguish the life degradation state of rolling bearing better.

show abstract

A robust indicator based on singular value decomposition for flaw feature detection from noisy ultrasonic signals

Cited by 6 publications

References 17 publications

A new selection method of singular values using correlation coefficients for detecting weak exponential damped sinusoidal signals under strong noise

A new selection method of singular values using correlation coefficients for detecting weak exponential damped sinusoidal signals under strong noise

Recognition of Key Information in Non-Stationary Signals Based on Wavelet Threshold Denoising and Back Propagation Neural Network Optimized by Manta Ray Foraging Optimization Algorithm

Rolling bearing performance degradation assessment based on singular value decomposition-sliding window linear regression and improved deep learning network in noisy environment

Contact Info

Product

Resources

About