Search

The fault diagnosis problems of rolling bearings commonly found in modern industrial equipment are becoming increasingly complex and challenging, and traditional fault diagnosis methods are unable to meet the requirements for high accuracy and strong robustness. To address this problem, this study proposes a high-performance rolling bearing fault diagnosis method based on adaptive feature mode decomposition (AFMD) and Transformer. Aiming at the problem that the decomposition performance of the traditional feature mode decomposition (FMD) is easily affected by its parameter settings, the coati optimization algorithm (COA) is applied to adaptively optimize the two key parameters of the FMD (the number of modes and the filter length). Further, due to the problems of slow convergence and easy to fall into local optimal solutions in the traditional COA, Tent chaotic mapping, dynamic refraction inverse learning and Levy flight strategy are introduced to improve the convergence speed and global search capability of the algorithm. Combining the above two methods and the excellent global feature learning ability of Transformer, the AFMD-Transformer fault diagnosis framework with high accuracy and strong robustness is constructed. The average experimental results on the two fault diagnosis datasets are 99.52% and 99.75%, respectively. The results show that our proposed diagnostic method outperforms currently popular methods in terms of accuracy and robustness. • An intelligent optimization algorithm is applied to solve the parameter setting problem in feature mode decomposition (FMD). • The coati optimization algorithm (COA) is improved using a hybrid strategy. • A new fault diagnosis framework is established by combining AFMD and Transformer, exhibiting high accuracy and robustness. [ABSTRACT FROM AUTHOR]

Research has proven that light treatment, specifically red light radiation, can provide more clinical benefits to human health. Our investigation was firstly conducted to characterize the tissue morphology of mouse breast post 660 nm laser radiation with low power and long-term exposure. RNA sequencing results revealed that light exposure with a higher intervention dosage could cause a number of differentially expressed genes compared with a low intervention dosage. Gene ontology analysis, protein–protein interaction network analysis, and gene set enrichment analysis results suggested that 660 nm light exposure can activate more transcription-related pathways in HC11 breast epithelial cells, and these pathways may involve modulating critical gene expression. To consider the critical role of the Wnt/T-catenin pathway in light-induced modulation, we hypothesized that this pathway might play a major role in response to 660 nm light exposure. To validate our hypothesis, we conducted qRT-PCR, immunofluorescence staining, and Western blot assays, and relative results corroborated that laser radiation could promote expression levels of β-catenin and relative phosphorylation. Significant changes in metabolites and pathway analysis revealed that 660 nm laser could affect nucleotide metabolism by regulating purine metabolism. These findings suggest that the Wnt/β-catenin pathway may be the major sensor for 660 nm laser radiation, and it may be helpful to rescue drawbacks or side effects of 660 nm light exposure through relative interventional agents. [ABSTRACT FROM AUTHOR]

This paper proposes an intelligent diagnosis method for rotating machinery faults based on improved variational mode decomposition (IVMD) and CNN to process the rotating machinery non-stationary signal. Firstly, to solve the problem of time-domain feature extraction for fault diagnosis, this paper proposes an improved variational mode decomposition method with automatic optimization of the number of modes. This method overcomes the problems of the traditional VMD method, in that each parameter is set by experience and is greatly influenced by subjective experience. Secondly, the decomposed signal components are analyzed by correlation, and then high correlated components with the original signal are selected to reconstruct the original signal. The continuous wavelet transform (CWT) is employed to extract the two-dimensional time–frequency domain feature map of the fault signal. Finally, the deep learning method is used to construct a convolutional neural network. After feature extraction, the two-dimensional time-frequency image is applied to the neural network to identify fault features. Experiments verify that the proposed method can adapt to rotating machinery faults in complex environments and has a high recognition rate. [ABSTRACT FROM AUTHOR]

Aiming at reducing the noise interference of digital speckle pattern interferometry (DSPI) phase map, a DSPI phase denoised method based on improved variational mode decomposition (IVMD) is proposed in this paper. Firstly, the IVMD method is adopted to decompose DSPI phase map and obtain the optimal mode components; Then, according to the mode components characteristics, an adaptively mode threshold method is proposed to process the mode components. Finally, the denoised mode components are reconstructed to obtain the denoised DSPI phase map. The result indicates that the proposed method can effectively filter out noise interference, and the peak signal-to-noise ratio (PSNR) is higher than other denoised methods. [ABSTRACT FROM AUTHOR]

A novel method based on variational mode decomposition (VMD) and cross-time–frequency spectrum (CTFS) is proposed for leak location in natural-gas pipelines. Leakage signals are decomposed into mode components by VMD, and an adaptive selection method using mutual information is proposed to process these mode components and obtain the sensitive components closely related to the leak. CTFS is applied to analyze the time–frequency distribution of sensitive mode components. The delay and the corresponding frequency information are extracted when CTFS reaches the maximum. The corresponding frequency is used to calculate the group velocity of wave speed, in combination with the dispersive curve. Finally, the time-delay information and wave speed can be used to determine leakage source. The proposed scheme has been experimentally validated; the results demonstrate that the average relative location errors are reduced to one-third when compared with the CTFS location method based on empirical mode decomposition (EMD). [ABSTRACT FROM AUTHOR]

When a gas pipeline leaks, the leak aperture and its position cannot be easily identified. This paper proposed a leak aperture recognition and location method based on root mean square (RMS) entropy of local mean deposition (LMD) and Wigner–Ville time-frequency analysis. Firstly, wavelet packet energy analysis is employed to remove noise and extract the primary energy leak bands to reconstruct. The reconstructed signal is then decomposed using LMD, and the RMS entropy of PF components is calculated. The RMS entropies of multiple groups are used to build the feature vector, and it is then input into support vector machines (SVMs) to achieve the aperture recognition. An adaptive method based on mutual information is proposed to select the sensitive product function (PF) components, and the time-frequency parameters of the sensitive PF components are then calculated using Wigner–Ville distribution (WVD) method. The time delays are obtained by analyzing the time-frequency parameters. Finally the leak location was achieved by combining time delay with leak signal propagation velocity. The experimental results show that the proposed method can effectively identify different leak apertures, and the leak location accuracy is better than that of the direct cross-correlation method. [ABSTRACT FROM AUTHOR]

In this study, a time-delay estimation method based on Ensemble Local Mean Decomposition (ELMD) method and high-order ambiguity function (HAF) is proposed for locating natural gas pipeline leaks. The leakage signals were decomposed using ELMD, and numerous production functions (PFs) were obtained. An adaptive selection method based on Kullback–Leibler (K–L) divergence was proposed to process these PF components and choose the characteristic PFs that contain most of the leakage information. The HAF was employed to analyze the instantaneous parameters of the characteristic PFs and calculate the difference in arrival time of characteristic frequencies. From the time difference and the signal propagation speed, the natural gas pipeline leakage location can be determined. The experiment results show that the proposed method can locate leaks with higher accuracy than cross-correlation method. [ABSTRACT FROM AUTHOR]

Abstract: In order to design systems generating large photonic band gaps (PBGs), in this paper we construct interesting quasi-one-dimensional (quasi-1D) periodic triangular, diamond, and tetrahedral networks composed of 1D waveguides and triangular fundamental loops. The optical frequency band structures and photonic attenuation behaviors of electromagnetic (EM) waves propagating in the three kinds of one- and two-segment-connected (1SC and 2SC) networks without dissipation are, respectively, investigated and we find that huge PBGs can be produced in the middle of a frequency period and the widths of the largest PBGs can be controlled by adjusting the matching ratio of waveguide length. When the ratio equals 2:1, the width of the hugest PBG resulted in tetrahedral network reaches 0.73 times of a frequency period and is about 1.16 times of the best result reported previously. The average attenuations of the largest PBGs are very strong and increase rapidly with the increment of the number of unit cell. This makes our designed networks with very few unit cells exhibit wonderful PBG features and they can be realized in experiments easily. It may be useful for the designing of optical devices with large PBG and strong attenuation. [Copyright &y& Elsevier]

In the existing tracking-by-detection paradigm, advanced approaches rely on appearance features to establish associations between current detections and trajectories. However, these methods are often plagued by issues such as sluggish tracking performance and suboptimal results, particularly when confronted with the unreliability of the appearance features. Considering these challenges, we propose a novel cascaded matching algorithm called the detection box area-based tracking algorithm (DBAT), which groups the detection boxes by area size and associates detections within each group in a cascaded manner. To enhance the accuracy of grouping, we introduce two crucial components to enhance the quality of detections: the compressed self-decoding module (CSDM) and the task collaboration module (TCM). To acquire more precise location information and augment feature richness, CSDM decomposes the input features into two one-dimensional feature encodings and one two-dimensional feature encoding. Subsequently, these feature encodings perform feature aggregation along both spatial directions to capture long-range dependencies and refine the accuracy of location information. Ultimately, these aggregated features engage with the original features, facilitating information fusion and elevating the overall feature representation. To alleviate potential conflicts between various tasks and bolster task-specific representations, TCM combines disparate receptive fields and decouples features through self-relationship and cross-relationship mappings, thereby concurrently enhancing learning across different tasks. Extensive experiments demonstrate that our proposed method achieves performance comparable to state-of-the-art methods on the MOT17, MOT20 and DanceTrack benchmark tests. • Aggregating features captures long-range correlations and location information. • Reducing task competition improves detection performance. • Cascaded association enhances tracking performance and reduces identity switches. [ABSTRACT FROM AUTHOR]

In order to resolve the problem of speckle interference phase map noise interference, a phase denoising method of digital speckle pattern interferometry based on empirical wavelet transform (EWT) and cross correlation is proposed. First, the EWT is used to decompose the speckle phase map, and a series of components is obtained. Then the noise-free components containing phase information are extracted based on the cross correlation. After that, combined with the particle swarm optimization algorithm and sine–cosine denoising method, an improved sine–cosine denoising method is proposed to process the noise components. Finally, the noiseless components are reconstructed to obtain the denoised speckle phase map. Simulation and experimental results show that the proposed method can effectively reduce the noise interference and obtain accurate phase information. [ABSTRACT FROM AUTHOR]

In this article, the H∞$$ {H}_{\infty } $$ safe control problem of continuous‐time affine nonlinear safety‐critical systems is studied based on the barrier function (BF) and adaptive dynamic programming (ADP). We show that the safety constraints in this article occur under the condition that the system initial state is unsafe and have not been adequately addressed in the existing work. First, the H∞$$ {H}_{\infty } $$ control problem is transformed into a zero‐sum game problem, and a new safe Hamilton–Jacobi–Isaacs equation is proposed by combining with the BF, which makes the unsafe behavior be punished in the learning process. In addition, a damping coefficient is introduced into the BF to punish the unsafe behavior more flexibly. Aiming at the requirement that the initial system state must be strictly constrained in the safe set, a new weight updating method based on ADP is proposed, which can reasonably avoid the influence of the BF on neural network (NN) parameters when the initial system state is unsafe. Furthermore, based on the Lyapunov stability theory, it is proved that the system states of safety‐critical systems and the NN parameters are uniformly ultimately bounded under the safety constraints and disturbances effect. Finally, the effectiveness of the proposed method is verified by two simulation examples. [ABSTRACT FROM AUTHOR]