Identification of liquor adulteration by Raman spectroscopy method based on ICNAFS.

The health of consumers can be impacted by the additives placed into the liquor. To address the issues of poor accuracy, low reliability, and complex operational procedures in identifying adulteration in existing liquor, an improved convex non-negative matrix factorization (ICNAFS) with an adaptive graph constraint for unsupervised feature extraction is proposed in this paper, with the goal of achieving rapid identification of adulteration in liquor by Raman spectroscopy through dimensionality reduction. For the sake to streamline the calculation process for effective feature extraction and increase the accuracy of the analyzed model, the proposed ICNAFS method incorporates two fundamental models, such as ridge regression and convex non-negative matrix factorization (NMF). In particular, dimensionality reduction of the original spectrum is initially conducted using Principal Component Analysis (PCA), Sequential Projection Algorithm (SPA), Convex Non-Negative Matrix Factorization with an Adaptive Graph Constraint (CNAFS), and ICNAFS respectively. k-means is subsequently employed to merge the four models for clustering analysis. The results suggest that the accuracy of the presented ICNAFS-assisted k-means model is higher than the other techniques, with a clustering accuracy of 98.67%, exhibiting a 4% improvement over the existing CNAFS, through examination of 150 sets of tainted liquor data from five categories of samples. This demonstrates the potency of the proposed ICNAFS-assisted k-means clustering model in conjunction with Raman spectroscopy as a method for detecting tainted liquor.

Steel Strip Defect Sample Generation Method Based on Fusible Feature GAN Model under Few Samples.

Due to the shortage of defect samples and the high cost of labelling during the process of hot-rolled strip production in the metallurgical industry, it is difficult to obtain a large quantity of defect data with diversity, which seriously affects the identification accuracy of different types of defects on the steel surface. To address the problem of insufficient defect sample data in the task of strip steel defect identification and classification, this paper proposes the Strip Steel Surface Defect-ConSinGAN (SDE-ConSinGAN) model for strip steel defect identification which is based on a single-image model trained by the generative adversarial network (GAN) and which builds a framework of image-feature cutting and splicing. The model aims to reduce training time by dynamically adjusting the number of iterations for different training stages. The detailed defect features of training samples are highlighted by introducing a new size-adjustment function and increasing the channel attention mechanism. In addition, real image features will be cut and synthesized to obtain new images with multiple defect features for training. The emergence of new images is able to richen generated samples. Eventually, the generated simulated samples can be directly used in deep-learning-based automatic classification of surface defects in cold-rolled thin strips. The experimental results show that, when SDE-ConSinGAN is used to enrich the image dataset, the generated defect images have higher quality and more diversity than the current methods do.

A new technology for rapid determination of isomers of hydroxybenzoic acid by terahertz spectroscopy.

This study investigated the feasibility of using terahertz (THz) technology for the rapid identification of isomers. The time-domain spectra of 2-hydroxybenzoic acid (2-HA), 3-hydroxybenzoic acid (3-HA), and 4-hydroxybenzoic acid (4-HA) were measured by a THz time-domain spectroscopy system (THz-TDS) in the range of 0.3-1.8 THz. Aiming at the isomer classification problem, a THz spectral data classification model based on a variational mode decomposition-particle swarm optimization-support vector machine (VMD-PSO-SVM) method was proposed. Empirical mode decomposition (EMD) and variational mode decomposition (VMD) were used to extract the first eight intrinsic mode functions (IMFs) of the time-domain signal. Principal component analysis (PCA) was used to extract the first 80 principal components of each modal component as the classification feature vector. The particle swarm optimization (PSO) and support vector machine (SVM) algorithms were used to construct 2-, 3-, and 4-HA classification models. We found that the prediction accuracy of the VMD-PSO-SVM model was significantly higher than that of EMD-PSO-SVM model regardless of the modal components. For both EMD and VMD, with the increase in the IMF number, the corresponding classification recognition accuracy tended to decrease. The results showed that the rapid identification model of hydroxybenzoic acid isomers based on THz spectroscopy and SVM was effective and feasible, providing an accurate and rapid method for the chemical synthesis and quality monitoring of biomedicine.

A Novel Fault Diagnosis Scheme for Rolling Bearing Based on Convex Optimization in Synchroextracting Chirplet Transform.

Synchroextracting transform (SET) developed from synchrosqueezing transform (SST) is a novel time-frequency (TF) analysis method. Its concentrated TF spectrum is obtained by applying a synchroextracting operator into TF transformation co-efficients on the TF plane. For this class of post-processing TF analysis methods, the main research focuses on the accurate estimation of instantaneous frequency (IF). However, the performance of TF analysis is greatly affected by the strong frequency modulation (FM) signal. In particular, the actual measured mechanical vibration signals always contain strong background noise, which decreases the resolution of TF representation, resulting in an inaccurate ridge extraction. To solve this problem, an improved penalty function based on the convex optimization scheme is firstly introduced for signal denoising. Based on the superiority of the linear chirplet transform (LCT) in dealing with modulated signals, the synchroextracting chirplet transform (SECT) is employed to sharpen the TF representation after the convex optimization denoising operation. To verify the effectiveness of the proposed method, the numerical simulation signals and the measured fault signals of rolling bearing are carried out, respectively. The results demonstrate that the proposed method leads to a better solution in rolling bearing fault feature extraction.

A Novel Demodulation Analysis Technique for Bearing Fault Diagnosis via Energy Separation and Local Low-Rank Matrix Approximation.

Bearing fault diagnosis is of utmost importance in the maintenance of mechanical equipment. The collected fault vibration signal generally presents a modulated nature due to the special structure and dynamic characteristics of the bearings. This paper introduces a novel demodulation analysis technique via energy separation and local low-rank matrix approximation (LLORMA) to address this type of signal. The amplitude envelope and instantaneous frequency of the signal can be calculated via an energy separation algorithm based on the Teager energy operator. We can confirm the bearing faults by comparing the peak frequencies of the Fourier spectrum of the amplitude envelope and instantaneous frequency with the theoretical bearing fault-related frequencies. However, this algorithm is only suitable for handling single-component signals. In addition, the powerful background noise has a serious effect on the demodulation results. To tackle these problems, a new signal decomposition method based on LLORMA is proposed to decompose the signal into several single-components and eliminate the noise simultaneously. After that, the single-component signal representing the fault characteristics can be identified via the high frequency feature of the modulated signal. The analysis of the simulated signal and the bearing outer race fault signal collected from a bearing-gear fault test rig indicate that the proposed technique has an excellent diagnostic performance for bearing fault signals.

A Novel Fault Feature Recognition Method for Time-Varying Signals and Its Application to Planetary Gearbox Fault Diagnosis under Variable Speed Conditions.

The existing time-frequency analysis (TFA) methods mainly highlight the time-frequency ridges of the interested components by optimizing the time-frequency plane to facilitate the extraction of the relevant components. Generalized demodulation (GD), order tracking (OT), and other methods are generally used in conjunction with the TFA methods to realize the transition from a time-varying signal to a stationary signal, and finally identify the fault feature through a time-frequency plane. Generally, it is necessary to clarify the accuracy of the estimated components such as the rotational frequency or the fault characteristic frequency (FCF) during the operation of the GD or OT methods. Unfortunately, it is not only difficult to extract and locate rotational frequency or FCF, but also complicated in the whole estimation process. In this paper, a simple yet readable method is proposed to reveal the fault feature of time-varying signals. First, the method only needs to extract an arbitrary instantaneous frequency (IF). This is different from the GD method which needs to estimate and locate all phase functions. Then, it converts all variable frequency curves into corresponding lines parallel to the frequency axis based on the extracted IF to determine the proportional relationship between the components. Finally, to further improve the readability of the final results, we reduce the dimension of the transformed time-frequency representation to generate a two-dimensional (2D) energy-frequency map with high resolution and the same proportion. Subsequently, the performance is validated by simulated and experimental data.

Trivariate Empirical Mode Decomposition via Convex Optimization for Rolling Bearing Condition Identification.

As a multichannel signal processing method based on data-driven, multivariate empirical mode decomposition (MEMD) has attracted much attention due to its potential ability in self-adaption and multi-scale decomposition for multivariate data. Commonly, the uniform projection scheme on a hypersphere is used to estimate the local mean. However, the unbalanced data distribution in high-dimensional space often conflicts with the uniform samples and its performance is sensitive to the noise components. Considering the common fact that the vibration signal is generated by three sensors located in different measuring positions in the domain of the structural health monitoring for the key equipment, thus a novel trivariate empirical mode decomposition via convex optimization was proposed for rolling bearing condition identification in this paper. For the trivariate data matrix, the low-rank matrix approximation via convex optimization was firstly conducted to achieve the denoising. It is worthy to note that the non-convex penalty function as a regularization term is introduced to enhance the performance. Moreover, the non-uniform sample scheme was determined by applying singular value decomposition (SVD) to the obtained low-rank trivariate data and then the approach used in conventional MEMD algorithm was employed to estimate the local mean. Numerical examples of synthetic defined by the fault model and real data generated by the fault rolling bearing on the experimental bench are provided to demonstrate the fruitful applications of the proposed method.

Optimized Dynamic Mode Decomposition via Non-Convex Regularization and Multiscale Permutation Entropy.

Dynamic mode decomposition (DMD) is essentially a hybrid algorithm based on mode decomposition and singular value decomposition, and it inevitably inherits the drawbacks of these two algorithms, including the selection strategy of truncated rank order and wanted mode components. A novel denoising and feature extraction algorithm for multi-component coupled noisy mechanical signals is proposed based on the standard DMD algorithm, which provides a new method solving the two intractable problems above. Firstly, a sparse optimization method of non-convex penalty function is adopted to determine the optimal dimensionality reduction space in the process of DMD, obtaining a series of optimal DMD modes. Then, multiscale permutation entropy calculation is performed to calculate the complexity of each DMD mode. Modes corresponding to the noise components are discarded by threshold technology, and we reconstruct the modes whose entropies are smaller than a threshold to recover the signal. By applying the algorithm to rolling bearing simulation signals and comparing with the result of wavelet transform, the effectiveness of the proposed method can be verified. Finally, the proposed method is applied to the experimental rolling bearing signals. Results demonstrated that the proposed approach has a good application prospect in noise reduction and fault feature extraction.

A Joint Fault Diagnosis Scheme Based on Tensor Nuclear Norm Canonical Polyadic Decomposition and Multi-Scale Permutation Entropy for Gears.

Gears are key components in rotation machinery and its fault vibration signals usually show strong nonlinear and non-stationary characteristics. It is not easy for classical time-frequency domain analysis methods to recognize different gear working conditions. Therefore, this paper presents a joint fault diagnosis scheme for gear fault classification via tensor nuclear norm canonical polyadic decomposition (TNNCPD) and multi-scale permutation entropy (MSPE). Firstly, the one-dimensional vibration data of different gear fault conditions is converted into a three-dimensional tensor data, and a new tensor canonical polyadic decomposition method based on nuclear norm and convex optimization called TNNCPD is proposed to extract the low rank component of the data, which represents the feature information of the measured signal. Then, the MSPE of the extracted feature information about different gear faults can be calculated as the feature vector in order to recognize fault conditions. Finally, this researched scheme is validated by practical gear vibration data of different fault conditions. The result demonstrates that the proposed scheme can effectively recognize different gear fault conditions.

Optimized Adaptive Local Iterative Filtering Algorithm Based on Permutation Entropy for Rolling Bearing Fault Diagnosis.

The characteristics of the early fault signal of the rolling bearing are weak and this leads to difficulties in feature extraction. In order to diagnose and identify the fault feature from the bearing vibration signal, an adaptive local iterative filter decomposition method based on permutation entropy is proposed in this paper. As a new time-frequency analysis method, the adaptive local iterative filtering overcomes two main problems of mode decomposition, comparing traditional methods: modal aliasing and the number of components is uncertain. However, there are still some problems in adaptive local iterative filtering, mainly the selection of threshold parameters and the number of components. In this paper, an improved adaptive local iterative filtering algorithm based on particle swarm optimization and permutation entropy is proposed. Firstly, particle swarm optimization is applied to select threshold parameters and the number of components in ALIF. Then, permutation entropy is used to evaluate the mode components we desire. In order to verify the effectiveness of the proposed method, the numerical simulation and experimental data of bearing failure are analyzed.