Time-domain Signal Research Articles

Damping-regulated GSR array method for multi-harmonic fault diagnosis under variable speed conditions

Abstract Bearing fault diagnosis under variable speed conditions is essential due to the complexities introduced by speed fluctuations. The accurate detection of multi-harmonic faults is critical for ensuring reliability in intricate operating environments. From the perspective of the beneficial effects of noise, in this study we propose a novel damping-regulated generalized stochastic resonance (GSR) array method designed for multi-harmonic fault diagnosis under variable speed conditions. First, we employ computed order tracking (COT) to transform non-stationary time-domain signals into stationary signals in the angular domain. A damping-regulated GSR oscillator is then introduced within this domain, forming the basis of our GSR array. By analyzing the system stationary response, we reveal the diagnostic performance in theory to assess the array’s capacity for enhancing multi-harmonic fault characteristics. Through simulations and experimental validation, our method demonstrates superior diagnostic accuracy, particularly in variable speed scenarios. It excels in preserving and enhancing weak multi-harmonic fault characteristics while offering significant advantages in high diagnostic robustness. These findings provide significant potential for practical applications in fault diagnostics across various engineering systems.

Impact Damage Localization in Composite Structures Using Data-Driven Machine Learning Methods.

Due to the uncertainty of material properties of plate-like structures, many traditional methods are unable to locate the impact source on their surface in real time. It is important to study the impact source-localization problem for plate structures. In this paper, a data-driven machine learning method is proposed to detect impact sources in plate-like structures and its effectiveness is tested on three plate-like structures with different material properties. In order to collect data on the localization of the impact source, four piezoelectric transducers and an oscilloscope were utilized to construct an experimental platform for impulse response testing. Meanwhile, the position of the impact source on the surface of the test plate is generated by manually releasing the steel ball. The eigenvalue of arrival time in the time domain signal is extracted to build data sets for machine learning. This paper uses the Back Propagation (BP) neural network to learn the difference in the arrival time of each sensor and predict the location of the impact source. The results demonstrate that the machine learning method proposed in this paper can predict the location of the impact source in the plate-like structure without relying on the material properties, with high test accuracy and robustness. The research work in this paper can provide experimental methods and testing techniques for locating impact damage in composite material structures.

Technical validation of the Zeto wireless, dry electrode EEG system.

Objective.Clinical adoption of innovative EEG technology is contingent on the non-inferiority of the new devices relative to conventional ones. We present the four key results from testing the signal quality of Zeto's WR 19 EEG system against a conventional EEG system conducted on patients in a clinical setting.Methods.We performed 30 min simultaneous recordings using the Zeto WR 19 (zEEG) and a conventional clinical EEG system (cEEG) in a cohort of 15 patients. We compared the signal quality between the two EEG systems by computing time domain statistics, waveform correlation, spectral density, signal-to-noise ratio and signal stability.Results.All statistical comparisons resulted in signal quality non-inferior relative to cEEG. (i) Time domain statistics, including the Hjorth parameters, showed equivalence between the two systems, except for a significant reduction of sensitivity to electric noise in zEEG relative to cEEG. (ii) The point-by-point waveform correlation between the two systems was acceptable (r>0.6; P<0.001). (iii) Each of the 15 datasets showed a high spectral correlation (r>0.99; P<0.001) and overlapping spectral density across all electrode positions, indicating no systematic signal distortion. (iv) The mean signal-to-noise ratio (SNR) of the zEEG system exceeded that of the cEEG by 4.82 dB, equivalent to a 16% improvement. (v) The signal stability was maintained through the recordings.Conclusion.In terms of signal quality, the zEEG system is non-inferior to conventional clinical EEG systems with respect to all relevant technical parameters that determine EEG readability and interpretability. Zeto's WR 19 wireless dry electrode system has signal quality in the clinical EEG space at least equivalent to traditional cEEG recordings.

Method and Experimental Research of Transmission Line Tower Grounding Body Condition Assessment Based on Multi-Parameter Time-Domain Pulsed Eddy Current Characteristic Signal Extraction

Pole tower grounding bodies are part of the normal structure of the power system, providing relief from fault currents and equalizing overvoltage channels. They are important devices; however, in the harsh environment of the soil, they are prone to corrosion or even fracture, which in turn affects the normal utilization of the transmission line, so accurately assessing the condition of grounding bodies of the power grid is critical. To assess the operational status of a grounding body in a timely manner, this paper proposes a multi-parameter pulsed eddy current (PEC) time-domain characteristic signal corrosion classification method for the detection of the state of a pole tower grounding body. The method firstly theoretically analysed the influence of multi-parameter changes on the PEC response time-domain feature signal caused by grounding body corrosion and extracts the decay time constant (DTC), and the decay time constant stabilization value (DTCSV) and time to stabilization (TTS) were obtained based on the DTC time domain characteristics for describing the corrosion of the grounding body. Subsequently, DTCSV and TTS were used as inputs to a support vector machine (SVM) to classify the corrosion of the grounding body. A simulation model was constructed to investigate the effect of multiparameter time on the DTCSV and TTS of the tower grounding body based on the single-variable method, and the multiparameter time-domain characterization method used for corrosion assessment was validated. Four defects with different corrosion levels were classified using the optimized SVM model, with an accuracy rate of 95%. Finally, a PEC inspection system platform was built to conduct classification tests on steel bars with different degrees of corrosion, and the results show that the SVM classification model based on DTCSV and TTS has a better discriminatory ability for different corrosive grounders and can be used to classify corrosion in the grounders of poles towers to improve the stability of power transmission.

Time-domain signatures of distinct correlated insulators in a moiré superlattice

Among expanding discoveries of quantum phases in moiré superlattices, correlated insulators stand out as both the most stable and most commonly observed. Despite the central importance of these states in moiré physics, little is known about their underlying nature. Here, we use pump-probe spectroscopy to show distinct time-domain signatures of correlated insulators at fillings of one (ν = −1) and two (ν = −2) holes per moiré unit cell in the angle-aligned WSe2/WS2 system. Following photo-doping, we find that the disordering time of the ν = −1 state is independent of excitation density (nex), as expected from the characteristic phonon response time associated with a polaronic state. In contrast, the disordering time of the ν = −2 state scales with 1/nex, in agreement with plasmonic screening from free holons and doublons. These states display disparate reordering behavior dominated either by first order (ν = −1) or second order (ν = −2) recombination, suggesting the presence of Hubbard excitons and free carrier-like holons/doublons, respectively. Our work delineates the roles of electron–phonon (e–ph) versus electron–electron (e–e) interactions in correlated insulators on the moiré landscape and establishes non-equilibrium responses as mechanistic signatures for distinguishing and discovering quantum phases.

Alzheimer's disease recognition based on waveform and spectral speech signal processing.

Alzheimer's disease (AD) is a neurodegenerative disorder with an irreversible progression. Currently, it is diagnosed using invasive and costly methods, such as cerebrospinal fluid analysis, neuroimaging, and neuropsychological assessments. Recent studies indicate that certain changes in language ability can predict early cognitive decline, highlighting the potential of speech analysis in AD recognition. Based on this premise, this study proposes an AD recognition multi-channel network framework, which is referred to as the ADNet. It integrates both time-domain and frequency-domain features of speech signals, using waveform images and log-Mel spectrograms derived from raw speech as data sources. The framework employs inverted residual blocks to enhance the learning of low-level time-domain features and uses gated multi-information units to effectively combine local and global frequency-domain features. The study tests it on a dataset from the Shanghai cognitive screening (SCS) digital neuropsychological assessment. The results show that the method we proposed outperforms existing speech-based methods, achieving an accuracy of 88.57%, a precision of 88.67%, and a recall of 88.64%. This study demonstrates that the proposed framework can effectively distinguish between the AD and normal controls, and it may be useful for developing early recognition tools for AD.

Study on the Relationship Between Vibration Signals in CFRP Grinding and Grinding Mechanism

AbstractIn the process of grinding carbon fiber reinforced polymer (CFRP), the changing state of acceleration vibration signal is closely related to the grinding surface quality. Therefore, the time domain and frequency domain signal changes in acceleration can be used to monitor and control the grinding process to improve machining quality and efficiency. In this paper, changes in acceleration frequency domain signals under different grinding angles are studied, and the relationship between the vibration forms of acceleration frequency domain signals at the beginning, middle and end of grinding as well as the surface morphology and quality of the corresponding grinding position is studied. Combined with the micro-morphology of the grinding area, it can be seen that, a high frequency and low amplitude vibration can improve the grinding surface quality, but a high amplitude will have a negative impact on the grinding quality. The acceleration vibration during grinding is more stable and the corresponding surface quality is better. With an increase in grinding angle, the maximum amplitude of acceleration first increases and then decreases. The length of the surface fiber also initially decreases and then increases. Lastly, the resin residue first increases and then decreases. The purpose of this study is to provide a theoretical basis for the subsequent grinding state adjustment.

Constraining the Galactic Structure Using Time Domain Gravitational Wave Signal from Double White Dwarfs Detected by Space Gravitational Wave Detectors

The gravitational wave (GW) signals from a large number of double white dwarfs (DWDs) in the Galaxy are expected to be detected by space GW detectors, e.g., the Laser Interferometer Space Antenna (LISA), Taiji, and Tianqin in the millihertz band. In this paper, we present an alternative method by directly using the time-domain GW signal detected by space GW detectors to constrain the anisotropic structure of the Galaxy. The information of anisotropic distribution of DWDs is naturally encoded in the time-domain GW signal because of the variation of the detectors’ directions and consequently the pattern functions due to their annual motion around the Sun. The direct use of the time-domain GW signal enables simple calculations, such as utilizing an analytical method to assess the noise arising from the superposition of random phases of DWDs and using appropriate weights to improve the constraints. We investigate the possible constraints on the scale of the Galactic thin disk and bulge that may be obtained from LISA and Taiji by using this method with mock signals obtained from population synthesis models. We further show the different constraining capabilities of the low-frequency signal (foreground) and the high-frequency signal (resolvable sources) via the Markov Chain Monte Carlo method, and find that the scale height and length of the Galactic thin disk and the scale radius of the bulge can be constrained to a fractional accuracy of ∼30%, 30%, 40% (or 20%, 10%, 40%) by using the low-frequency (or high-frequency) signal detected by LISA or Taiji.

High dynamic-range dimming control for CFMO-OFDM-based VLC systems utilizing a grouped direct current scheme

Since visible light communication (VLC) uses light-emitting diode (LED) as transmitters, it has dual functions of illumination and communication. To meet flexible lighting and energy-saving needs, we propose a novel, to the best of our knowledge, dimming control scheme for spectrally efficient clipping-free multilayer optical orthogonal frequency division multiplexing (CFMO-OFDM)-based VLC systems. In order to achieve high dynamic-range dimming control for LEDs, the time-domain CFMO-OFDM signals are first grouped based on frequency-domain subcarrier distribution. Then, the periodic grouped direct current (DC) signal can be flexibly chosen and added to the above CFMO-OFDM signals. Simulation results demonstrate that, compared with classical asymmetrical hybrid optical OFDM (AHO-OFDM) and DC bias-based dimming methods, the proposed scheme offers higher dimming range, better flexibility, and lower implementation complexity, as well as superior bit error rate (BER) performance.

Wavelet-Based Analysis of Motor Current Signals for Detecting Obstacles in Train Doors

Trains used in urban mass passenger transit often have side entrance doors through which passengers can rapidly enter and exit the train. These doors are typically electrically powered and automated. Many incidents have occurred in which a passenger is trapped and injured while passing through the doors as they are closing. Existing solutions rely on sensitive-edge sensors and current signal peak detection in the time domain to detect door obstructions. However, these methods have notable limitations: sensors are expensive, and sensor failure can result in safety risks, while time-domain signal analysis is prone to noise, potentially leading to false peak detection. The proposed efficient and cost-effective method enhances safety by implementing the torque control of a DC motor which limits the door closing force to prevent potential injuries. In addition, it reduces reliance on traditional edge sensors, which are prone to failure and may result in undetected obstructions. By using a robust time–frequency domain approach, the system ensures more accurate detection, minimizing potential injury risks. An obstruction of the door causes a corresponding change in the motor current. These changes can be detected by using the discrete wavelet transform to decompose the current signal. The norm and peak of the current are used as obstacle detection features, and appropriate threshold values are obtained from a simulation. The simulation results were validated through an experiment. The proposed novel system effectively detects forces between 100 and 200 N (indicating the presence of an object) within 0.3 s and complies with EN14752 safety standard. It can also differentiate between soft and hard objects trapped in train doors.

ResNet50-3Cur-HGCN: a novel multimodal hybrid curvature space approach to bearing fault diagnosis

Abstract Bearings are widely used in various fields due to their advantages of high load-bearing capacity, low friction loss, and minimal vibration and impact. In response to the current issues in the process of fault signal identification using graph neural networks, such as the insufficient expression of node data types, the inadequate exploration of feature information carried by the graph structure, and the singularity of mathematical space operations on graph data, this paper proposes a ResNet50-HGCN model based on the combination of Residual Network-50 (ResNet50) and Heterogeneous Graph Convolutional Neural Network (HGCN). The model is trained in a threefold mixed curvature (3 Curvature, 3Cur) space environment for semi-supervised learning, aiming to achieve accurate classification of bearing fault signals. Specifically, first, the time-domain signals and the components of the time-frequency map obtained from the Wavelet Synchrosqueezing Transform (WSST) are used as the bimodal node information for HGCN. Then, the graph network is merged in the 3Cur space for training and weighted validation, obtaining the predicted category labels of the model under the 3Cur configuration space. Finally, experimental data analysis is conducted, sparse confusion matrices for predicted categories are drawn, and four types of accuracy-related evaluation metrics are calculated. The experimental results show that the proposed ResNet50-3Cur-HGCN classification model outperforms other models in the experiment, achieving an accuracy of 97.71%, which verifies the advantage of the method in terms of accuracy and effectiveness. It also provides a good methodological reference for bearing fault diagnosis approaches centered on graph neural networks.&amp;#xD;

Optical Frequency Sweeping Nonlinearity Measurement Based on a Calibration-free MZI

Frequency sweeping linearity is essential for Frequency-Modulated Continuous Wave (FMCW) Light Detection and Ranging (LIDAR), as it impacts the ranging resolution and accuracy of the system. Pre-distortion methods can correct for frequency sweeping nonlinearity; however, residual minor nonlinearities can still degrade the system ranging resolution, especially at far distances. Therefore, the precise measurement of minor nonlinearities is particularly essential for long-range FMCW LIDAR. This paper proposes a calibration-free MZI for measuring optical frequency sweeping nonlinearity, which involves alternately inserting two short polarization-maintaining fibers with different delays into one arm of an MZI, and after two rounds of beat collection, the optical frequency sweep curve of the light source is accurately measured for nonlinearity evaluation. Using the proposed method, the nonlinearity of a frequency-swept laser source is measured to be 0.2113%, and the relative nonlinearity is 5.3560 × 10−5. With the measured frequency sweep curve, we simulate the beat signal and compare it with the collected beat signal in time and frequency domain, to verify the accuracy of the proposed method. A test conducted at 24.1 °C, 30.4 °C, 39.5 °C and 44.0 °C demonstrate the method’s insensitivity to temperature fluctuations. Based on the proposed MZI, a tunable laser is pre-distorted and then used as light source of a FMCW lidar. A wall at 45 m and a building at 1.2 km are ranged by the lidar respectively. Before and after laser pre-distortion, the FWHM of echo beat spectrum are 25.635 kHz and 9.736 kHz for 45 m, 747.880 kHz and 22.012 kHz for 1.2 km.

A Lightweight Kernel Density Estimation and Adaptive Synthetic Sampling Method for Fault Diagnosis of Rotating Machinery with Imbalanced Data

Rotating machinery is widely used across various industries, making its reliable operation crucial for industrial production. However, in real-world settings, intelligent fault diagnosis faces challenges due to imbalanced fault data and the complexity of neural network models. These challenges are particularly pronounced when defining decision boundaries accurately and managing limited computational resources in real-time machine monitoring. To address these issues, this study presents KDE-ADASYN-based MobileNet with SENet (KAMS), a lightweight convolutional neural network designed for fault diagnosis in rotating machinery. KAMS effectively handles data imbalances commonly found in industrial applications and is optimized for real-time monitoring. The model employs the Kernel Density Estimation Adaptive Synthetic Sampling (KDE-ADASYN) algorithm for oversampling to balance the data, applies fast Fourier transform (FFT) to convert time-domain signals into frequency-domain signals, and utilizes a 1D-MobileNet network enhanced with a Squeeze-and-Excitation (SE) block for feature extraction and fault diagnosis. Experimental results across datasets with varying imbalance ratios demonstrate that KAMS achieves excellent performance, maintaining nearly 90% accuracy even on highly imbalanced datasets. Comparative experiments further demonstrate that KAMS not only delivers exceptional diagnostic performance but also significantly reduces network parameters and computational resource requirements.

Bromine isotope splitting in vibrational spectra of bromoform by time-resolved transient transmission spectroscopy.

The femtosecond pump-probe technique, i.e. the transient transmission spectroscopy, has been used for the first time, to detect the vibrational spectra of symmetric fundamentals ν2 and ν3 in bromoform and chloroform. The spectra were obtained by fast Fourier transforms of the time domain signals. For both, CHCl3 and CHBr3, there are four isotopologues contributing to the spectra, due to the existence of two stable isotopes; chlorine, 35Cl and 37Cl, and bromine, 79Br and 81Br, respectively. While for chloroform the isotope splitting of the ν3 spectral band can be observed even in the spontaneous Raman scattering, for bromoform it is not detectable. Herewith we show that using the time domain spectroscopy and the windowed Fourier transform method we can provide the high resolution spectrum of the ν3 fundamental in bromoform, in which the contributions of all isotopologues are well distinguishable. The data have been collected for few volume concentrations of the studied liquids diluted in the neutral solvent CCl4. It is shown that the intensity pattern of the spectra evolves with decreasing concentration and for the ν3 fundamental it reaches the natural abundance pattern at a very high dilution. The simple theoretical model, which treats the molecules in a liquid as interacting oscillators, allows us to explain the dependence of the shape of the spectrum on the strength of the intermolecular interactions.

High-Precision Digital-to-Time Converter with High Dynamic Range for 28 nm 7-Series Xilinx FPGA and SoC Devices

Over the last ten years, the need for high-resolution time-domain digital signal production has grown exponentially. More than ever, applications call for a digital-to-time converter (DTC) that is extremely accurate and precise. Skew compensation and camera shutter operation represent just a few examples of such applications. The advantages of adopting a flexible and rapid time-to-market strategy focused on fast prototyping using programmable logic devices—such as field-programmable gate arrays (FPGAs) and system-on-chip (SoC)—have become increasingly evident. These benefits outweigh those of performance-focused yet expensive application-specific integrated circuits (ASICs). Despite the availability of various architectures, the high non-recurring engineering (NRE) costs make them unsuitable for low-volume production, especially in research or prototyping environments. To address this trend, we introduce an innovative DTC IP-Core with a resolution, also known as least significant bit (LSB), of 52 ps, compatible with all Xilinx 7-Series FPGAs and SoCs. Measurements have been performed on a low-end Artix-7 XC7A100TFTG256-2, guaranteeing a jitter lower than 50 ps r.m.s. and offering a high dynamic range up to 56 ms. With resource utilization below 1% and a dynamic power dissipation of 285 mW for our target FPGA, the design maintains excellent differential and integral nonlinearity errors (DNL/INL) of 1.19 LSB and 1.56 LSB, respectively.

Multi-channel fused vision transformer network for bearing fault diagnosis under different working conditions

Abstract Many of the current fault diagnosis methods rely on time-domain signals. While the richest information are contained in these signals, their complexity poses challenges to network learning and limits the ability to fully characterize them. To address these issues, a novel multi-channel fused vision transformer network (MFVTN) is proposed in this paper. Firstly, the overlapping patch embedding module is introduced to overlap the time-domain map with edge information, preserving the global continuous features of the time-domain map and adding positional encoding for sorting. This integration helps the vision transformer merge detailed features and construct the global mapping. Secondly, multiple dimensional time domain signal features are extracted and fused in parallel, enabling multi-domain fault diagnosis of bearings. In order to enhance the network ability to extract domain-invariant features, an adversarial training strategy combined with Wasserstein distance is utilized. The results demonstrate that the diagnostic accuracy of the proposed MFVTN can reach 98.2%.

An Automatic Modulation Recognition Algorithm Based on Time–Frequency Features and Deep Learning with Fading Channels

Automatic modulation recognition (AMR) stands as a crucial core technology within the realm of signal processing and perception, playing a significant part in harsh electromagnetic environments. The time–frequency image (TFI) of communication signals can manifest modulation characteristics and serve as a foundation for signal modulation recognition and classification. However, under the influence of the electromagnetic environment, communication signals are exposed to varying degrees of interference, which poses a challenge to the recognition of modulation types. Taking into account the effects of interference and channel fading, this paper introduces a communication signal modulation recognition algorithm based on deep learning (DL) and time–frequency analysis. This approach employs short-time Fourier transform (STFT) to generate time–frequency diagrams from time-domain signals. Subsequently, it binarizes the image and feeds it as input data to the neural network. Our research presents a composite deep convolutional neural network (CNN) architecture known as the composite dense-residual neural network (CDRNN). This architecture focuses on enhancing the feature extraction and identification, aiming to achieve accurate recognition of modulation types in harsh electromagnetic environments. Finally, simulation results validate that the proposed deep learning algorithm holds remarkable advantages in boosting the accuracy of modulation type recognition with better adaptability. The algorithm shows better performance even in harsh electromagnetic environments. When the signal-to-noise ratio (SNR) is 18 dB, the recognition accuracy can reach 92.1%.

Spatiotemporal mode-locked intelligent fiber laser based on cascaded photonic lanterns.

An intelligent controlled spatiotemporal mode-locked (STML) fiber laser based on a photonic lantern (PL) is proposed and experimentally demonstrated. A pair of in-house developed PLs is spliced into the cavity in a back-to-back structure. This PL-based structure functions as a mode multiplexer/demultiplexer to generate higher-order spatial modes. Owing to the strong polarization-dependent loss structural characteristics, the structure also serves as an equivalent saturable absorber based on nonlinear polarization rotation. The STML fiber laser delivers pulsed high-order modes at a center wavelength of 1555.4 nm with a pulse duration of 6.99 ps. With the deployment of a motorized polarization controller and a time-domain signal feedback system, the laser enables intelligent control via a stochastic global polarization search algorithm. Experimental results indicate that the proposed laser can effectively evade disturbances from complex intracavity states and accurately lock onto the fundamental mode-locking state within only 2 min.

Optical and Geometrical Properties from Terahertz Time-Domain Spectroscopy Data.

Terahertz waves are nondestructive and non-ionizing to synthetic and natural materials, including polymeric and biological materials. As a result, terahertz-based spectroscopy has emerged as a suitable technique to uncover fundamental molecular mechanisms and material properties in this electromagnetic spectrum regime. In terahertz time-domain spectroscopy (THz-TDS), the material's optical properties are resolved using the raw time-domain signals collected from the sample and air reference data depending on accurate prior knowledge of the sample geometry. Alternatively, different spectral analysis algorithms can extract the complex index of refraction of optically thick or optically thin samples without specific thickness knowledge. A THz-TDS signal without apparent Fabry-Pérot oscillations is commonly associated with optically thin samples, whereas the terahertz signal of optically thick samples exhibits distinct Fabry-Pérot oscillations. While several extraction algorithms have been reported a priori, the steps from reducing the time-domain signal to calculating the complex index of refraction and resolving the correct thickness can be daunting and intimidating while obscuring important steps. Therefore, the objective is to decipher, demystify, and demonstrate the extraction algorithms for Fabry-Pérot-absent and -present terahertz signals for various polymers with different molecular structure classifications and nonlinear optical crystal zinc telluride. The experimental results were in good agreement with previously published values while elucidating the contributions of the molecular structure to the stability of the algorithms. Finally, the necessary condition for manifesting Fabry-Pérot oscillations was delineated.

A novel TCN-GRU based open set method for unknown damage diagnosis

Abstract In the aerospace and high-speed rail industries, carbon fiber reinforced polymer (CFRP) has seen widespread application. CFRP plates and connectors in operation are often subjected to impacts that can cause damage. The unpredictable nature of the impacts introduces uncertainties in both the location and extent of the damage, posing significant challenges to traditional supervised learning models, which often struggle with missed detections or misclassifications when identifying unknown damages. To address the issue, a deep learning model based on temporal convolutional network-gated recurrent unit (TCN-GRU) is proposed. TCN extracts features from the raw time domain signals, and GRU selectively retains the significant features and completes sequence modeling. A center loss function is incorporated into the fully connected layer to improve the effects of intra-class aggregation and inter-class separation. An unknown detection module is introduced to realize the identification and classification of unknown damages based on a predefined threshold. The experimental results indicate that the proposed method can achieve effective unknown damage diagnosis in the open set case. This study provides a feasible solution for open set unknown damage diagnosis in CFRP plates and connectors.