Search Results (13)

Reliable multi-sensor bearing fault diagnosis is challenged by temporal leakage caused by window-level random splitting, limited modeling of cross-sensor dependencies, and inadequate integration of raw temporal dynamics with time-frequency representations. To address these issues, this study proposes a leakage-resistant multi-sensor diagnosis framework that combines a partition-before-windowing evaluation protocol with adaptive time-frequency graph learning and reliability-aware fusion. Continuous vibration records are first divided into disjoint temporal regions with guard intervals and overlap auditing to suppress time-neighbor leakage. The model then extracts complementary features from a raw-signal branch and a dual-resolution log-STFT branch, while adaptive graph learning captures sample-dependent inter-sensor couplings and sensor reliability weighting highlights informative channels. A cross-gated fusion module further integrates temporal and graph-domain representations in a sample-adaptive manner for final classification. Experiments on a reconstructed nine-class benchmark derived from the HUSTbearing dataset show that the proposed method achieves a Macro-Accuracy of 0.973, a Macro-Recall of 0.964, and a Macro-F1 of 0.954, outperforming representative raw-signal and STFT-based baselines under the same leakage-resistant protocol. These results demonstrate that jointly modeling multi-scale time-frequency structure, dynamic sensor relationships, and reliable evaluation yields an effective and interpretable solution for intelligent bearing fault diagnosis under complex operating conditions. Full article

Classifying power quality disturbances (PQDs) under strong noise conditions remains challenging for existing deep learning models. These models typically separate denoising from feature extraction, often rely on attention mechanisms that operate along only a single dimension, and tend to achieve high accuracy at the cost of high complexity, which limits their performance under low signal-to-noise ratio conditions and hinders practical deployment. To address these limitations, this paper proposes AD-PDAF-Net, which organically integrates three key mechanisms through a co-design strategy. Unlike conventional methods that depend on preprocessing, an adaptive soft thresholding denoising layer is embedded into a lightweight residual network to progressively suppress noise during feature extraction, thereby unifying denoising with feature learning. A parallel dual attention module independently refines features along the channel and temporal dimensions, then adaptively fuses them using learnable weights to capture both frequency domain and temporal characteristics of disturbances. The lightweight network entry replaces aggressive downsampling with small convolutions to preserve transient details, and a bidirectional long short-term memory network (BiLSTM) efficiently captures temporal dependencies. Evaluated on a dataset of 25 disturbance categories defined in IEEE Std 1159-2019, the model achieves a classification accuracy of 97.26% and a Kappa coefficient of 97.02% under 20 dB white Gaussian noise, along with an accuracy of 98.78% under mixed noise conditions. The model has only 0.36 million parameters and a computational cost of just 1.50 GFLOPS. Through this co-design, AD-PDAF-Net achieves both high noise robustness and high classification accuracy with minimal computational overhead, offering an effective solution for time series signal recognition in resource constrained environments. Full article

To address the challenges in identifying effective fault features and achieving sufficient diagnostic accuracy and robustness in variable-speed printing press bearings, where complex mixed-condition vibration signals exhibit non-stationarity, strong nonlinearity, ambiguous time-frequency characteristics, and overlapping fault features across multiple operating conditions, this paper proposes an adaptive optimization signal decomposition method combined with dual-modal time-series and image deep feature fusion for variable-speed multi-condition bearing fault diagnosis. First, to overcome the strong parameter dependency and significant noise interference of traditional adaptive decomposition algorithms, the Crested Porcupine Optimization Algorithm is introduced to adaptively search for the optimal noise amplitude and integration count of ICEEMDAN for effective signal decomposition. IMF components are then screened and reorganized based on correlation coefficients and variance contribution rates to enhance fault-sensitive information. Second, multidimensional time-domain features are extracted in parallel to construct time-frequency images, forming time-sequence-image bimodal inputs that enhance fault representation across different dimensions. Finally, a dual-branch deep learning model is developed: the time-sequence branch employs gated recurrent units to capture feature evolution trends, while the image branch utilizes SE-ResNet18 with embedded channel attention mechanisms to extract deep spatial features. Multimodal feature fusion enables classification recognition. Validation using a bearing self-diagnosis dataset from variable-speed hybrid operation and the publicly available Ottawa variable-speed bearing dataset demonstrates that this method achieves high-accuracy fault identification and strong generalization capabilities across diverse variable-speed hybrid operating conditions. Full article

As the core refrigeration equipment in cooling systems, the operational state of chiller units is crucial for ship support, equipment cooling, and mission stability. However, because of their sensitivity and the complexity of operating environments, obtaining large volumes of complete, fault-labeled data is difficult in practical engineering appli-cations. This limitation makes it challenging for traditional data-driven approaches to deliver accurate fault diagnoses. Furthermore, data collected from different devices or under varying operating conditions often differ significantly in both feature dimensions and distributions, i.e., data heterogeneity, which further complicates model transfer. To address these challenges, this study proposes a deep transfer learning–based fault di-agnosis method designed to leverage abundant knowledge from the source domain while adaptively learning features of the target domain. Given the persistent difficulties in collecting sufficient high-quality labeled fault data, traditional data-driven models continue to face restricted diagnostic performance on target equipment. At the same time, data heterogeneity across devices or operating conditions intensifies the challenge of cross-domain knowledge transfer. To overcome these issues, this study develops a heterogeneous transfer learning method that integrates a dual-channel autoencoder, domain adversarial training, and pseudo-label self-training. This combination enables precise small-sample knowledge transfer from the source to the target domain. Specifi-cally, the dual-channel autoencoder is first applied to align heterogeneous feature di-mensions. Then, a Gradient Reversal Layer (GRL) and a domain discriminator are in-troduced to extract domain-invariant features. In parallel, high-confidence pseu-do-labeled samples from the target domain are incorporated into joint training to im-prove generalization and robustness. Experimental results confirm that the method achieves high fault diagnosis accuracy in typical industrial application scenarios, ena-bling effective identification of common faults in various types of chiller units under conventional operating conditions, the proposed method achieves higher accuracy and F1-scores in multi-class fault diagnosis tasks compared with both traditional approaches and existing transfer learning methods. These findings provide a novel perspective for advancing the intelligent operation and maintenance of chiller units. Full article

Continuous Test-Time Adaptation aims to adapt a source model to continuously and dynamically changing target domains. However, previous studies focus on adapting to each target domain independently, treating them as isolated, while ignoring the interplay of interference and promotion between domains, which limits the model’s sustained capability, often causing it to become trapped in local optima. This study highlights this critical issue and identifies two key factors that limit the model’s sustained capability: (1) The update of parameters lacks constraints, where domain-sensitive parameters capture domain-specific knowledge, leading to unstable channel representations and interference from old domain knowledge and hindering the learning of domain-invariant knowledge. (2) The decision boundary lacks constraints, and distribution shifts, which carry significant domain-specific knowledge, cause features to become dispersed and prone to clustering near the decision boundary. This is particularly problematic during the early stages of domain shifts, where features are more likely to cross the boundary. To tackle the two challenges, we propose a Dual Constraints method: First, we constrain updates to domain-sensitive parameters by minimizing the representation changes in domain-sensitive channels, alleviating the interference among domain-specific knowledge and promoting the learning of domain-invariant knowledge. Second, we introduce a constrained virtual decision boundary, which forces features to move away from the original boundary, and with a virtual margin to prevent features from crossing the decision boundary due to domain-specific knowledge interference caused by domain shifts. Extensive benchmark experiments show our framework outperforms competing methods. Full article

Bolted connections are extensively utilized in aerospace, civil, and mechanical systems for structural assembly. However, inevitable structural vibrations can induce bolt loosening, leading to preload reduction and potential structural failure. Early-stage preload degradation, particularly during initial loosening, is often undetectable by conventional monitoring methods due to limited sensitivity and poor noise resilience. To address these limitations, this study proposes an intelligent bolt preload monitoring framework that combines electromechanical impedance (EMI) signal analysis with a parallel deep learning architecture. A multiphysics-coupled model of flange joint connections is developed to reveal the nonlinear relationships between preload degradation and changes in EMI conductance spectra, specifically resonance peak shifts and amplitude attenuation. Based on this insight, a parallel convolutional neural network (P-CNN) is designed, employing dual branches with 1 × 3 and 1 × 7 convolutional kernels to extract local and global spectral features, respectively. The architecture integrates dilated convolution to expand frequency–domain receptive fields and an enhanced SENet-based channel attention mechanism to adaptively highlight informative frequency bands. Experimental validation on a flange-bolt platform demonstrates that the proposed P-CNN achieves 99.86% classification accuracy, outperforming traditional CNNs by 20.65%. Moreover, the model maintains over 95% accuracy with only 25% of the original training samples, confirming its robustness and data efficiency. The results demonstrate the feasibility and scalability of the proposed approach for real-time, small-sample, and noise-resilient structural health monitoring of bolted connections. Full article

The design of materials with programmable degradation profiles presents a fundamental challenge in pattern recognition across molecular space, requiring the identification of complex structure–property relationships within an exponentially large chemical domain. This paper introduces a novel physics-informed deep learning framework that integrates multi-scale molecular sensing data with reinforcement learning algorithms to enable intelligent characterization and prediction of polymer degradation dynamics. Our method combines three key innovations: (1) a dual-channel sensing architecture that fuses spectroscopic signatures from Graph Isomorphism Networks with temporal degradation patterns captured by transformer-based models, enabling comprehensive molecular state detection across multiple scales; (2) a physics-constrained policy network that ensures sensor measurements adhere to thermodynamic principles while optimizing the exploration of degradation pathways; and (3) a hierarchical signal processing system that balances multiple sensing modalities through adaptive weighting schemes learned from experimental feedback. The framework employs curriculum-based training that progressively increases molecular complexity, enabling robust detection of degradation markers linking polymer architectures to enzymatic breakdown kinetics. Experimental validation through automated synthesis and in situ characterization of 847 novel polymers demonstrates the framework’s sensing capabilities, achieving a 73.2% synthesis success rate and identifying 42 structures with precisely monitored degradation profiles spanning 6 to 24 months. Learned molecular patterns reveal previously undetected correlations between specific spectroscopic signatures and degradation susceptibility, validated through accelerated aging studies with continuous sensor monitoring. Our results establish that physics-informed constraints significantly improve both the validity (94.7%) and diversity (0.82 Tanimoto distance) of generated molecular structures compared with unconstrained baselines. This work advances the convergence of intelligent sensing technologies and materials science, demonstrating how physics-informed machine learning can enhance real-time monitoring capabilities for next-generation sustainable materials. Full article

Driven by dual carbon targets, the scale of wind power integration has surged dramatically. However, its strong volatility causes insufficient short-term prediction accuracy, severely constraining grid security and economic dispatch. To address three key challenges in extracting temporal characteristics of strong volatility, adaptive fusion of multi-source features, and enhancing model interpretability, this paper proposes a Time-Domain Dual-Channel Adaptive Learning Model (TDDCALM). The model employs dual-channel feature decoupling: one Transformer encoder layer captures global dependencies while the raw state layer preserves local temporal features. After TCN-based feature compression, an adaptive weighted early fusion mechanism dynamically optimizes channel weights. The ACON adaptive activation function autonomously learns optimal activation patterns, with fused features visualized through visualization techniques. Validation on two wind farm datasets (A/B) demonstrates that the proposed method reduces RMSE by at least 8.89% compared to the best deep learning baseline, exhibits low sensitivity to time window sizes, and establishes a novel paradigm for forecasting highly volatile renewable energy power. Full article

To address the challenge of low diagnostic accuracy in rolling bearing fault diagnosis under varying operating conditions, this paper proposes a novel method integrating the synchronized wavelet transform (SWT) with an enhanced Vision Transformer architecture, referred to as ResCAA-ViT. The SWT is first applied to process raw vibration signals, generating high-resolution time–frequency maps as input for the network model. By compressing and reordering wavelet transform coefficients in the frequency domain, the SWT enhances time–frequency resolution, enabling the clear capture of instantaneous changes and local features in the signals. Transfer learning further leverages pre-trained ResNet50 parameters to initialize the convolutional and residual layers of the ResCAA-ViT model, facilitating efficient feature extraction. The extracted features are processed by a dual-branch architecture: the left branch employs a residual network module with a CAA attention mechanism, improving sensitivity to critical fault characteristics through strip convolution and adaptive channel weighting. The right branch utilizes a Vision Transformer to capture global features via the self-attention mechanism. The outputs of both branches are fused through addition, and the diagnostic results are obtained using a Softmax classifier. This hybrid architecture combines the strengths of convolutional neural networks and Transformers while leveraging the CAA attention mechanism to enhance feature representation, resulting in robust fault diagnosis. To further enhance generalization, the model combines cross-entropy and mean squared error loss functions. The experimental results show that the proposed method achieves average accuracy rates of 99.96% and 96.51% under constant and varying load conditions, respectively, on the Case Western Reserve University bearing fault dataset, outperforming other methods. Additionally, it achieves an average diagnostic accuracy of 99.25% on a real-world dataset of generator non-drive end bearings in wind turbines, surpassing competing approaches. These findings highlight the effectiveness of the SWT and ResCAA-ViT-based approach in addressing complex variations in operating conditions, demonstrating its significant practical applicability. Full article

Sliding bearings are vital components in modern industry, exerting a crucial influence on equipment performance, with wear being one of their primary failure modes. In addressing the issue of wear diagnosis in sliding bearings, this paper proposes an intelligent diagnostic method based on a multiscale gated convolutional neural network (MGCNN). The proposed method allows for the quantitative inference of the maximum wear depth (MWD) of sliding bearings based on online vibration signals. The constructed model adopts a dual-path parallel structure in both the time and frequency domains to process bearing vibration signals, ensuring the integrity of information transmission through residual network connections. In particular, a multiscale gated convolution (MGC) module is constructed, which utilizes convolutional network layers to extract features from sample sequences. This module incorporates multiple scale channels, including long-term, medium-term, and short-term cycles, to fully extract information from vibration signals. Furthermore, gated units are employed to adaptively assign weights to feature vectors, enabling control of information flow direction. Experimental results demonstrate that the proposed method outperforms the traditional CNN model and shallow machine learning model, offering promising support for equipment condition monitoring and predictive maintenance. Full article

Due to the variation in the image capturing process, the difference between source and target sets causes a challenge in unsupervised domain adaptation (UDA) on person re-identification (re-ID). Given a labeled source training set and an unlabeled target training set, this paper focuses on improving the generalization ability of the re-ID model on the target testing set. The proposed method enforces two properties at the same time: (1) camera invariance is achieved through the positive learning formed by unlabeled target images and their camera style transfer counterparts; and (2) the robustness of the backbone network feature extraction is improved, and the accuracy of feature extraction is enhanced by adding a position-channel dual attention mechanism. The proposed network model uses a classic dual-stream network. Comparative experimental results on three public benchmarks prove the superiority of the proposed method. Full article

The task of pitch estimation is an essential step in many audio signal processing applications. In this paper, we propose a data-driven pitch estimation network, the Dual Attention Network (DA-Net), which processes directly on the time-domain samples of monophonic music. DA-Net includes six Dual Attention Modules (DA-Modules), and each of them includes two kinds of attention: element-wise and channel-wise attention. DA-Net is to perform element attention and channel attention operations on convolution features, which reflects the idea of "symmetry". DA-Modules can model the semantic interdependencies between element-wise and channel-wise features. In the DA-Module, the element-wise attention mechanism is realized by a Convolutional Gated Linear Unit (ConvGLU), and the channel-wise attention mechanism is realized by a Squeeze-and-Excitation (SE) block. We explored three kinds of combination modes (serial mode, parallel mode, and tightly coupled mode) of the element-wise attention and channel-wise attention. Element-wise attention selectively emphasizes useful features by re-weighting the features at all positions. Channel-wise attention can learn to use global information to selectively emphasize the informative feature maps and suppress the less useful ones. Therefore, DA-Net adaptively integrates the local features with their global dependencies. The outputs of DA-Net are fed into a fully connected layer to generate a 360-dimensional vector corresponding to 360 pitches. We trained the proposed network on the iKala and MDB-stem-synth datasets, respectively. According to the experimental results, our proposed dual attention network with tightly coupled mode achieved the best performance. Full article

The demand for bandwidth-intensive and delay-sensitive services is surging daily with the development of 5G technology, resulting in fierce competition for scarce radio resources. Power domain Nonorthogonal Multiple Access (NOMA) technologies can dramatically improve system capacity and spectrum efficiency. Unlike existing NOMA scheduling that mainly focuses on fairness, this paper proposes a power control solution for uplink hybrid OMA and PD-NOMA in dual dynamic environments: dynamic and imperfect channel information together with the random user-specific hierarchical quality of service (QoS). This paper models the power control problem as a nonconvex stochastic, which aims to maximize system energy efficiency while guaranteeing hierarchical user QoS requirements. Then, the problem is formulated as a partially observable Markov decision process (POMDP). Owing to the difficulty of modeling time-varying scenes, the urgency of fast convergency, the adaptability in a dynamic environment, and the continuity of the variables, a Deep Reinforcement Learning (DRL)-based method is proposed. This paper also transforms the hierarchical QoS constraint under the NOMA serial interference cancellation (SIC) scene to fit DRL. The simulation results verify the effectiveness and robustness of the proposed algorithm under a dual uncertain environment. As compared with the baseline Particle Swarm Optimization algorithm (PSO), the proposed DRL-based method has demonstrated satisfying performance. Full article