Search Results (66)

Fault diagnosis methods based on deep learning have made certain progress in recent years. However, in actual industrial scenarios, there are severe strong background noise and limited computing resources, which poses challenges to the practical application of fault diagnosis models. In response to the above issues, this paper proposes a novel noise-resistant and lightweight fault diagnosis framework with nonlinear timestep degenerative greedy strategy (NTDGS) and dual residual horizontal feature pyramid (DRHFPN) for fault diagnosis in strong noise scenarios. This method takes advantage of the strong fitting ability of deep learning methods to model the agent in reinforcement learning by the ways of parameterization, fully leveraging the advantages of both deep learning and reinforcement learning methods. NTDGS is further developed to adaptively adjust the action sampling strategy of the agent at different training stages, improving the convergence speed of the network. To enhance the noise resistance of the network, DRHFPN is constructed, which can filter out interference noise at the feature map level by fusing local feature details and global semantic information. Furthermore, the feature map weighting attention mechanism (FMWAM) is designed to enhance the weak feature extraction ability of the network through adaptive weighting of the feature maps. Finally, the performance of the proposed method is evaluated in different datasets and strong noise environments. Experiments show that in various fault diagnosis scenarios, the proposed method has better noise resistance, higher fault diagnosis accuracy, and fewer parameters compared to other methods. Full article

Fault diagnosis of battery energy storage systems (BESSs) in dynamic operating conditions presents significant challenges due to complex spatiotemporal patterns and measurement noise. This research proposes a novel thermal fault diagnosis framework for BESSs based on Bayesian inference and a Kalman filter. Firstly, PLS-based spatiotemporal feature extraction is designed to capture temporal dependencies. Based on Bayesian global exploration and Kalman real-time weight adaptation, a dual-stage optimization strategy is proposed to derive a multiscale detection index with the dominant statistic, the residual statistic, and the module voltage similarity. A time window-based cumulative contribution strategy is constructed for precise cell localization. Finally, the experimental validation on a Li-ion battery pack demonstrates the proposed method’s superior performance: 96.92–99.90% anomaly detection rate, false alarm rate ranging from 0.10% to 7.22%, detection delays of 1–27 s, and 100% accuracy in fault localization. The proposed framework provides a comprehensive solution for safety management of BESSs and is significant for battery life and energy sustainability. Full article

As the primary discharge channel for fault currents, substation grounding grids are crucial for ensuring the safe and stable operation of power systems. Due to its non-destructive and efficient nature, the pulsed eddy current (PEC) method has become a research hotspot in grounding grid detection in recent years. However, during the detection process, the signal is severely interfered with by substation noise, seriously affecting data quality and interpretation accuracy. To address the problem of suppressing both power frequency and random noise, this paper proposes a composite denoising method that combines bipolar cancellation, minimum noise fraction (MNF), and mask-guided self-supervised denoising. First, based on the periodic characteristics of power frequency noise, a bipolar pulse excitation and differential averaging process is designed to effectively filter out power frequency interference. Subsequently, an MNF algorithm is introduced to identify and reconstruct random noise, improving signal purity. Furthermore, a mask-guided self-supervised denoising model is constructed, using a segmentation convolutional neural network to extract signal-noise masks from noisy data, achieving refined suppression of residual noise. Comparative experiments with simulation and actual substation noise data show that the proposed method outperforms existing typical noise reduction algorithms in terms of signal-to-noise ratio improvement and waveform fidelity, significantly improving the availability and interpretation reliability of pulsed eddy current data. Full article

To address the issue of output anomalies that frequently occur in position sensors of permanent magnet synchronous motors within electromechanical actuation systems operating in harsh environments and can lead to degradation in system performance or operational interruptions, this paper proposes an integrated method for fault diagnosis and fault-tolerant control based on the Cubature Kalman Filter (CKF). This approach effectively combines state reconstruction, fault diagnosis, and fault-tolerant control functions. It employs a CKF observer that utilizes innovation and residual sequences to achieve high-precision reconstruction of rotor position and speed, with convergence assured through Lyapunov stability analysis. Furthermore, a diagnostic mechanism that employs dual-parameter thresholds for position residuals and abnormal duration is introduced, facilitating accurate identification of various fault modes, including signal disconnection, stalling, drift, intermittent disconnection, and their coupled complex faults, while autonomously triggering fault-tolerant strategies. Simulation results indicate that the proposed method maintains excellent accuracy in state reconstruction and fault tolerance under disturbances such as parameter perturbations, sudden load changes, and noise interference, significantly enhancing the system’s operational reliability and robustness in challenging conditions. Full article

In this paper, we propose an ensemble approach for the intelligent fault diagnosis of machinery, which consists of six feature selection methods and classifiers. In the proposed approach, six filters, based on distinct metrics, are utilized. Each filter is combined with an improved sparse representation classifier (ISRC) to form a base model, in which the ISRC is an improved version of a sparse representation classifier and has the advantages of high classification accuracy and being less time consuming than the unimproved version. For each base model, the filter selects a feature subset that is used to train and test the ISRC, where the two hyper-parameters involved in the filter and ISRC are optimized by the binary particle swarm optimization algorithm. The outputs of six base models are aggregated through the cumulative reconstruction residual (CRR), where the CRR is devised to replace the commonly used voting strategy. The effectiveness of the proposed method is verified on six mechanical datasets involving information about bearings and gears. In particular, we conduct a detailed comparison between CRR and voting and carry out an intensive exploration into the question of why CRR is superior to voting in the ensemble model. Full article

The enhancement of weak-fault signal characteristics in rolling bearings under strong background noise interference has always been a challenging problem in rotating machinery fault diagnosis. Research indicates that multivariate statistical indicators such as skewness and kurtosis can characterize the fault features of vibration signals. However, when the fault features in the signal are weak and severely affected by noise, the characterization capability of these indicators diminishes, significantly compromising diagnostic accuracy. To address this issue, this paper proposes a novel multivariate statistical filtering (MSF) method for multi-band filtering, which can effectively screen the target fault information bands in vibration signals during bearing faults. The core idea involves constructing a multivariate matrix of fused-fault multidimensional features by integrating fault and healthy signals, and then utilizing eigenvalue distance metrics to significantly characterize the spectral differences between fault and healthy signals. This enables the selection of frequency bands containing the most informative fault features from the segmented frequency spectrum. To address the inherent in-band residual noise in the MSF-processed signals, this paper further proposes the Hilbert differential Teager energy operator (HDTEO) based on MSF to suppress the filtered in-band noise, thereby enhancing transient fault impulses more effectively. The proposed method has been validated using both public datasets and laboratory datasets. Results demonstrate its effectiveness in accurately identifying fault characteristic frequencies, even under challenging conditions such as incipient bearing faults or severely weak vibration signatures caused by strong background noise. Finally, comparative experiments confirm the superior performance of the proposed approach. Full article

This study investigates the subsurface rooting of the Merija anticline in the Missour Basin, Morocco, with a focus on copper mineralization exploration. A sequential geophysical workflow was implemented, combining gravity surveys, electrical resistivity (ER), and induced polarization (IP) methods. The gravity data, acquired along spaced profiles extending from outcropping areas to Quaternary-covered zones, clearly delineated the structural continuity of the anticline beneath the cover. The application of trend filtering in covered areas allowed the removal of regional effects, successfully isolating residual anomalies associated with the buried continuation of the anticline. Interpolated Bouguer anomaly maps highlighted a major regional fault, interpreted as controlling the deep rooting of the anticline. A resistivity profile was then deployed perpendicular to this fault, providing detailed imaging of the anticline’s geometry and lithological contrasts. Complementary IP profiles conducted near the mine site targeted the detection of chargeability anomalies associated with copper mineralization dominated by malachite, confirming the electrical signature of copper mineralization, particularly within the sandstone and conglomerate formations of the Lower Cretaceous. To validate the geophysical interpretations, a drilling campaign was conducted, which confirmed the presence of the identified lithological units and the anticline rooting, as revealed by geophysical data. This approach provides a robust framework for copper exploration in the Merija area and can be adapted to similar geological contexts elsewhere. Full article

In unmanned systems, especially in large-scale and complex ones, sensor and communication failures occur from time to time and are hard to avoid. Therefore, this paper studies the fault detection problem of a class of unknown nonlinear singular uncertain time-varying delay Markov jump systems (UNSUTVDMJSs). Firstly, the corresponding sliding mode controller (SMC) is designed by using the equivalent control principle, and the unknown nonlinearity is equivalently replaced by changing the system input. Then, a fault detection filter adapted to this system is designed, thereby obtaining the unknown nonlinear stochastic singular uncertain Augmented filter residual system (UNSSUAFRS) model. To obtain the sufficient conditions for the random admissibility of this augmented system, a weak infinitesimal generator was used to design the required Lyapunov-Krasovskii functional. With the help of the Lyapunov principle and $H_{\infty }$ performance analysis method, the sufficient conditions for the random admissibility of UNSSUAFRS under the $H_{\infty }$ performance index $\gamma$ were derived. Finally, with the aid of the designed residual evaluation function and threshold, simulation analysis was conducted on the examples of DC servo motors and numerical calculation examples to verify the effectiveness and practicability of this fault detection filter. Full article

Circulating current control is a critical part of the Modular Multilevel Converter (MMC) control system. Existing control methods rely on arm current information obtained from complex current measurement devices. However, these devices are susceptible to failures, which can lead to distorted arm currents, increased peak arm current values, and higher losses. In extreme cases, this can result in system instability. This paper first analyzes four typical arm current measurement faults, i.e., constant gain faults, amplitude deviation faults, phase shift faults, and stuck faults. Then, a Kalman Filter (KF)-based fault detection method is proposed, which allows for the simultaneous monitoring status of all six arm current measurements. Moreover, to facilitate fault detection, the Moving Root Mean Square (MRMS) value of the observation residual is defined, which effectively detects faults while suppressing noise. The entire fault detection process takes less than 20 ms. Finally, the feasibility and effectiveness of the proposed method are validated through MATLAB/Simulink simulations and experimental results. Full article

Accurate and robust navigation is fundamental to Unmanned Aerial Vehicle (UAV) remote sensing operations. However, the susceptibility of UAV navigation sensors to diverse interference and malicious attacks can severely degrade positioning accuracy and compromise mission integrity. Addressing these vulnerabilities, this paper presents an integrated framework combining sensor anomaly detection with a Dynamic Adaptive Extended Kalman Filter (DAEKF) and federated filtering algorithms to bolster navigation resilience and accuracy for UAV remote sensing. Specifically, mathematical models for prevalent UAV sensor attacks were established. The proposed framework employs adaptive thresholding and residual consistency checks for the real-time identification and isolation of anomalous sensor measurements. Based on these detection outcomes, the DAEKF dynamically adjusts its sensor fusion strategies and noise covariance matrices. To further enhance the fault tolerance, a federated filtering architecture was implemented, utilizing adaptively weighted sub-filters based on assessed trustworthiness to effectively isolate faults. The efficacy of this framework was validated through rigorous experiments that involved real-world flight data and software-defined radio (SDR)-based Global Positioning System (GPS) spoofing, alongside simulated attacks. The results demonstrate exceptional performance, where the average anomaly detection accuracy exceeded 99% and the precision surpassed 98%. Notably, when benchmarked against traditional methods, the proposed system reduced navigation errors by a factor of approximately 2-3 under attack scenarios, which substantially enhanced the operational stability of the UAVs in challenging environments. Full article

In complex underwater environments, the stability of navigation for autonomous underwater vehicles (AUVs) is critical for mission success. To enhance the reliability of the AUV-integrated navigation system, fault detection technology was investigated. Initially, the causes and classifications of faults within the integrated navigation system were analyzed in detail, and these faults were categorized as either abrupt or gradual, based on variations in sensor output characteristics under fault conditions. To overcome the limitations of the residual chi-square method in detecting gradual faults, a cumulative residual detection approach with error coefficient amplification was proposed. The algorithm enhances gradual fault detection by using eigenvalue analysis and constructing fault-frequency-based error amplification coefficients with non-parametric techniques. Furthermore, to improve the detection of gradual faults, artificial intelligence-based fault detection methods were also explored. Specifically, the particle swarm optimization (PSO) algorithm was employed to optimize the hyperparameters of a long short-term memory (LSTM) neural network, leading to the development of a PSO-LSTM fault detection model. In this model, the fault detection function was formulated by comparing the predictions generated by the PSO-LSTM model with those derived from the Kalman filter. The experimental results demonstrated that the fault detection function formulated by PSO-LSTM exhibited enhanced sensitivity to gradual faults and enabled the timely isolation of faulty sensors. In unfamiliar marine regions, the PSO-LSTM method demonstrates greater stability and avoids the need to recalibrate detection thresholds for each sea area—an important advantage for AUV autonomous navigation in complex environments. Full article

Electrostatic detection is a highly accurate way to monitor system performance failures at an early stage. However, due to the weak electrostatic signal, it can be easily interfered with under complex real-world conditions, leading to a reduction in its monitoring capability. During the electrostatic monitoring of rolling bearings, noise can easily drown out the effective signal, making it difficult to extract fault characteristics. In order to solve this problem, a sparse representation based on cluster-contraction stagewise orthogonal matching pursuit (CcStOMP) is proposed to extract the fault features in the electrostatic signals of rolling bearings. The method adds a clustering contraction mechanism to the stagewise orthogonal matching pursuit (StOMP) algorithm, performs secondary filtering based on atom similarity clustering on the selected atoms in the atom search process, updates the support set, and finally solves the weights and updates the residuals, so as to reconstruct the original electrostatic signals and extract the fault feature components of rolling bearings. The method maintains fast convergence while analysing the extraction effect by comparing the measured signals of rolling bearing outer ring and bearing roller faults with the traditional StOMP algorithm, and the results show that the CcStOMP algorithm has obvious advantages in accurately extracting the fault features in the electrostatic monitoring signals of rolling bearings. Full article

The article deals with the properties of fault detection filters when applying their structure to a class of linear, continuous-time systems, with dynamics being specified by the system matrix of the ostensible Metzler structure. The proposed solution is reduced to the use of diagonal stabilization in the synthesis of the state observer and uses the decomposition of the ostensible Metzler matrix. The approach creates a unified framework that covers the compactness of parametric constraints on Metzler matrices and their quadratic stability. Due to the complexity of such constraints, the design conditions are formulated using sharp linear matrix inequalities. For potential application in network control structures, the problem is formulated and solved for linear discrete-time ostensible positive systems. Finally, a linearized model of the B747-100/200 aircraft is used to validate the proposed method. The numerical solution and simulation results show that the proposed approach provides superior sensitivity of the fault detection filter in detecting faults, compared to synthesis methods that do not guarantee the positivity of the filter gain. Full article

A method is proposed for fault classification in milling machines using advanced image processing and machine learning. First, raw data are obtained from real-world industries, representing various fault types (tool, bearing, and gear faults) and normal conditions. These data are converted into two-dimensional continuous wavelet transform (CWT) images for superior time-frequency localization. The images are then augmented to increase dataset diversity using techniques such as rotating, scaling, and flipping. A contrast enhancement filter is applied to highlight key features, thereby improving the model’s learning and fault detection capability. The enhanced images are fed into a modified AlexNet model with three residual blocks to efficiently extract both spatial and temporal features from the CWT images. The modified AlexNet architecture is particularly well-suited to identifying complex patterns associated with different fault types. The deep features are optimized using ant colony optimization to reduce dimensionality while preserving relevant information, ensuring effective feature representation. These optimized features are then classified using a support vector machine, effectively distinguishing between fault types and normal conditions with high accuracy. The proposed method provides significant improvements in fault classification while outperforming state-of-the-art methods. It is thus a promising solution for industrial fault diagnosis and has potential for broader applications in predictive maintenance. Full article

Due to the complexity of the missile air data system (ADS) and the harshness of the environment in which its sensors operate, the effectiveness of traditional fault diagnosis methods is significantly reduced. To this end, this paper proposes a method fusing the model and neural network based on unscented Kalman filter (UKF) and Inception V3 to enhance fault diagnosis performance. Initially, the unscented Kalman filter model is established based on an atmospheric system model to accurately estimate normal states. Subsequently, in order to solve the difficulties such as threshold setting in existing fault diagnosis methods based on residual observers, the UKF model is combined with a neural network, where innovation and residual sequences of the UKF model are extracted as inputs for the neural network model to amplify fault characteristics. Then, multi-scale features are extracted by the Inception V3 network, combined with the efficient channel attention (ECA) mechanism to improve diagnostic results. Finally, the proposed algorithm is validated on a missile simulation platform. The results show that, compared to traditional methods, the proposed method achieves higher accuracy and maintains its lightweight nature simultaneously, which demonstrates its efficiency and potential of fault diagnosis in missile air data systems. Full article