For fault diagnosis under unseen working conditions (WCs), it is crucial to extract general knowledge unrelated to data distribution from available source data and identify transferable discriminative features. However, WC-related information is often tightly coupled with health state (HS)-related information, making it difficult to directly distinguish their contributions, posing challenges to fault diagnosis. To address this issue, a novel approach named WC-decoupled and invariant-feature fusion transformer (WCD-IFFT) is proposed, which aims to minimize the impact of WCs by extracting transferable features closely related to HSs. Specifically, two key components are designed to decouple WC-related features from HS-related features: orthogonality separation and decouple loss. Additionally, to enrich the semantics of HS-related features, time-domain and Fourier phase features are mapped into a unified space and fused, combining the instantaneous changes of time-domain signals with frequency-domain distribution information to enhance the feature representation capability. Extensive experiments on cross-domain fault diagnosis tasks demonstrate the effectiveness of the proposed method.

Tackling the challenge of open-switch (OS) fault diagnostics in grid-tied three-level neutral point clamped (NPC) inverters with parameter uncertainty, this paper introduces a fault diagnosis method that integrates a variable forgetting factor bias-compensation recursive least squares (VFFBCRLS) algorithm with a novel discrete disturbance sliding mode observer (DSMO) for three-level inverters. The proposed approach initially employs a VFFBCRlS algorithm to obtain the uncertain parameters of the inverter. Building upon this foundation, a novel discrete DSMO is introduced to obtain the output currents rapidly and accurately. Then, an adaptive fault detection variable is constructed based on the norm of the residual between the measured and the estimated currents, ensuring the accuracy and robustness of the detection algorithm. Finally, a precise identification of OS faults in grid-tied inverters is achieved through the establishment of a localization mechanism. The hardware-in-the-loop (HIL) test results provide validation for the efficacy and robustness of the proposed method.

This article introduces an innovative approach to simultaneously estimate time-domain and spatiotemporal faults in nonlinear distributed parameter systems (NDPSs)nonlinear distributed parameter systems (NDPSs) under external disturbances. First, the establishment of an iterative learning observer that accounts for both temporal and spatial changes is presented. Next, a fault estimation law is devised utilizing a distinct full iterative learning (FIL)full iterative learning (FIL) technique, facilitating rapid and precise estimation of fault signals while mitigating the impact of external disturbances. Furthermore, the adoption of the $\lambda $ -norm method aids in simplifying the determination of convergence conditions and gain matrix calculations. Lastly, comprehensive simulation results validate the efficacy of the developed approach, underscoring its adeptness in efficiently and precisely estimating faults across both time and spatiotemporal domains.

In this article, we propose a feature distillation-based uniformity few-shot domain adaptation (FUFD), for cross-domain fault diagnosis with sample shortage. To address the the few-shot problem, a uniformity prototypical contrastive network is designed to improve the data sensitivity of the model. Compared to the vanilla prototypical network, the learned prototypes contain more information about fault classes by encoding semantic structure information into the feature space while dynamically estimating the distribution concentration around each class prototype. Uniformity and correlation principles are introduced to alleviate prototype collapse: the uniformity principle ensures balanced prototype distribution, while the correlation principle enhances the diversity and distinctiveness of prototypical features. In addition, a cross-domain feature distillation-based domain adaptation module is designed to address significant domain shift. This module softens the class-specific information to capture more domain-consistent information and avoid overfitting to source working condition. Finally, experiments and ablation studies on cross-domain bearing fault diagnosis tasks with limited samples validate the effectiveness of FUFD and its individual modules in enhancing few-shot cross-domain fault diagnosis performance.

A certainty and transferability guided few-shot domain adaptation network is proposed to address few-shot open-set cross-domain fault diagnosis in this article. The proposed method is composed of a feature extractor, a certainty-guided prototypical contrastive module and a transferability weighting domain adaptation module. The certainty-guided prototypical contrastive module based on samples informative importance is designed to enhance the data sensitivity with limited samples while achieving well class separation for open-set scenarios. The module infers informative importance of samples to guide method learn more effective representations. Meanwhile, correlation and uniformity principles are incorporated to alleviate prototype collapse. The transferability weighting domain adaptation module is designed to address great domain gaps and negative transfer caused by asymmetrical label spaces. The module quantifies sample transferability and down-weights the irrelevant samples based on their transferability scores. Experimental results on few-shot open-set cross-domain bearing fault diagnosis tasks demonstrated the superior and effectiveness of the proposed method.

The remaining useful life (RUL) prediction is one of the most important tasks in the prognostics and health management of industrial equipment. The statistical model-based method is widely used for RUL prediction, but it depends on sufficient prior knowledge and appropriate degradation assumptions, which limits its use in the case of the complexity degradation process or insufficient prior knowledge. Motivated by this, we propose a hybrid model to describe the degradation process, which is composed of some given degradation models. This can be conducive to describing complex degradation trajectories and further improving the flexibility of the degradation model. Then, a sparsity mechanism is proposed to automatically fuse these candidate degradation models based on the sparse variational Bayesian method. As a result, we can find the most appropriate degradation mechanism for a given degradation process without sufficient prior knowledge. To the best of the authors' knowledge, it is the first time to use such a fusion mechanism to describe the complex degradation process. A numerical example, a practical example, and three public datasets are used to verify the effectiveness and merits of the proposed method.

In this article, a distributed fault estimation (DFE) approach for switched interconnected nonlinear systems (SINSs) with time delays and external disturbances is proposed using a novel segmented iterative learning scheme (SILS). First, through the utilization of interrelated information among subsystems, a distributed iterative learning observer is developed to enhance the accuracy of fault estimation results, which can realize the fault estimation of all subsystems under time delays and external disturbances. Simultaneously, to facilitate rapid fault information tracking and significantly reduce sensitivity to interference, a new SILS-based fault estimation law is constructed by combining the idea of segmented design with the method of variable gain. Then, an assessment of the convergence of the established fault estimation methodology is conducted, and the configurations of observer gain matrices and iterative learning gain matrices are duly accomplished. Finally, simulation results are showcased to demonstrate the superiority and feasibility of the developed fault estimation approach.

Accurate fault detection and localization are essential for single-phase three-level (SPTL) rectifier systems with high reliability requirements. However, power switch faults, grid-side current sensor (CS) faults, and DC-link voltage sensor (VS) faults can all contribute to distorted output in the rectifier system, posing challenges for existing diagnostic methods tailored for single-type faults, as they struggle to distinguish between these various faults. Therefore, this study proposes a comprehensive diagnosis technology for open-circuit (OC) faults, CS faults, and VS faults of SPTL rectifiers on the basis of a reduced-order observer. To achieve this, the method begins by expanding and transforming the state equation of the rectifier with faults, ensuring complete decoupling of the OC fault vector from the initial system states and sensor faults. Subsequently, an assessment of the initial system state, CS faults, and VS faults is achieved via the design of a reduced-order observer. Using these estimation results, fault detection variable and its adaptive thresholds is designed, along with fault-distinguishing variables to differentiate between sensor faults and OC faults. Simultaneously, sensor fault identification method and OC fault location method are introduced. Finally, the validity and resilience of the comprehensive diagnostic approach are confirmed through hardware-in-the-loop (HIL) test results under diverse scenarios.

Unsupervised domain adaptation is widely used for fault diagnosis under variable working conditions. However, loss oscillation and slow convergence, which are caused by the dynamically varying alignment of targets during domain adaptation, are ignored. Therefore, a Gaussian mixture variational based transformer domain adaptation (GMVTDA) fault diagnosis method is proposed. A feature extractor based on transformer layers is designed to capture long-term dependency information and local features. Subsequently, a domain alignment term is proposed to project the features learned from both working conditions into the common assistance distribution and make them follow the same distribution after the alignment process. Additionally, considering that fault diagnosis is a multiclassification process, a Gaussian mixture is utilized to build the common assistance distribution. Ultimately, the proposed GMVTDA is applied to bearing fault diagnosis under variable working conditions, and the experimental results prove its effectiveness.

Deep state space model (DSSM) are a temporal model that is actively researched. It combines deep neural networks with classic state space model (SSM) so that it can be used for prediction and identification of dynamic system. However, DSSM treats observation as multiple 1-D variables instead of one multidimensional variable, which does not make the most of correlation information in observation. On the other hand, DSSM cannot combine the mechanism of the system. Therefore, we propose a new model called mechanism-assisted deep state space model (MA-DSSM). We use the multidimensional time-varying SSM to describe the temporal structure of system observation. System observation is assumed to be a multidimensional random variable. Two recurrent neural networks are used to extract static and dynamic features of the system. These measures enhance the prediction performance of the model. In addition, we design conditional mask which can finely and flexibly combine prior knowledge in the SSM with DSSM. So that MA-DSSM can combine data with mechanism and further improve performance. This article introduces the detailed structure of MA-DSSM, the calculation of training loss and the process of prediction. Three numerical experiments are used to verify the performance of MA-DSSM. The experimental results show that MA-DSSM has the best prediction performance with the assistance of system mechanism. The prediction performance is also better than the comparison model without the assistance of system mechanism.

Most existing domain generalization fault diagnosis methods concentrate on learning domain-invariant features or global feature distribution alignment. Nevertheless, this could lose vital clues related to the fault categories, and make adapting to unknown working conditions challenging. To this end, a novel approach termed feature adaptive modulation and health state prototype consistency learning (FAMPL) is proposed. Specifically, FAMPL incorporates a feature adaptive modulation module designed to generate modulation parameters, which are utilized to perform affine transformations on the acquired features, yielding modulation features. This approach aims to capture essential clues associated with specific working conditions. To further enhance the ability to distinguish between different fault categories, a specialized health state prototype learning strategy has been developed. This approach significantly refines the model's capacity for feature discrimination, making it more adept at accurately identifying and categorizing various fault types. Numerous cross-domain fault diagnosis experiments have demonstrated the superiority of FAMPL.

The identification of gene regulatory networks (GRN) from gene expression time series data is a challenge and open problem in system biology. This paper considers the structure inference of GRN from the incomplete and noisy gene expression data, which is a not well-studied issue for GRN inference. In this paper, the dynamical behavior of the gene expression process is described by a stochastic nonlinear state-space model with unknown noise information. A variational Bayesian (VB) framework are proposed to estimate the parameters and gene expression levels simultaneously. One of the advantages of this method is that it can easily handle the missing observations by generating the prediction values. Considering the sparsity of GRN, the smoothed gene data are modeled by the extreme gradient boosting tree, and the regulatory interactions among genes are identified by the importance scores based on the tree model. The proposed method is tested on the artificial DREAM4 datasets and one real gene expression dataset of yeast. The comparative results show that the proposed method can effectively recover the regulatory interactions of GRN in the presence of missing observations and outperforms the existing methods for GRN identification.

Time Series Classification (TSC) is an essential task in Time Series Data (TSD) analysis. Ensemble-based approaches now achieve the best performance on TSC tasks. However, integrating numerous different models makes them highly suffer from heavy preprocessing. Even worse, non-deep-learning ensemble-based methods suffer from substantial computational costs due to lacking GPU acceleration. Multi-scale information in TSD can improve TSC performance. However, Existing TSD classifiers employing multi-scale information struggle with heavy preprocessing and cannot help other TSD classifiers obtain multi-scale feature extraction capabilities. Inspired by these, we proposed a performance enhancement framework called multi-scale ensemble booster (MEB), helping existing TSD classifiers achieve performance leaps. In MEB, we proposed an easy-to-combine network structure without changing any of their structure and hyperparameters, only needed to set one hyperparameter, consisting of multi-scale transformation and multi-output decision fusion. Then, a probability distribution co-evolution strategy is proposed to attain the optimal label probability distribution. We conducted numerous ablation experiments of MEB on 128 univariate datasets and 29 multivariate datasets and comparative experiments with 11 state-of-the-art methods, which demonstrated the significant performance improvement ability of MEB and the most advanced performance of the model enhanced by MEB, respectively. Furthermore, to figure out why MEB can improve model performance, we provided a chain of interpretability analyses.https://github.com/foryichuanqi/Multi-Scale-Ensemble-Booster-for-Improving-Existing-Time-Series-Data-Classifiers.

Although existing domain-adaptive person re-identification (re-ID) methods have achieved competitive performance, most of them highly rely on the reliability of pseudo-label prediction, which seriously limits their applicability as noisy labels cannot be avoided. This paper designs a Transformer framework based on body part-level domain alignment to solve the above-mentioned issues in domain-adaptive person re-ID. Different parts of the human body (such as head, torso, and legs) have different structures and shapes. Therefore, they usually exhibit different characteristics. The proposed method makes full use of the dissimilarity between different human body parts. Specifically, the local features from the same body part are aggregated by the Transformer to obtain the corresponding class token, which is used as the global representation of this body part. Additionally, a Transformer layer-embedded adversarial learning strategy is designed. This strategy can simultaneously achieve domain alignment and classification of the class token for each human body part in both target and source domains by an integrated discriminator, thereby realizing domain alignment at human body part level. Compared with existing domain-level and identity-level alignment methods, the proposed method has a stronger fine-grained domain alignment capability. Therefore, the information loss or distortion that may occur in the feature alignment process can be effectively alleviated. The proposed method does not need to predict pseudo labels of any target sample, so the negative impact caused by unreliable pseudo labels on re-ID performance can be effectively avoided. Compared with state-of-the-art methods, the proposed method achieves better performance on the datasets that are in line with real-world scene settings. The source codes of this paper will be available at https://github.com/lhf12278/BPDA.