Remaining useful life (RUL) prediction plays a pivotal role in prognostics and health management (PHM) systems, which enhances the reliability of operating equipment and reduces maintenance costs. With the advent of Industrial Internet of Things (IIoT) technology, it becomes feasible to obtain preformance-degradation data that precisely mirrors the health status of equipment. This facilitates real-time process monitoring of device status and promotes intelligent predictive maintenance methods, thereby achieving more accurate RUL estimations. Nonetheless, IIoT systems often suffer from sensing or communication failures in practical industrial scenarios, leading to fragmented multivariate time-series data with missing observations from partial edge devices, which severely restricts the performance of RUL prediction. To address this issue, a graph-based representation learning method is skillfully proposed for RUL prediction with missing values. Specifically, the observations and features in multivariate time-series are regarded as two distinct nodes types in a semi-bipartite graph. Furthermore, the observed values are viewed as real edges whereas the dependency between timestamps and the correlation between features as virtual edges. In this scheme, the representation learning for multivariate time-series is expressed as the graph-level tasks. The acquired representations in the observation missing mode possess the capability to extract meaningful information and unveil underlying patterns from fragmented data, which bolsters the performance of RUL estimations through direct imputation or end-to-end architecture. The effectiveness and robustness of the proposed method are validated through the results of comparative experiments under different missing patterns using the C-MAPSS dataset, demonstrating its capability to enhance predictive maintenance in IIoT-enabled systems while reducing unexpected failures and maintenance costs.

In modern industrial processes, class imbalance occurs when there is a significant disparity in the number of instances between different classes. Current approaches for handling this problem cannot work effectively due to the invalid instance replenishment strategy for rare categories and even exacerbate class imbalance issues. Therefore, this work presents a novel semi-supervised fault diagnosis (FD) method to address imbalances in FD data by leveraging extensive unlabeled samples. Inspired by adversarial discriminative domain adaptation learning, the proposed approach includes a distribution alignment model for extracting domain-invariant fault features from unlabeled data. Additionally, a soft threshold selection strategy is introduced to strategically select unlabeled fault samples, ensuring an abundance of samples for rare categories and enriching their distribution. Extensive experiments on the two industrial process datasets, including a real-world hot rolling of steel process and a well-established public Tennessee Eastman process, demonstrate the effectiveness of the proposed method in alleviating imbalances and utilizing unlabeled samples, establishing its superiority over existing methods. The code is publicly available onhttps://github.com/Ticuby/SFDM.

Detection and classification are two significant tasks for industrial fault diagnosis. However, conventional methods typically treat these tasks as separate and independent problems, and necessitate a retraining process when new fault samples or classes are collected. Therefore, an incremental support vector data description scheme using Gaussian kernel function is pro-posed for industrial process fault diagnosis in a unified frame-work. In this framework, the decision boundary is updated incrementally only based on the specific original support vectors and newly collected samples. An adaptive threshold and a restructured radius are proposed to promote accuracy in the fault detection. In the classification procedure, the hyperspheres for all known classes are constructed by decision tree. The new sample that does not belong to any known class is identified as an unknown class. Without a time-consuming retraining process, the proposed diagnosis method with the incremental learning capability can synchronously achieve the fault detection and classification task. Experimental results demonstrate the effectiveness and superiority in terms of diagnosis performance.Note to Practitioners—In practical industrial processes, new fault samples are collected and new fault classes emerge continually. Under such scenarios, fault diagnosis with incremental learning capability is becoming increasingly important. This work proposes a unified incremental framework for fault diagnosis based on support vector data description, which is able to achieve fault detection and classification synchronously. For fault detection, an adaptive threshold and a restructured radius are developed. For fault classification, hypersphere-shaped boundaries of all fault classes are given via the decision tree-based strategy. The proposed method is updated to include new fault samples and fault classes without dimensionality reduction or any distributional assumptions.

Object detection in remote sensing images is crucial in numerous fields; however, it becomes highly challenging under adverse weather circumstances. Given that previous remote sensing image object detection methods were designed based on normal weather conditions and ideal datasets, they are not beneficial for detection under real-world haze conditions and with class-imbalanced data. In this work, an adaptive dehazing centroid contrastive network (DHC-Net) is proposed to address the aforementioned issues. This network consists of an adaptive dehazing module and a centroid-guided contrastive learning approach. The adaptive dehazing module learns the image content to generate adaptive dehazing parameters, thus alleviating the influence of haze on the quality of remote sensing images. The centroid-guided contrastive learning approach is particularly designed to address the issue of imbalanced datasets. Integrating centroid vectors with actual samples in each training batch guarantees that each class is sampled at least once, effectively preventing the undersampling of minority classes. Moreover, dynamic weighted sampling based on prediction confidence guides the model to give priority to smaller classes, remarkably improving its ability to handle imbalanced data. Extensive experiments on the DOTA-v2.0, DOTA-v2.0Haze, RTTS, and HazeNet datasets demonstrate that DHC-Net is outstanding in handling haze conditions in remote sensing data, substantially enhancing target detection accuracy, even in the presence of imbalanced object classes. The source code will be available athttps://github.com/Linghuaqian1/DHC_Net

Learning under feature evolution data streams has attracted widespread attention in recent years. Existing methods usually assume that the model predicts and learns from all instances in the data stream. However, when the data stream rate is faster than the model update rate, the model can only learn from some instances. Therefore, this assumption may not always hold in practical scenarios. Additionally, existing methods often update based only on the current instance, ignoring the impact of data stream changes, which further limits their application in practical data streams. This paper proposes a novel learning paradigm to solve this problem: Online Learning under Feature Evolution data streams with A Fast Rate, called OLFE-FR. Specifically, OLFE-FR introduces the concept of relative rate to adaptively determine the prediction mode and update node of the model in the data stream. Additionally, OLFE-FR proposes an adaptive learning rate adjustment strategy based on the upper bound of dynamic regret minimization. This strategy enables the model to find a suitable learning rate based on weights change induced by known data stream variations before using the instance update. Theoretical analysis and experimental results show that OLFE-FR can effectively handle feature evolution data streams with a fast rate.

Evolving monitoring requirements may necessitate the addition of new sensors or the exclusion of old ones. Unfortunately, traditional data-driven fault diagnosis methods usually hold the assumption that the number of sensors remains constant throughout the monitoring process, so they need to be retrained with intractable computation to account for the varying sensor behaviors. This article designs a fault diagnosis method that deals with varying sensor behaviors in an online fashion. First, we list potential sensor varying behaviors by providing definitions of sensor states and sensor state transitions. Then, this article proposes the incremental varying sensory data-driven fault diagnosis model (IVSM). IVSM is able to update in an incremental manner under varying sensory data, with a theoretical performance guarantee. The primary objective of IVSM is to continuously map the heterogeneous sensory data within different time into a unified subspace, thereby enabling the direct measurement of heterogeneous and varying sensory data. Subsequently, it constructs a fault identification classifier within this unified subspace to determine the presence of faulty conditions in the systems. Its effectiveness and efficiency are verified by experimental results obtained from two public industrial systems and one practical industrial plant.

Federated learning (FL) and zero-shot learning have been becoming increasingly popular due to the data-privacy protection and the diagnosis of unseen faults in the industrial fault diagnosis. However, most existing diagnosis methods have the consistency assumption of distributions across different clients under multisource domain scenarios and cannot effectively diagnose both seen and unseen faults. Therefore, to diagnose both seen and unseen faults without data sharing and with distribution discrepancies across different clients, a federated generalized zero-sample fault diagnosis (GZSFD) paradigm is proposed in this article. In the client side, a stacked autoencoder (AE)-based feature extractor is introduced in each client for low-level features. In the cloud server, a feature-level distribution alignment scheme is developed to alleviate discrepancies for more discriminative high-level features. Moreover, a bidirectional AE (BAE) with reconstruction and cross-reconstruction streams is designed to enhance the feature-semantic consistency. Finally, a gating model based on BAE is proposed to identify online samples and mitigates the misclassification of unseen samples. Results on two practical industrial cases show that the proposed method achieves the improvement in federated GZSFD and effectively handles distribution discrepancies across different clients.

Health state prediction and fault time prediction are two key tasks in the fault prediction field. However, existing fault prediction techniques perform these tasks hierarchically and separately without considering the time-varying dynamics of the system operation process, which reduces the prediction efficiency and accuracy. Therefore, a Gaussian process regression prediction method based on the update strategy is proposed for the dual tasks of aeroengines. In this method, for new samples collected continuously, the predictive distributions are deduced and the model parameters are updated. Specifically, through variable selection and multivariable fusion technology, the most beneficial variables corresponding to the health state are used to construct a shared health index for health state and fault time prediction. The proposed health index can better characterize the health state. Then, by the update strategies including single-point and multipoint update strategies, a unified Gaussian process regression framework with newly collected samples information is obtained. Thereby, the health index and fault time prediction are realized synchronously. Experimental results on the commercial modular aero-propulsion system simulation dataset demonstrate that the proposed method outperforms state-of-the-art ones.

The scarcity of fault samples is a common scenario in the field of fault diagnosis. In the context of mechanical fault diagnosis, the emergence of new working conditions and fault modes renders the availability of samples of target (unseen) faults for model training unfeasible, thus limiting the performance of data-driven methods. Consequently, zero-sample learning and diagnosis of mechanical faults is a challenging task. In this regard, this article proposes a multiattribute learning model, inspired by the zero-shot learning paradigm, for zero-sample mechanical fault diagnosis. The key lies in the shared multiclass attribute classifiers. During the attribute learning process, a convolutional neural network is developed to construct multiclass attribute classifiers, which serve as a mapping between visual features and semantic features. These classifiers are transferred from readily available faults to enhance the capability of diagnosing unseen faults. By minimizing the difference among the fault attributes, the diagnosis of unseen faults is achieved, which includes fault location, size, working load, etc. Experiments on two real datasets verify the efficacy and the superiority of the proposed method.

Classifying different types of defects in light-emitting diode (LED) packaging poses a challenging and complex task. The difficulty in collecting diverse defect types presents a challenge characterized by a long-tailed distribution. Furthermore, this challenge is compounded by the complexity of the packing process, which involves a wide range of intricate defect types, making it difficult for the model to learn accurate decision boundaries. To tackle these issues, we propose the probabilistic aggregation network (PraNet) for detecting defects in LED packaging. First, the model learns to perform binary classification between normal and defect, then leverages binary classification confidence as a prior fusion for multiclass decision-making, mitigating the problem of low recall rates in the tail categories of the long-tailed distribution. Furthermore, by adapting the objective of supervised contrastive learning, we construct explicit decision boundaries in the feature space. Finally, we introduce a curriculum learning method to facilitate training coordination among all networks and loss functions. This method is verified on a real industrial process which includes 13 types of defects. Experimental results demonstrate the effectiveness of our proposed PraNet method in LED packaging defect classification task with imbalanced data, achieving an accuracy of 98.36% for normal binary classification and 93.47% for defects multiclass classification.

Most current domain adaptation methods for fault diagnosis focus on single target domain. However, test data often comes from multiple target domains, as machines work under different operating conditions, subsequently generating a more complex and extensive distribution of target data. Unfortunately, single target domain adaptation methods are not adaptive for multiple target domains adaptation (MTDA), which results in transfer performance degradation. To this end, a novel disentanglement learning with adaptive centroid alignment is proposed for MTDA. Specifically for disentanglement learning, two encoders and two classifiers are constructed independently for fault-related and domain-related feature extractions and classifications. Followed by the dual-adversarial strategy, only fault-related but domain-irrelevant features are extracted. Furthermore, to achieve the category alignment, we also propose an adaptive centroid alignment strategy, so that the feature centroids of the same fault category in different domains are enforced to be close to each other. Extensive experiments demonstrate the superiority of our proposed method compared with other popular approaches.

Zero-sample fault diagnosis (ZSFD) achieves remarkable success and has attracted considerable attention. However, existing methods suffer from the limitations in fault attribute labeling and feature extraction, leading to poor generalization and robustness. To be specific, fault attribute labeling that depends on expert knowledge is time-consuming and laborious. Fault feature extraction is carried out in one projecting space, and the useful knowledge of target data is ignored, which results in nonoptimal ZSFD performance. To tackle the above problems, a novel ZSFD based on transfer learning is proposed. First, a shared knowledge dictionary that automatically learns from source data with labels is transferred into the target data, which reduces the dependence of ZSFD on fault description. Second, a novel multiclass space projection model is designed to obtain the discriminative fault features. Finally, the pseudolabel mechanism is introduced to excavate the interclass and intraclass information in the target domain. The experiment results on the Tennessee–Eastman process and a real hot roll of steel process show the effectiveness of our method as well as its superiority.

Many real-world streaming data are sequentially collected over time and with skew-distributed classes. In this situation, online learning models may tend to favor samples from majority classes, making the wrong decisions for those from minority classes. Previous methods try to balance the instance number of different classes or assign asymmetric cost values. They usually require data-buffers to store streaming data or pre-defined cost parameters. This study alternatively shows that the imbalance of instances can be implied by the imbalance of gradients. Then, we propose the Online Harmonizing Gradient Descent (OHGD) for one-pass online classification. By harmonizing the gradient magnitude occurred by different classes, the method avoids the bias of the proposed method in favor of the majority class. Specifically, OHGD requires no data-buffer, extra parameters, or prior knowledge. It also handles imbalanced data streams the same way that it would handle balanced data streams, which facilitates its easy implementation. On top of a few common and mild assumptions, the theoretical analysis proves that OHGD enjoys a satisfying sub-linear regret bound. Extensive experimental results demonstrate the high efficiency and effectiveness in handling imbalanced data streams.

This study investigates the sample value imbalance problem of process monitoring. A fault detection approach based on variable selection and support vector data description (SVDD) is developed for efficient process monitoring. First, Kullback–Leibler divergence serves as the variable selection algorithm, which highlights the most beneficial information about the concerned faults. The attained variables are segmented by block division to avoid faults information being covered in single space monitoring, so that the relevant variables and the most beneficial information are concentrated in the same block. Then, Kernel principal component analysis is applied in each block to address the challenge that variables may still be high-dimensional and nonlinear. After that, the monitoring result is given based on the proposed SVDD with a restructured radius index, which is more sensitive to the fault. As demonstrated from experimental results on the Tennessee Eastman process, this method is effective and outperforms counterparts with higher mean fault detection rate. Note to Practitioners—Recently, multivariate statistical process monitoring (MSPM) has attracted much attention. In general, MSPM incorporates all variables for the large-scale process. However, only a small number of variables are fault-dependent. Namely, the sample value imbalance problem is encountered in application. In this scenario, the monitoring performance degrades and the online computational complexity increases. To this end, a SVDD-based fault detection method, which considers the fault-related variables, is proposed for process monitoring. The proposed method is verified by the Tennessee Eastman process and it is more sensitive to the concerned fault.

Incremental learning-based fault diagnosis is effective to learn from continuous industrial data on an ongoing basis. However, in the case of nonstationary industrial processes, new data distribution often gradually shifts away from that of historical data, due to equipment aging and manufacturing strategies. Thus, conventional incremental methods with the identical independent distribution (i.i.d.) assumption may no longer promise satisfied diagnosis performance. This article concerns the conditional drift phenomenon, a relaxation of the i.i.d. assumption, in which the conditional distribution of industrial data changes within different time. From a mathematical point of view, we first give the problem formulation of conditional drift and introduce a target mapping strategy for drift adaptation, under the minimum risk criteria. Then, following this strategy, an incremental diagnosis model with adaptation ability is designed. Particularly, a transformation matrix keeps matching the distributions of historical and new data. Thus, our method can quickly adapt to the conditional drift and be more robust against evolving environment. The proposed method is applied for diagnosing faults in two industrial processes to demonstrate its effectiveness.

A robotic system for harvesting tomatoes in greenhouses is designed. Effective recognition of ripen tomatoes from complex background is the key technology of the harvesting robotic system. In this work, the color feature of ripen tomatoes is employed. The ripen tomato is segmented by K-means clustering using the L*a*b* color space. To extract a single integrity ripen tomato, mathematical morphology method is used to denoise and handle the situations of tomato overlapping and shelter. Experimental results show the effectiveness of the proposed method.