Bayesian Hierarchical Graph Neural Networks With Uncertainty Feedback for Trustworthy Fault Diagnosis of Industrial Processes.

Deep learning (DL) methods have been widely applied to intelligent fault diagnosis of industrial processes and achieved state-of-the-art performance. However, fault diagnosis with point estimate may provide untrustworthy decisions. Recently, Bayesian inference shows to be a promising approach to trustworthy fault diagnosis by quantifying the uncertainty of the decisions with a DL model. The uncertainty information is not involved in the training process, which does not help the learning of highly uncertain samples and has little effect on improving the fault diagnosis performance. To address this challenge, we propose a Bayesian hierarchical graph neural network (BHGNN) with an uncertainty feedback mechanism, which formulates a trustworthy fault diagnosis on the Bayesian DL (BDL) framework. Specifically, BHGNN captures the epistemic uncertainty and aleatoric uncertainty via a variational dropout approach and utilizes the uncertainty information of each sample to adjust the strength of the temporal consistency (TC) constraint for robust feature learning. Meanwhile, the BHGNN method models the process data as a hierarchical graph (HG) by leveraging the interaction-aware module and physical topology knowledge of the industrial process, which integrates data with domain knowledge to learn fault representation. Moreover, the experiments on a three-phase flow facility (TFF) and secure water treatment (SWaT) show superior and competitive performance in fault diagnosis and verify the trustworthiness of the proposed method.

PLC-Informed Distributed Game Theoretic Learning of Energy-Optimal Production Policies.

This article describes a novel concept to optimize manufacturing systems distributively through data-based learning. We propose a game-theoretic (GT) learning set-up that is incorporated with accessible control code of the programmable logic controller (PLC) to accelerate the optimal policies learning procedures, instead of learning everything from scratch. Therefore, we offer to process the accessible and available control code into a GT-based learning framework which is subsequently optimized in a fully distributed manner. To this end, we employ the recently developed framework of state-based potential games (PGs) and prove that under mild conditions PLC-informed (PLCi) learning forms a state-based PG framework. We conduct the experiment on a laboratory scale testbed in numerous production scenarios. The experiment's results highlight the major potential of using the PLCi GT-learning, which is the reduction of energy consumption of the production timescales and improvement of production efficiency while nearly halven the learning times.

Interaction-Aware Graph Neural Networks for Fault Diagnosis of Complex Industrial Processes.

Fault diagnosis of complex industrial processes becomes a challenging task due to various fault patterns in sensor signals and complex interactions between different units. However, how to explore the interactions and integrate with sensor signals remains an open question. Considering that the sensor signals and their interactions in an industrial process with the form of nodes and edges can be represented as a graph, this article proposes a novel interaction-aware and data fusion method for fault diagnosis of complex industrial processes, named interaction-aware graph neural networks (IAGNNs). First, to describe the complex interactions in an industrial process, the sensor signals are transformed into a heterogeneous graph with multiple edge types, and the edge weights are learned by the attention mechanism, adaptively. Then, multiple independent graph neural network (GNN) blocks are employed to extract the fault feature for each subgraph with one edge type. Finally, each subgraph feature is concatenated or fused by a weighted summation function to generate the final graph embedding. Therefore, the proposed method can learn multiple interactions between sensor signals and extract the fault feature from each subgraph by message passing operation of GNNs. The final fault feature contains the information from raw data and implicit interactions between sensor signals. The experimental results on the three-phase flow facility and power system (PS) demonstrate the reliable and superior performance of the proposed method for fault diagnosis of complex industrial processes.

Curriculum-Based Deep Reinforcement Learning for Quantum Control.

Deep reinforcement learning (DRL) has been recognized as an efficient technique to design optimal strategies for different complex systems without prior knowledge of the control landscape. To achieve a fast and precise control for quantum systems, we propose a novel DRL approach by constructing a curriculum consisting of a set of intermediate tasks defined by fidelity thresholds, where the tasks among a curriculum can be statically determined before the learning process or dynamically generated during the learning process. By transferring knowledge between two successive tasks and sequencing tasks according to their difficulties, the proposed curriculum-based DRL (CDRL) method enables the agent to focus on easy tasks in the early stage, then move onto difficult tasks, and eventually approaches the final task. Numerical comparison with the traditional methods [gradient method (GD), genetic algorithm (GA), and several other DRL methods] demonstrates that CDRL exhibits improved control performance for quantum systems and also provides an efficient way to identify optimal strategies with few control pulses.

Gradient Monitored Reinforcement Learning.

This article presents a novel neural network training approach for faster convergence and better generalization abilities in deep reinforcement learning (RL). Particularly, we focus on the enhancement of training and evaluation performance in RL algorithms by systematically reducing gradient's variance and, thereby, providing a more targeted learning process. The proposed method, which we term gradient monitoring (GM), is a method to steer the learning in the weight parameters of a neural network based on the dynamic development and feedback from the training process itself. We propose different variants of the GM method that we prove to increase the underlying performance of the model. One of the proposed variants, momentum with GM (M-WGM), allows for a continuous adjustment of the quantum of backpropagated gradients in the network based on certain learning parameters. We further enhance the method with the adaptive M-WGM (AM-WGM) method, which allows for automatic adjustment between focused learning of certain weights versus more dispersed learning depending on the feedback from the rewards collected. As a by-product, it also allows for automatic derivation of the required deep network sizes during training as the method automatically freezes trained weights. The method is applied to two discrete (real-world multirobot coordination problems and Atari games) and one continuous control task (MuJoCo) using advantage actor-critic (A2C) and proximal policy optimization (PPO), respectively. The results obtained particularly underline the applicability and performance improvements of the methods in terms of generalization capability.

Distributed Adaptive Leader-Following Consensus for Nonlinear Multiagent Systems With Actuator Failures Under Directed Switching Graphs.

This article studies the distributed adaptive failures compensation output-feedback consensus for a class of nonlinear multiagent systems (MASs) with multiactuator failures allowing unmatched redundancy under directed switching graphs. With estimated information of neighbors, a novel distributed reference generator is designed. To compensate the unmeasured state variables of each agent, a reduced-order dynamic gain filter is constructed. Based on the generator and filter, and using the recursive design method, a distributed adaptive protocol is designed, where the adaptive technique is used to compensate the actuator failures. The proposed scheme can significantly relax conditions on the communication graph, which allows the graph to be disconnected at any time instant. The number of introduced variables in the filter and its dimension is greatly reduced and, thus, reduces the numerical challenge. The output-feedback consensus for nonlinear MASs with actuator failures and possible unmatched actuator redundancy is addressed for the first time. The consensus error can converge to an arbitrarily small set not affected by actuator failures, and the resulting closed-loop system is semiglobally stable. Finally, simulation results are given to illustrate the effectiveness of the proposed method.

Robust Asymptotic Fault Estimation of Discrete-Time Interconnected Systems With Sensor Faults.

In this article, a robust asymptotic fault estimation (RAFE) design is proposed for discrete-time interconnected systems with sensor faults. By constructing a singular augmented system, an equivalent description of the considered interconnected systems is presented. Then, a novel RAFE observer is proposed for the singular augmented system. Furthermore, gain matrices of the RAFE observer are calculated based on multiconstrained design. Simulation results are illustrated to show the feasibility of the presented approaches.

Distributed Self-Optimization of Modular Production Units: A State-Based Potential Game Approach.

This article presents a novel approach for distributed optimization of production units based on potential game (PG) theory and machine learning. The core of our approach is split into two parts: the first part concentrates on the conceptual treatment of modular installed production units in terms of a PG scenario. The second part focuses on the development and incorporation of suitable learning algorithms to finally form an intelligent autonomous system. In this context, we model the production environment as a state-based PG where each actuator of each module has the role of an agent in the game aiming to maximize its utility value by learning the optimal process behavior. The benefit of the additional state information is visible in the performance of the algorithm making the environment dynamic and serving as a connector between the players. We propose a novel learning algorithm based on a global interpolation method that is applied to a laboratory scale modular bulk good system. The thorough analysis of the encouraging results yields to highly interesting insights into the learning dynamics and the process itself. The benefits of our distributed optimization approach are the plug-and-play functionality, the online capability, fast adaption to changing production requirements, and the possibility of an IEC 61131 conforming to PLC implementation.

A Comparative Study of Deep Neural Network-Aided Canonical Correlation Analysis-Based Process Monitoring and Fault Detection Methods.

Multivariate analysis is an important kind of method in process monitoring and fault detection, in which the canonical correlation analysis (CCA) makes use of the correlation change between two groups of variables to distinguish the system status and has been greatly studied and applied. For the monitoring of nonlinear dynamic systems, the deep neural network-aided CCA (DNN-CCA) has received much attention recently, but it lacks a general definition and comparative study of different network structures. Therefore, this article first introduces four deep neural network (DNN) models that are suitable to combine with CCA, and the general form of DNN-CCA is given in detail. Then, the experimental comparison of these methods is conducted through three cases, so as to analyze the characteristics and distinctions of CCA aided by each DNN model. Finally, some suggestions on method selection are summarized, and the existed open issues in the current DNN-CCA form and future directions are discussed.

Extended Relevance Vector Machine-Based Remaining Useful Life Prediction for DC-Link Capacitor in High-Speed Train.

Remaining useful life (RUL) prediction is a reliable tool for the health management of components. The main concern of RUL prediction is how to accurately predict the RUL under uncertainties. In order to enhance the prediction accuracy under uncertain conditions, the relevance vector machine (RVM) is extended into the probability manifold to compensate for the weakness caused by evidence approximation of the RVM. First, tendency features are selected based on the batch samples. Then, a dynamic multistep regression model is built for well describing the influence of uncertainties. Furthermore, the degradation tendency is estimated to monitor degradation status continuously. As poorly estimated hyperparameters of RVM may result in low prediction accuracy, the established RVM model is extended to the probabilistic manifold for estimating the degradation tendency exactly. The RUL is then prognosticated by the first hitting time (FHT) method based on the estimated degradation tendency. The proposed schemes are illustrated by a case study, which investigated the capacitors' performance degradation in traction systems of high-speed trains.

Deep Convolutional Clustering-Based Time Series Anomaly Detection.

This paper presents a novel approach for anomaly detection in industrial processes. The system solely relies on unlabeled data and employs a 1D-convolutional neural network-based deep autoencoder architecture. As a core novelty, we split the autoencoder latent space in discriminative and reconstructive latent features and introduce an auxiliary loss based on k-means clustering for the discriminatory latent variables. We employ a Top-K clustering objective for separating the latent space, selecting the most discriminative features from the latent space. We use the approach to the benchmark Tennessee Eastman data set to prove its applicability. We provide different ablation studies and analyze the method concerning various downstream tasks, including anomaly detection, binary and multi-class classification. The obtained results show the potential of the approach to improve downstream tasks compared to standard autoencoder architectures.

Design of a Discrete-Time Fault-Tolerant Quantum Filter and Fault Detector.

This paper solves the problem of discrete-time fault-tolerant quantum filtering for a class of laser-atom open quantum systems subject to the stochastic faults. We show that by using the discrete-time quantum measurements, optimal estimates of both the atomic observables and the classical fault process can be simultaneously determined in terms of recursive quantum stochastic difference equations. A dispersive interaction quantum system example is used to demonstrate the proposed filtering approach.

Performance-Based Fault Detection and Fault-Tolerant Control for Nonlinear Systems With T-S Fuzzy Implementation.

This article addresses the performance-based fault detection (FD) and fault-tolerant control (FTC) issues for nonlinear systems. For this purpose, in the first part of this article, the performance-based FD and FTC scheme is investigated with the aid of the nonlinear factorization technique. To be specific, the controller parameterization for nonlinear systems is first discussed. The so-called fault-tolerant margin is introduced as an indicator of the system fault-tolerant ability. Then, the FD scheme aiming at estimating and detecting the stability performance degradation of the closed-loop system caused by the system faults is developed. Furthermore, to recover the system performance, the performance-based FTC strategy is presented. In the second part of this article, the design approach of the performance-based FD and FTC scheme is studied by applying the Takagi-Sugeno fuzzy dynamic modeling technique. The achieved results are demonstrated in the end by a case study on the three-tank system.

Bidirectional deep recurrent neural networks for process fault classification.

In this study, a new approach for time series based condition monitoring and fault diagnosis based on bidirectional recurrent neural networks is presented. The application of bidirectional recurrent neural networks essentially provide a viewpoint change on the fault diagnosis task, which allows to handle fault relations over longer time horizons helping in avoiding critical process breakdowns and increasing the overall productivity of the system. To further enhance the capability, we propose a novel procedure of data preprocessing and restructuring which enforces the generalization and a more efficient data utilization and consequently yields more efficient network training, especially for sequential fault classification task. The proposed Bidirectional Long Short Term Memory network outperforms standard recurrent architectures including vanilla recurrent neural networks, Long Short Term Memories and Gated Recurrent Units. We apply the proposed approach to the Tennessee Eastman benchmark process to test the effectiveness of the mentioned deep architectures and provide a detailed comparative analysis. The experimental results for binary as well as multi-class classification show the superior average fault detection capability of the bidirectional Long Short Term Memory Networks compared to the other architectures and to results from other state-of-the-art architectures found in the literature.

Fault detection for piecewise affine systems with application to ship propulsion systems.

In this paper, the design approach of non-synchronized diagnostic observer-based fault detection (FD) systems is investigated for piecewise affine processes via continuous piecewise Lyapunov functions. Considering that the dynamics of piecewise affine systems in different regions can be considerably different, the weighting matrices are used to weight the residual of each region, so as to optimize the fault detectability. A numerical example and a case study on a ship propulsion system are presented in the end to demonstrate the effectiveness of the proposed results.

A KPI-based process monitoring and fault detection framework for large-scale processes.

Large-scale processes, consisting of multiple interconnected subprocesses, are commonly encountered in industrial systems, whose performance needs to be determined. A common approach to this problem is to use a key performance indicator (KPI)-based approach. However, the different KPI-based approaches are not developed with a coherent and consistent framework. Thus, this paper proposes a framework for KPI-based process monitoring and fault detection (PM-FD) for large-scale industrial processes, which considers the static and dynamic relationships between process and KPI variables. For the static case, a least squares-based approach is developed that provides an explicit link with least-squares regression, which gives better performance than partial least squares. For the dynamic case, using the kernel representation of each subprocess, an instrument variable is used to reduce the dynamic case to the static case. This framework is applied to the TE benchmark process and the hot strip mill rolling process. The results show that the proposed method can detect faults better than previous methods.

Real-Time Fault Detection Approach for Nonlinear Systems and its Asynchronous T-S Fuzzy Observer-Based Implementation.

This paper is concerned with a real-time observer-based fault detection (FD) approach for a general type of nonlinear systems in the presence of external disturbances. To this end, in the first part of this paper, we deal with the definition and the design condition for an L ∞ / L 2 type of nonlinear observer-based FD systems. This analytical framework is fundamental for the development of real-time nonlinear FD systems with the aid of some well-established techniques. In the second part, we address the integrated design of the L ∞ / L 2 observer-based FD systems by applying Takagi-Sugeno (T-S) fuzzy dynamic modeling technique as the solution tool. This fuzzy observer-based FD approach is developed via piecewise Lyapunov functions, and can be applied to the case that the premise variables of the FD system is nonsynchronous with the premise variables of the fuzzy model of the plant. In the end, a case study on the laboratory setup of three-tank system is given to show the efficiency of the proposed results.

A data-driven multiplicative fault diagnosis approach for automation processes.

This paper presents a new data-driven method for diagnosing multiplicative key performance degradation in automation processes. Different from the well-established additive fault diagnosis approaches, the proposed method aims at identifying those low-level components which increase the variability of process variables and cause performance degradation. Based on process data, features of multiplicative fault are extracted. To identify the root cause, the impact of fault on each process variable is evaluated in the sense of contribution to performance degradation. Then, a numerical example is used to illustrate the functionalities of the method and Monte-Carlo simulation is performed to demonstrate the effectiveness from the statistical viewpoint. Finally, to show the practical applicability, a case study on the Tennessee Eastman process is presented.

Fuzzy state/disturbance observer design for T-S fuzzy systems with application to sensor fault estimation.

A novel fuzzy-observer-design approach is presented for Takagi-Sugeno fuzzy models with unknown output disturbances. In order to decouple the unknown output disturbance, an augmented fuzzy descriptor model is constructed by supposing the disturbance to be an auxiliary state vector. A fuzzy state-space observer is next designed for the augmented fuzzy descriptor system, and the simultaneous estimates of the original state and disturbance are thus obtained. The proposed observer technique is further applied to estimate sensor faults. Finally, a numerical example is given to illustrate the design procedure, and the simulation results show the desired tracking performance. The preknowledge of the disturbance and fault is not necessary for our design. Moreover, the considered disturbance and sensor fault can be in any form.