With the continuous increase in complexity and expense of industrial systems, there is less tolerance for performance degradation, productivity decrease, and safety hazards, which greatly necessitates to detect and identify any kinds of potential abnormalities and faults as early as possible and implement real-time fault-tolerant operation for minimizing performance degradation and avoiding dangerous situations. During the last four decades, fruitful results have been reported about fault diagnosis and fault-tolerant control methods and their applications in a variety of engineering systems. The three-part survey paper aims to give a comprehensive review of real-time fault diagnosis and fault-tolerant control, with particular attention on the results reported in the last decade. In this paper, fault diagnosis approaches and their applications are comprehensively reviewed from model- and signal-based perspectives, respectively.

/pdf/a-survey-of-fault-diagnosis-and-fault-tolerant-techniques-9uuf1uqysf.pdf

A Survey of Fault Diagnosis and Fault-Tolerant Techniques—Part I: Fault Diagnosis With Model-Based and Signal-Based Approaches

Deep learning and its applications to machine health monitoring

In this paper, two online schemes for an integrated design of fault-tolerant control (FTC) systems with application to Tennessee Eastman (TE) benchmark are proposed. Based on the data-driven design of the proposed fault-tolerant architecture whose core is an observer/residual generator based realization of the Youla parameterization of all stabilization controllers, FTC is achieved by an adaptive residual generator for the online identification of the fault diagnosis relevant vectors, and an iterative optimization method for system performance enhancement. The performance and effectiveness of the proposed schemes are demonstrated through the TE benchmark model.

Real-Time Implementation of Fault-Tolerant Control Systems With Performance Optimization

In modern manufacturing systems and industries, more and more research efforts have been made in developing effective machine health monitoring systems. Among various machine health monitoring approaches, data-driven methods are gaining in popularity due to the development of advanced sensing and data analytic techniques. However, considering the noise, varying length and irregular sampling behind sensory data, this kind of sequential data cannot be fed into classiﬁcation and regression models directly. Therefore, previous work focuses on feature extraction/fusion methods requiring expensive human labor and high quality expert knowledge. With the development of deep learning methods in the last few years, which redeﬁne representation learning from raw data, a deep neural network structure named Convolutional Bi-directional Long Short-Term Memory networks (CBLSTM) has been designed here to address raw sensory data. CBLSTM ﬁrstly uses CNN to extract local features that are robust and informative from the sequential input. Then, bi-directional LSTM is introduced to encode temporal information. Long Short-Term Memory networks(LSTMs) are able to capture long-term dependencies and model sequential data, and the bi-directional structure enables the capture of past and future contexts. Stacked, fully-connected layers and the linear regression layer are built on top of bi-directional LSTMs to predict the target value. Here, a real-life tool wear test is introduced, and our proposed CBLSTM is able to predict the actual tool wear based on raw sensory data. The experimental results have shown that our model is able to outperform several state-of-the-art baseline methods.

/pdf/learning-to-monitor-machine-health-with-convolutional-bi-2vok1ch88q.pdf

Learning to Monitor Machine Health with Convolutional Bi-Directional LSTM Networks.

Standard partial least squares (PLS) serves as a powerful tool for key performance indicator (KPI) monitoring in large-scale process industry for last two decades. However, the standard approach and its recent modifications still encounter some problems for fault diagnosis related to KPI of the underlying process. To cope with these difficulties, an improved PLS (IPLS) approach is presented in this paper. IPLS is able to decompose the measurable process variables into the KPI-related and unrelated parts, respectively. Based on it, the corresponding test statistics are designed to offer meaningful fault diagnosis information and thus, the corresponding maintenance actions can be further taken to ensure the desired performance of the systems. In order to demonstrate the effectiveness of the proposed approach, a numerical example and Tennessee Eastman (TE) benchmark process are respectively utilized. It can be seen that the proposed approach shows satisfactory results not only for diagnosing KPI-related faults but also for its high fault detection rate.

Improved PLS Focused on Key-Performance-Indicator-Related Fault Diagnosis

The control of automobile electronic throttle (AET) systems is a challenging task owing to multiple nonlinearities, i.e., transmission friction, gear backlash, limp-home spring set, and external load disturbance. In this paper, a practical tracking control scheme of an AET system is developed using a continuous fast nonsingular terminal sliding mode (CFNTSM) technique based on uncertainty observer. By using the prescribed CFNTSM surface and fast terminal sliding mode-based reaching element, the proposed control implementation guarantees the fast error convergence characteristic and high tracking accuracy under parameter uncertainties and perturbations. Furthermore, due to the adoption of the finite-time exact observer (FEO) for the lumped uncertainty estimation in the controller, the ease of the selection of the control gains is well-achieved since they only depend on the uncertainty estimation error. The closed-loop stability and finite-time convergence are presented based on the Lyapunov stability theory. Experimental verifications are conducted to validate the remarkable performance of the proposed control, in terms of the step and sinusoidal tracking as well as anti-disturbance ability.

Continuous Fast Nonsingular Terminal Sliding Mode Control of Automotive Electronic Throttle Systems Using Finite-Time Exact Observer

This paper proposes a novel adaptive terminal sliding-mode (ATSM) control scheme for a Steer-by-Wire (SBW) vehicle. It is shown that the developed ATSM controller can drive the closed-loop error dynamics to converge to zero in a finite time, where adaptive laws are applied to estimate the uncertain bounds of the system parameters and disturbances in Lyapunov sense. Compared with the sliding-mode-based SBW control systems, the proposed ATSM control not only assures the finite-time error convergence and strong robustness with respect to parameter uncertainties and varying driving conditions, but also requires no prior knowledge of the system parameters and road information. Experimental results from an SBW vehicle are demonstrated to verify the remarkable performance of the proposed control in terms of the robustness, error convergence, and road disturbance attenuation, in comparison with other control strategies.

Design and Implementation of Adaptive Terminal Sliding-Mode Control on a Steer-by-Wire Equipped Road Vehicle

Recently, a bond-graph-based fault detection and isolation (FDI) framework has been developed with a new concept of global analytical redundancy relations (GARRs) (Low, Wang, Arogeti, and Luo, 2009, 2010; Low, Wang, Arogeti, and Zhang, 2010). This new concept allows the fault diagnosis for hybrid systems which consist of both continuous dynamics and discrete modes. A failure of a safety critical system such as the steering system of an automated guided vehicle may cause severe damage. Such failure can be avoided by an early detection and estimation of faults. In this paper, the newly developed FDI method is studied in details using an electrohydraulic steering system of an electric vehicle. The steering system and faults are modeled as a hybrid dynamic system by the hybrid bond graph (HBG) modeling technique. GARRs are then derived systematically from the HBG model with a specific causality assignment. Fault detection, isolation, and estimation are applied, experimental setup is described, and results are discussed.

Fault Detection Isolation and Estimation in a Vehicle Steering System

This book systematically presents a comprehensive framework and effective techniques for in-depth analysis, clear design procedure, and efficient implementation of diagnosis and prognosis algorithms for hybrid systems. It offers an overview of the fundamentals of diagnosis\prognosis and hybrid bond graph modeling. This book also describes hybrid bond graph-based quantitative fault detection, isolation and estimation. Moreover, it also presents strategies to track the system mode and predict the remaining useful life under multiple fault condition. A real world complex hybrid systema vehicle steering control systemis studied using the developed fault diagnosis methods to show practical significance. Readers of this book will benefit from easy-to-understand fundamentals of bond graph models, concepts of health monitoring, fault diagnosis and failure prognosis, as well as hybrid systems. The reader will gain knowledge of fault detection and isolation in complex systems including those with hybrid nature, and will learn state-of-the-art developments in theory and technologies of fault diagnosis and failure prognosis for complex systems.

Model-based Health Monitoring of Hybrid Systems

Prognosis of hybrid systems is a challenging problem because multiple faults may happen simultaneously at a mode where these faults have different detectability. In other words, at the fault-initiating mode, some of the faults are detectable while others are nondetectable. As a result, decision making based on only one observation of abnormal behavior is not reliable under this condition. This paper focuses on the development of a model-based prognosis framework for hybrid systems where a dynamic fault isolation scheme is proposed to facilitate the prognostic tasks. The degradation behavior of each faulty component is mode dependent and can be estimated by a hybrid differential evolution algorithm. Thereafter, the remaining useful life of the faulty component that varies with different operating modes is calculated by using both the estimated degradation model and the user-selected failure threshold. Experiments are carried out to validate the key concepts of the developed methods, and results suggest the effectiveness.

Ming Yu

Papers

Continuous Fast Nonsingular Terminal Sliding Mode Control of Automotive Electronic Throttle Systems Using Finite-Time Exact Observer

Design and Implementation of Adaptive Terminal Sliding-Mode Control on a Steer-by-Wire Equipped Road Vehicle

Fault Detection Isolation and Estimation in a Vehicle Steering System

Model-based Health Monitoring of Hybrid Systems

Model-Based Prognosis for Hybrid Systems With Mode-Dependent Degradation Behaviors