Robust Fault Detection Method Research Articles

Hybrid robust fault detection and isolation of satellite reaction wheel actuators

In this paper, a combined robust fault detection and isolation scheme is studied for satellite system subject to actuator faults, external disturbances, and parametric uncertainties. The proposed methodology incorporates a residual generation module, including a bank of filters, into an intelligent residual evaluation module. First, residual filters are designed based on an improved nonlinear differential algebraic approach so that they are not affected by external disturbances. The residual evaluation module is developed based on the suggested series and parallel forms. Further, a new ensemble classification scheme defined as blended learning integrates heterogeneous classifiers to enhance the performance. A wide range of simulations is carried out in a high-fidelity satellite simulator subject to the constant and time-varying actuator faults in the presence of disturbances, manoeuvres, uncertainties, and noises. The obtained results demonstrate the effectiveness of the proposed robust fault detection and isolation method compared to the traditional nonlinear differential algebraic approach.

Robust fault detection for chemical processes based on dynamic low-rank matrix and optimized LSTM

The data collected by sensors in modern chemical process systems are always contaminated by industrial noise, so robust fault detection is an important technology to ensure process safety. However, the existing robust detection techniques have limitations in extracting dynamic and time series features of chemical process data. In this paper, a robust fault detection method based on dynamic low-rank matrix and optimized LSTM is proposed for dynamic chemical processes under noise background. First, a new low-rank matrix decomposition method, dynamic principal component pursuit (DPCP), is proposed for the dynamic characteristics of process data containing noise. The PCP is improved using the time delay excursion method, i.e., the DPCP is constructed by embedding an augmented dynamic matrix with a time lag factor on the PCP and solved using the alternating direction multiplier method. The purpose is to strip the useless noise from the useful detection information and extract the dynamic low-rank matrix. Second, an optimized LSTM (OPLSTM) is proposed for the time series features of dynamic low-rank matrix. vector factors for feature selection of memory cells in the LSTM are reweighted to balance the time series features of industrial processes and avoid heavy reliance on erroneous fault information. In addition, support vector data description (SVDD) is used to describe the hypersphere with clear boundary, and distance-based statistic is constructed to achieve fault detection with high fault detection rate and low false alarm rate. Finally, to assess the effectiveness of the proposed method, we performed extensive experiments on the Tennessee Eastman Process and the electrolytic aluminum process, and compared them with multiple methods. The experimental results show that the proposed method is effective and robust for chemical processes fault detection in the background of industrial noise.

Robust Sensor Fault Detection for a Single-Phase Pulse Width Modulation Rectifier

Maintaining safe and efficient operation in a single-phase pulse width modulation (PWM) rectifier that employs current sensors relies heavily on accurate sensor readings. However, several factors such as environmental conditions, aging, or damage can lead to sensor faults. Therefore, it is imperative to implement robust fault detection methods to ensure reliable system operation. The use of unknown input observer techniques is one such method that involves analyzing the differences between actual and estimated states to detect and identify faults in the system. This paper presents the development of a fault detection method that employs an unknown input observer with high sensitivity to faults and disturbance rejection to achieve robust fault detection. The method involves modeling the system as a state-space model and designing an observer to estimate the system’s state variables based on input and output measurements. The deviations between the actual and estimated states are then analyzed to detect and identify sensor faults, without the need for additional hardware, making it a cost-effective solution. Hardware-in-the-loop tests confirm the effectiveness of the proposed method.

Non-convex logarithm embedding subspace weighted graph approach to fault detection with missing measurements

In this work, we propose a robust fault detection method to address the problem of increase in missing and/or spurious alarms due to different degrees of loss or corruptions in data. Specifically, given historical data, our aim is to recover a clear low-space representation from corrupted data so that the detection control limits are updated, whenever required. To this end, a novel non-convex logarithm embedding subspace weighted graph detection method is presented. The method compensates the original missing nodes by embedding a non-convex logarithm regularizer, and then constructs an undirected graph model of the compensated nodes. Additionally, the dual constraints of l2,1-norm regularization and specific weights are introduced into the loss function of the graph model to further improve its robustness. Finally, the statistics and detectable criterion of the proposed method are given. Extensive simulations conducted on a real-world hot strip mill process and a multi-phase flow process demonstrate that the proposed method displays more robust detection than other state-of-the-art methods in the presence of outliers and missing measurements.

A Robust Fault Detection Method Based on Neighborhood Preserving Embedding

Industrial data are in general corrupted by noises and outliers, which do not meet the application assumptions in neighborhood preserving embedding (NPE). Many existing NPElike algorithms are not robust, overly consider the local features of the data and cannot capture the key features of the data. To this end, multi-level neighborhood preserving embedding (MNPE) is incorporated into the self-paced learning (SPL) framework in this study. Compared with single-projection feature extraction algorithms, MNPE consists of two projections: the nonreduced and reduced dimensionality projections. The nonreduced dimensionality projection can remove the redundant features that are unrelated to the key features of the data. The reduced dimensionality projection can reduce the dimensionality of the data and further extract the features of the data. Moreover, the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">L21</i> -norm can enhance the robustness of the proposed algorithm, and SPL framework can avoid the local optimal solution problem caused by non-convex optimization. Extensive experiments have been conducted on multiple image classification datasets to demonstrate that the proposed method is more effective than other state-of-the-art methods. The numerical simulation and blast furnace fault detection experiment results further validated the effectiveness of the proposed method.

Unsupervised Fault Detection With a Decision Fusion Method Based on Bayesian in the Pumping Unit

Since a large amount of data can be obtained in the oil production process nowadays and the operation environment is increasingly complicated, it is necessary to research unsupervised and robust fault detection methods for improving safety. In this paper, an online Bayesian-based technique with a novel decision fusion algorithm is proposed for unsupervised fault detection in the pumping unit. First, a new strategy to detect the working condition of the pumping unit by dynamometer card as well as five process measured variables is proposed. To deal with high-dimension data and outliers in dynamometer card, a robust Douglas-Peucker algorithm is developed for obtaining compressed data. A chord ratio index evaluating deviation degree of observations is defined, which can be used for removing outliers during approximation. Two norms are introduced for choosing the threshold in the proposed Douglas-Peucker algorithm. Moreover, a Bayesian-based online change point detection model is attempted for detecting univariate faults in the pumping unit. A decision fusion method derived from Bayesian probability formula is proposed for fusing univariate fault detection results. At last, the power of the proposed method is evaluated by numerical simulations and a real oil production process.

Real-Time and Robust Hydraulic System Fault Detection via Edge Computing

We consider fault detection in a hydraulic system that maintains multivariate time-series sensor data. Such a real-world industrial environment could suffer from noisy data resulting from inaccuracies in hardware sensing or external interference. Thus, we propose a real-time and robust fault detection method for hydraulic systems that leverages cooperation between cloud and edge servers. The cloud server employs a new approach that includes a genetic algorithm (GA)-based feature selection that identifies feature-to-label correlations and feature-to-feature redundancies. A GA can efficiently process large search spaces, such as solving a combinatorial optimization problem to identify the optimal feature subset. By using fewer important features that require transmission and processing, this approach reduces detection time and improves model performance. We propose a long short-term memory autoencoder for a robust fault detection model that leverages temporal information on time-series sensor data and effectively handles noisy data. This detection model is then deployed at edge servers that provide computing resources near the data source to reduce latency. Our experimental results suggest that this method outperforms prior approaches by demonstrating lower detection times, higher accuracy, and increased robustness to noisy data. While we have a 63% reduction of features, our model obtains a high accuracy of approximately 98% and is robust to noisy data with a signal-to-noise ratio near 0 dB. Our method also performs at an average detection time of only 9.42 ms with a reduced average packet size of 179.98 KB from the maximum of 343.78 KB.

Robust Fault Detection and Isolation of Discrete-Time LPV Systems Combining Set-theoretic UIO and Invariant Sets

This paper proposes a mixed active/passive robust fault detection and isolation (FDI) method for discrete-time linear paramter varying (LPV) systems based on set-theoretic unknown input observers (SUIO) and invariant sets. The robustness against system uncertainties in FDI of LPV systems can be guaranteed by actively decoupling or passively bounding their effect on residual signal. Furthermore, the quadratic H∞ stability condition of the LPV-form state-estimation-error dynamics is established based on a group of linear matrix inequalities (LMIs). Under the precondition of stability, a family of residual sets are constructed to establish set-separation guaranteed fault isolation (FI) conditions using invariant sets off-line. As long as the occurred faults satisfy the guaranteed FI conditions, they can be isolated from each other. At the end, a numerical example is used to illustrate the effectiveness of the proposed method.

Robust Fault Detection and Set-Theoretic UIO for Discrete-Time LPV Systems With State and Output Equations Scheduled by Inexact Scheduling Variables

This paper proposes a novel robust fault detection (FD) approach and designs a set-theoretic unknown input observer (SUIO) for linear parameter-varying (LPV) systems with both state and output equations scheduled by inexact scheduling variables. First, for such LPV systems, we propose a novel robust FD method by combing the set theory with the unknown input observer, which considers the bounds of measurement errors of scheduling variables to generate FD-oriented sets. In general, as long as sensors with sufficiently high precision are equipped to measure the scheduling variables, the bounds of measurement errors of scheduling variables can be less conservative than those direct bounds of scheduling variables, which can reduce robust FD conservatism in this way. Second, we give the unknown input decoupling condition of SUIO for such LPV systems and propose an SUIO design method under this condition for robust state estimation (SE). Besides, stability conditions for the proposed methods are established via matrix inequalities. At the end of this paper, a case study is used to illustrate the effectiveness of the proposed methods.

Invariant set-based robust fault detection and optimal fault estimation for discrete-time LPV systems with bounded uncertainties

ABSTRACTThis paper proposes an invariant set-based robust fault detection (FD) and optimal fault estimation (FE) method for discrete-time linear parameter varying (LPV) systems with bounded uncertainties. Firstly, a novel invariant-set construction method for discrete-time LPV systems is proposed if and only if the system is poly-quadratically stable, which need not satisfy the condition that there must exist a common quadratic Lyapunov function for all vertex matrices of the system compared to the traditional invariant-set construction methods. Furthermore, by using a shrinking procedure, we provide minimal robust positively invariant (mRPI) set approximations that are always positively invariant at each step of iteration and allow a priori desired precision to obtain a high sensitivity of FD. Owing to the existence of invariant set-based FD phase, the assumption that the initial faults should be bounded by a given set can be avoided for FE. We compute an optimal parametric matrix gain by minimising the Frobenius norm-based size of the corresponding FE set to obtain the optimal FE performance. Theoretically, any trajectory in the FE tube can be chosen as a specific-value estimation for the real fault signals. Finally, a vehicle dynamics system is used to illustrate the effectiveness of the proposed method.

Combining set-theoretic UIO and invariant sets for optimal guaranteed robust fault detection and isolation

This paper proposes an optimal guaranteed robust fault detection and isolation (FDI) method combing the set-theoretic unknown input observer (SUIO) and the robust positively invariant (RPI) sets for linear time-invariant (LTI) systems. The optimality of the proposed FDI method is achieved under a two-layer framework. The first layer allows to design a single optimal FDI-oriented SUIO with RPI set separation-based guaranteed FDI conditions to reduce the FDI conservatism of single observer. The second layer consists in designing an optimal observer configuration for a bank of optimal SUIOs to reduce the FDI conservatism of all the SUIOs together. Finally, the effectiveness of the proposed guaranteed robust FDI method is illustrated by using a four-tank system.

A Newly Robust Fault Detection and Diagnosis Method for High-Speed Trains

Incipient faults in high-speed trains are usually masked by noises and disturbances from process and sensors, which severely increases the difficulty of incipient fault detection and diagnosis. By introducing Hellinger distance into multivariate statistical analysis framework, this paper develops a robust detection and diagnosis method for incipient faults under the principal component analysis. The proposed method can detect all incipient sensor faults in traction systems of high-speed trains in real time by comparing reference probability density functions (PDFs) with the online estimated PDFs. According to the fault detection information, an accurate fault diagnosis can be achieved online through Bayesian inference. Key advantages of the proposed method are its salient robustness to unknown noises and disturbances, as well as the high sensitivity to incipient faults. In addition, the proposed method does not require any information on system models of high-speed trains or any human intervention. The effectiveness of the proposed method has been firstly proven by mathematical derivations and then been verified by numerical simulations. Finally, the proposed method has been applied to the practical experiment platform of the high-speed trains.

Conservatism comparison of set-based robust fault detection methods: Set-theoretic UIO and interval observer cases

This paper focuses on comparing fault detection (FD) conservatism of the set-theoretic unknown input observer (SUIO) based method and the interval observer (IO) based method. The former is implemented by combining both the set theory and the unknown input observer (UIO), where the set theory and the UIO are used to decouple disturbances and noises in passive and active ways, respectively. The latter is based on the set theory and the Luenberger-structure observer, which completely relies on the set-based passive decoupling of disturbances and noises to obtain FD robustness. The objective of this paper is to compare their FD conservatism using systematic mathematical analysis based on zonotopes and a mathematical criterion comparing their FD conservatism is obtained. Moreover, by establishing a general residual framework for conservatism analysis, the proposed method comparing the FD conservatism is also extended to compare the FD conservatism of a wide variety of set-based robust FD methods. At the end of this paper, a two-tank system example is used to illustrate the effectiveness of the obtained results.

A novel robust fault detection method for induction motor rotor by using unknown input observer

ABSTRACTThe purpose of this paper is to provide a robust solution for fault detection in the induction motor. The structure of nonlinear unknown input observer is investigated for observer design in this paper. Recently, rotor defect is considered as the most important non-electric fault in induction motors. One of the challenges in rotor fault detection is the uncertainty in the mechanical load of the motor and the harmonics in the electric power. Accordingly, these cases are considered as disturbances for unknown input observer and their effects on the residual signal are eliminated. The results of the simulation show the performance of the observer to diagnose the induction motor rotor fault in different scenarios.

Fault detection in distillation column using NARX neural network

Fault detection in the process industries is one of the most challenging tasks. It requires timely detection of anomalies which are present with noisy measurements of a large number of variable, highly correlated data with complex interactions and fault symptoms. This study proposes the robust fault detection method for the distillation column. Fault detection and diagnosis (FDD) for process monitoring and control has been an effective field of research for two decades. This area has been used widely in sophisticated engineering design applications to ensure the proper functionality and performance diagnosis of advanced and complex technologies. Robust fault detection of the realistic faults in distillation column in dynamic condition has been considered in this study. For early detection of faults, the model is based on nonlinear autoregressive with exogenous input (NARX) network. Tapped delays lines (TDLs) have been used for the input and output sequences. A case study was carried out with three different fault scenarios, i.e., valve sticking at reflux and reboiler, and tray upset. These faults would cause the product degradation. The normal data (no fault) is used for the training of neural network in all three cases. It is shown that the proposed algorithm can be used for the detection of both internal and external faults in the distillation column for dynamic system monitoring and to predict the probability of failure.

A fault detection optimization method based on chaos adaptive artificial fish swarm algorithm on distributed control system

For the purpose of improving fault detection accuracy of aero-engine distributed control system, an optimal design method based on chaos adaptive artificial fish swarm algorithm for distributed control system fault detection observer was proposed. First, in order to better simulate the actual distributed control system, dual channel multiple packet transmission was converted to switching system and short time-varying delay was modeled as an uncertainty in system models. Then, a fault observer is designed to obtain the state-space model of error system and the chaos adaptive artificial fish swarm algorithm is introduced. Furthermore, the ratio value of the residual signal transfer function, respectively, to the unknown disturbance signal and fault signal, are regarded as the objective function, and the minimum of the objective function is optimized to get the optimal observer matrix by chaos adaptive artificial fish swarm algorithm with the constraint of error system stability. Finally, the structure and operating principle of aero-engine distributed control system semi-physical platform is introduced, and comparison is made between the simulation of fault detection optimization method and the tradition robust fault detection method on the platform. The simulation results verify that the proposed method can efficiently reduce fault positive ratio and fault negative ratio simultaneously, which can expand the selection scope of threshold value and improve accuracy of fault detection.

A Robust Fault Detection Method using a Zonotopic Kaucher Set-membership Approach

This paper presents a robust fault detection method using a zonotopic Kaucher set-membership method. The fault detection approach is based on checking the consistency between the model and the data. Consistency is given if there is an intersection between the feasible parameter set and the nominal parameter set. To allow efficient computation the feasible set is approximated by a zonotope. Due to the usage of Kaucher interval arithmetic the results are mathematically guaranteed. The proposed approach is assessed using an illustrative application based on a well-known four-tank case study. The study shows that it is possible to detect even small errors in a noisy setting.

Detection of intermittent faults for linear stochastic systems subject to time‐varying parametric perturbations

In this study, the robust detection problem of intermittent faults (IFs) for linear stochastic systems subject to time-varying parametric perturbations is addressed. The authors consider the case that an IF appears and disappears non-deterministically, and lasts for random periods of active time with unknown magnitudes. A novel robust fault detection method is presented to detect all the appearing time and the disappearing time of an IF. Based on the output of an observer-type residual generator, a novel robust residual is constructed by utilising a sliding-time window. Two hypothesis tests are provided to detect all the appearing and the disappearing time, respectively. Moreover, the detectability of the IF by using the proposed robust detection scheme is defined in a probabilistic sense, and a sufficient detectability condition is presented within the given framework. Capacities of false alarm rates and missed detection rates are rigorously analysed. Finally, the application of the presented scheme is illustrated on a simplified radial flight control system and the results show that the IF can be effectively detected in the presence of perturbations and noises.

Application of Residual-Based EWMA Control Charts for Detecting Faults in Variable-Air-Volume Air Handling Unit System

An online robust fault detection method is presented in this paper for VAV air handling unit and its implementation. Residual-based EWMA control chart is used to monitor the control processes of air handling unit and detect faults of air handling unit. In order to provide a level of robustness with respect to modeling errors, control limits are determined by incorporating time series model uncertainty in EWMA control chart. The fault detection method proposed was tested and validated using real time data collected from real VAV air-conditioning systems involving multiple artificial faults. The results of validation show residual-based EWMA control chart with designing control limits can improve the accuracy of fault detection through eliminating the negative effects of dynamic characteristics, serial correlation, normal transient changes of system, and time series modeling errors. The robust fault detection method proposed can provide an effective tool for detecting the faults of air handling units.

Robust Fault Detection and Diagnosis for Multiple-Model Systems with Uncertainties

In this paper, a robust fault detection and diagnosis (FDD) method is proposed for multiple-model systems with modeling uncertainties. A compensation step is introduced to modify the mixed states and their variances obtained through the interacting multiple model (IMM) approximation and to solve the uncertainty problem. The degree of compensation is governed by a modification parameter determined by the orthogonality principle, which means that the estimation error calculated in the sub-filter using the true system models should be orthogonal to the residual error vector. To avoid over compensation in the unmatched models, a minimization procedure is used to derive the overall modification parameter. When the modification parameter is equal to one, the proposed method reduces to the IMM algorithm. An experiment is conducted through the ball and tube system to demonstrate the effectiveness of the proposed method.