Active Fault Diagnosis Method Research Articles

Active Fault Diagnosis for Stochastic Systems Within Bayesian Minimum Risk Decision Framework

In this paper, the active fault diagnosis (AFD) problem is investigated for a class of discrete-time stochastic systems. We consider multiple fault modes of the system, and the different types of misdiagnoses of these faults lead to different losses. A novel AFD framework is developed based on the Bayesian minimum risk decision framework. The misdiagnosis risk is derived to characterize the total losses caused by the misdiagnoses of all considered faults. The input is specifically designed to improve diagnosis performance by minimizing the risk of misdiagnosis. Since it is intractable to express the misdiagnosis risk in the closed form, we deduce an upper bound on it to establish a novel input design criterion based on the Bhattacharyya distance. The optimal input signal can be obtained for AFD via concave quadratic programming. For diagnosis performance enhancement and computation simplification purposes, we propose two new semi-online implementation manners of the input design for stochastic systems with and without multi-fault modes switching respectively. Given the optimal input signal and the proposed implementation manner, faults are diagnosed based on the minimum misdiagnosis risk decision principle. The effectiveness and superiority of the proposed AFD method are demonstrated by simulation results on a four-tank system.

Active fault diagnosis for linear stochastic systems subject to chance constraints

In this paper, the problem of active fault diagnosis (AFD) is investigated for a class of linear stochastic systems subject to chance constraints, where the input signals are specifically designed to enhance diagnosis performance by discriminating between multi-fault modes. As the misdiagnosis probability cannot be expressed in the closed form, a reachable upper bound on it is derived to construct a novel input design criterion for AFD based on the Cauchy–Schwarz divergence. A sufficient condition is provided to decompose the multivariate chance constraints into a set of univariate ones. A three-stage optimization approach is proposed to solve the input design problem incorporating chance constraints for AFD. In the upper stage, the best iteration parameter is determined for subsequent iterative optimization to obtain the least conservative risk allocation scheme. In the middle stage, the total risk of violating the multivariate constraints is optimally allocated to a set of univariate ones through iterative optimization, which can improve diagnosis performance. In the lower stage, a set of univariate chance constraints is equivalently converted into expectation constraints under the allocated risk, and then the input design problem under the constructed design criterion is solved to global optimality. The optimal input signals are injected, and the AFD decision is made based on real-time measurement information according to the minimum probability of misdiagnosis decision principle. Finally, simulation results on a four-tank system demonstrate the effectiveness and superiority of the proposed AFD method.

Four-state active fault diagnosis method for cascaded H-bridge multilevel inverter

The cascaded multilevel inverters have many semiconductor switches with high failure rate, which require timely diagnosis techniques. It is a difficult point that the diagonally positioned switches have similar fault characteristics. By analyzing the output characteristics under fault, it is found that the position and the cause of similar faults are related to the permutation of carrier in pulse width modulation. A four-state active fault diagnosis method is proposed in this article. First, the permutation of carriers is reconstructed to activate the fault behavior after the fault is detected. Under the specific four modulation states, the fault characteristics can be fully behaved. Next, based on the machine learning algorithm, the effective fault features are extracted from the voltage signals in four modulation states and the fault location is achieved. The experiments show that the proposed method could effectively distinguish similar faults, improve diagnostic accuracy and have better robustness.

Minimal Detectable and Isolable Faults of Active Fault Diagnosis

This article aims to compute guaranteed minimal detectable and isolable faults of set-based active fault diagnosis methods for discrete linear time-invariant systems. First, guaranteed detectable and isolable faults of set-based active fault diagnosis methods are defined in this article. Second, an augmented vector is defined to integrate the effect of both inputs and faults such that the couplings of inputs and faults can be handled effectively. Third, guaranteed minimal detectable and isolable faults are computed by solving a mixed integer quadratic fractional programming problem under a proposed theoretic framework. At the end of this article, an example is used to illustrate the effectiveness of the proposed solution.

Active fault diagnosis under hybrid bounded and Gaussian uncertainties

In this paper, an active fault diagnosis method is proposed to diagnose faults of linear time-invariant systems under the effect of hybrid bounded and Gaussian uncertainties, which can guarantee an arbitrary confidence level of fault diagnosis. Zonotopes are used to characterize bounded uncertainties. Gaussian uncertainties are handled as ellipsoids by the given confidence coefficient. The confidence domains of output are formulated as the Minkowski sum of a zonotope and an ellipsoid. The fault diagnosis task is done based on the separation of all the confidence domains of different fault situations, which leads to a Mathematical Program with Complementarity Constraints (MPCC). A branch-and-bound method and a smoothing method are provided to solve the program. The effectiveness of the proposed active fault diagnosis method is illustrated by two examples at the end of the paper.

Active fault diagnosis for a class of closed-loop systems via parameter estimation

This paper addresses an active fault diagnosis problem for a class of discrete-time closed-loop system with stochastic noise. By introducing the theories of system identification, a novel active fault diagnosis method is developed to detect and isolate the faults. An important advantage of the proposed method is that there is no need to cut off the original input signal, which is necessary in most active fault diagnosis methods. Firstly, due to the features of the faults, we transform the problem of fault diagnosis into a problem of model selection by estimating model parameters. Then, the sufficient condition for active fault diagnosability is analysed, and the property that auxiliary input signal can enhance the fault diagnosability is given. Finally, simulation studies are carried out to demonstrate the effectiveness and applicability of the proposed method.

Active Fault Diagnosis for Linear Systems: Within a Signal Processing Framework

Compared with passive fault diagnosis (PFD) techniques, the research on active fault diagnosis (AFD) techniques is insufficient. The existing AFD methods generally include set-based method, Bayesian approach, and so on. Few AFD results within the framework of signal processing have been published. In this article, a new AFD framework, which is composed of auxiliary input generation, fault diagnosis observer, and signal analysis unit, is introduced into AFD analysis for the first time, within the signal processing framework. Auxiliary signal generation unit is used to generate appropriate auxiliary input signal according to the requirements of fault diagnosis. Then, a set of fault diagnosis observers are designed to generate residuals corresponding to the input channels one to one. Two signal processing methods, correlation analysis and ESPRIT algorithm, are applied to extract the relationship between auxiliary signal and residual signal for fault diagnosis. Numerical examples are given to illustrate the effectiveness of the proposed method.

Actuator fault tolerant control via active fault diagnosis for a remotely operated vehicle

Unmanned underwater vehicles operate in unstructured and harsh environments, and fault diagnosis together with fault tolerant control play a crucial role to prevent the mission from failing. When a fault occurs in presence of unknown disturbances, such as unknown marine current or hitting an obstacle, actuator fault diagnosis becomes more challenging, because the unexpected forces that arise from disturbances may be mistaken for actuator faults. In this paper, we propose an active fault diagnosis method to discern between an actuator fault and any other disturbance: the former is modelled as a multiplicative input fault, and the latter is an unknown, additive, time varying force. Once the actuator faults have been isolated and identified, their estimations are fed to the control allocation algorithm to perform fault tolerant control allocation. Simulation results are performed with a realistic non-linear model to show the effectiveness of the proposed strategy.

Observer-based asymptotic active fault diagnosis: A two-layer optimization framework

This paper proposes a novel asymptotic online robust active fault diagnosis (AFD) framework for discrete linear time-invariant (LTI) systems using a bank of set-valued observers (SVO), each SVO designed to match a healthy/faulty actuator mode. Different from the existing set-based AFD methods, the proposed AFD method designs an input at each time instant by solving a two-layer optimization problem. At each time instant, an input is designed such that the output sets estimated by the SVOs at the next time instant keep away from each other as far as possible. With this idea, faults can be diagnosed after a sufficiently long period by designing and injecting inputs into the system online. The key point of the proposed AFD method is to establish a logic describing the process keeping a group of output estimation sets asymptotically away from each other as far as possible, which is done by formulating a two-layer optimization problem to simultaneously optimize the center distances and sizes of output estimation sets. At the end, a case study is used to illustrate the effectiveness of the proposed method.

A Novel Online Active Fault Diagnosis Method Based on Invariant Sets

This letter proposes a novel online active fault diagnosis (AFD) method based on invariant sets for linear time-invariant (LTI) systems with bounded disturbances and noises. In general, the system has the healthy mode and several faulty modes. During system operation, with respect to the healthy and faulty modes, the corresponding healthy and faulty output estimation sets are computed, respectively. Thus, the task of AFD can be converted to separate the healthy and faulty output estimation sets, which can consequently distinguish the current system mode and implement AFD. The existing AFD methods implement the objective by designing an input sequence to separate the healthy and faulty output estimation sets. However, it is difficult to use these methods to simultaneously combine AFD and fault-tolerant control. Thus, we propose a new AFD method in this letter to online design inputs step by step, which has potential to overcome some weakness of the existing AFD methods. Particularly, we implement this AFD objective by online designing inputs at each step to maximize the distances of the healthy and faulty output estimation sets. Moreover, this online input design problem can finally be transformed into the resolution of two mixed integer programming (MIP) problems with a similar structure. At the end, a quadruple-tank benchmark system is used to illustrate the effectiveness of the proposed method by comparing with an existing set-based AFD method.

Active fault-diagnosis method using adaptive allocator and fault-tolerant adaptive control system design

A new approach to the active fault diagnosis (AFD) for input redundant plants is presented in this paper. A test signal which helps the diagnosis is injected to the plant in addition to nominal control input in the AFD typically. One feature of the proposed AFD is adaptive allocation of the test signal. The adaptive allocator which distributes the test signal and injects it to the redundant actuators is introduced in this paper. A fault-diagnosis (FD) system that estimates fault location and its magnitude from adaptive parameter of the adaptive allocator is constructed with a simple neural network (NN) model. Furthermore, a A fault-tolerant adaptive control system which includes the adaptive test signal allocator is designed based on the model reference adaptive control (MRAC) technique. The adaptive laws of the adaptive allocator and controller are derived by using a suitable Lyapunov function. By performing experiments using a two-input redundant plant, we show the effectiveness and applicability of the proposed AFD and MRAC system.

Active Fault Diagnosis Method for Vehicles in Platoon Formation

This paper presents an active fault diagnosis (AFD) method with reduced excitation for detection and identification of sensor faults of vehicles in a platoon formation. By introducing a probing signal into the platooning, it will allow an active excitation of the system, reveling a residual component, with the same frequency, that can be explored to obtain a fault identification of specific system faults. A supervisor is introduced to monitor the platoon behavior and activate the auxiliary input whenever the system natural excitation is insufficient for a clear fault diagnosis. This solution will allow the fault diagnosis to behave as active or passive through the adaptive signal provided by the supervisor. A dual Youla-Jabr-Bongiorno-Kucera (YJBK) matrix transfer function, also known as fault signature matrix (FSM) is investigated to get a fault diagnosis. In order to obtain an online identification of specific faults in the system, a Taylor approximation of the FSM is pursued. Computational simulations with a high-fidelity full-vehicle model, provided by CarSim , are carried out to demonstrate the effectiveness of the proposed active approach. A direct comparison between an active and a passive behavior in the same scenario shows that the active fault diagnosis method outperforms the passive approach whenever the dynamic behavior does not provide sufficient excitation. Furthermore, the excitation supervisor is able to significantly reduce the amount of artificial excitation introduced into the system ensuring a more energy efficient active fault diagnosis.

Parametric Fault Diagnosis of an Active Gas Bearing

Recently research into active gas bearings has had an increase in popularity. There are several factors that can make the use of gas bearings favourable. Firstly gas bearings have extremely low friction due to the usage of gas as the lubricant which reduce the needed maintenance. Secondly gas bearings is a clean technology which makes it possible to use for food processing, air condition and applications with similar requirements. Active gas bearings are therefore useful for applications where downtime is expensive and dirty lubricants such as oil are inapplicable. In order to keep as low downtime as possible it is important to be able to determine when a fault occurs. Fault diagnosis of active gas bearings is able to minimize the necessary downtime by making certain the system is only taken offline when a fault has occurred. Usually industry demands the removal of any sensor redundancy in systems. This makes it impossible to isolate faults using passive fault diagnosis. Active fault diagnosis methods have been shown able to isolate faults when there is no sensor redundancy. This makes active fault diagnosis methods relevant for industrial systems. It is in this paper shown possible to apply active fault diagnosis to diagnose parametric faults on a controllable gas bearing. The fault diagnosis is based on a statistical detector which is able to quantify the quality of the diagnosis scheme.

A Survey of Active Fault Diagnosis Methods

During the last three decades, there has been a growing interest in active fault diagnosis (AFD) and several important results have been reported in the literature. This survey aims to provide a compact up-to-date overview of main research directions and to propose a basic classification of AFD methods. The contributions are presented almost in the chronological order as they appeared in the literature and they are divided into two groups that differ in the way the uncertainties are modeled. The survey is concluded with the description of main trends, some application results, and future challenges in the area of AFD.

Input Design for Online Fault Diagnosis of Nonlinear Systems with Stochastic Uncertainty

Fault diagnosis is crucial for ensuring stable and reliable operation of high-performance systems in the presence of abnormal events. System uncertainties often make discrimination between normal and faulty behavior a challenging task. This paper presents an active fault diagnosis (AFD) method for nonlinear systems with stochastic uncertainty. AFD involves the optimal design of system inputs for discriminating between multiple model hypotheses that correspond to various operational scenarios. The proposed AFD method relies on minimizing the probability of error in hypothesis selection subject to hard input and state chance constraints. Moment-based approximations for a bound on the probability of error in hypothesis selection as well as for chance constraint evaluation are introduced in order to derive a tractable surrogate AFD problem that is amenable to online implementations. The performance of the AFD method for offline and online fault diagnosis is demonstrated on a continuous bioreactor with multipl...

Active Fault Isolation: A Duality-Based Approach via Convex Programming

This paper presents the mathematical conditions and the associated design methodology of an active fault diagnosis technique for continuous-time linear systems. Given a set of faults known a priori, the system is modeled by a finite family of linear time-invariant systems, accounting for one healthy and several faulty configurations. By assuming bounded disturbances and using a residual generator, an invariant set and its projection in the residual space (i.e., its limit set) are computed for each system configuration. Each limit set, related to a single system configuration, is parameterized with respect to the system input. Thanks to this design, active fault isolation can be guaranteed by the computation of a test input, either constant or periodic, such that the limit sets associated with different system configurations are separated, and the residual converges toward one limit set only. In order to alleviate the complexity of the explicit computation of the limit set, an implicit dual representation is adopted, leading to efficient procedures, based on quadratic programming , for computing the test input. The developed methodology offers a competent continuous-time solution to the optimization-based computation of the test input via Hahn–Banach duality. Simulation examples illustrate the application of the proposed active fault diagnosis methods and its efficiency in providing a solution, even in relatively large state-dimensional problems.

Degree of fault isolability and active fault diagnosis for redundantly actuated vehicle system

This paper proposes a degree of fault isolability concept and active fault diagnosis method for redundantly actuated vehicle systems. Fault isolability is a structural property related to system dynamics and composition of actuators and sensors. Existing research on testing fault isolability has involved checking whether the system is isolable, i.e., binary in nature. A continuous value rather than a binary metric is needed to evaluate how isolable a given system fault is based on a specific measurement set. After fault components are isolated, the fault type and magnitude are estimated by analyzing residual vectors. In a redundantly actuated system, the number of controls/actuators is greater than the system mobility. Thus, the control input distribution to achieve a given control objective is not unique. In the case of a fault, the active fault diagnosis system adjusts the control input distribution to diagnose the fault. Thus, much more system information can be identified by additional excitation through a redundantly actuated system, which improves the fault diagnosis performance. Simulation results of a four-wheel independently driven and steered vehicle model validated the proposed degree of fault isolability and the effectiveness of the proposed active fault diagnosis method.

Active Fault Diagnosis in Sampled-data Systems

The focus in this paper is on active fault diagnosis (AFD) in closed-loop sampleddata systems. Applying the same AFD architecture as for continuous-time systems does not directly result in the same set of closed-loop matrix transfer functions. For continuous-time systems, the LFT (linear fractional transformation) structure in the connection between the parametric faults and the matrix transfer function (also known as the fault signature matrix) applied for AFD is not directly preserved for sampled-data system. As a consequence of this, the AFD methods cannot directly be applied for sampled-data systems. Two methods are considered in this paper to handle the fault signature matrix for sampled-data systems such that standard AFD methods can be applied. The first method is based on a discretization of the system such that the LFT structure is preserved resulting in the same LFT structure in the fault signature matrix as obtained for continuous-time systems. The other method is an approximation method, where the same structure is obtained for small parametric faults.

Active fault diagnosis by controller modification

Two active fault diagnosis methods for additive or parametric faults are proposed. Both methods are based on controller reconfiguration rather than on requiring an exogenous excitation signal, as it is otherwise common in active fault diagnosis. For the first method, it is assumed that the system considered is controlled by an observer-based controller. The method is then based on a number of alternate observers, each designed to be sensitive to one or more additive faults. Periodically, the observer part of the controller is changed into the sequence of fault sensitive observers. This is done in a way that guarantees the continuity of transition and global stability using a recent result on observer parameterization. An illustrative example inspired by a field study of a drag racing vehicle is given. For the second method, an active fault diagnosis method for parametric faults is proposed. The method periodically adds a term to the controller that for a short period of time renders the system unstable if a fault has occurred, which facilitates rapid fault detection. An illustrative example is given.

AN OBSERVER PARAMETERIZATION APPROACH TO ACTIVE FAULT DIAGNOSIS WITH APPLICATIONS TO A DRAG RACING VEHICLE

An active fault diagnosis method for additive, parametric or multiplicative faults is proposed. It is assumed that the system considered is controlled by an observer based controller. The method is then based on a number of alternate observers, each designed to be sensitive to one or more faults. Periodically, the observer part of the controller is changed into the sequence of fault sensitive observers. This is done in a way that guarantees continuity of transition and global stability using a recent result on observer parameterization. The proposed active fault diagnosis method distinguishes itself from the existing literature in terms of not relying on an exogenous excitation signal. An illustrative example based on a drag racing vehicle is given.