Diagnosis Of Discrete Event Systems Research Articles

Fault diagnosis of Discrete Event Systems under uncertain initial conditions

A new technique is presented for diagnosing faults in a Discrete Event System (DES) when the state of the system under diagnosis (SUD) is uncertain upon commencement of the diagnosis process. Specifically, a diagnoser is developed that detects, identifies, and isolates faults in a DES relative to an initial belief state: a set of states in which the SUD could be located at the diagnoser’s time of activation. The diagnoser does not need to be synchronously initialized with the SUD. Rather, it can be activated anytime before or after a fault in the SUD occurs, as it does not require knowledge of the SUD’s behavior pre-activation. This form of analysis can be understood as fault diagnosis under uncertain initial conditions. The construction procedure for the diagnoser is given and the new concept of semi-asynchronous diagnosability is introduced. The latter requires all failures in the SUD to be diagnosable by observing the behaviors of the SUD after the diagnoser is activated. Moreover, necessary and sufficient conditions for semi-asynchronous diagnosability of a DES are provided. The semi-asynchronous diagnosis technique presented herein is compared with classical synchronous diagnosis approaches. Illustrative examples are provided in order to explain the introduced concepts and the developed approach.

Silent closure based pair verifier for fault pattern diagnosis of discrete event systems

AbstractThis paper addresses fault pattern diagnosis of discrete event systems, involving fault pattern detection and diagnosability. A fault pattern is modelled as a finite automaton whose accepted language is the objective to be diagnosed, representing the occurrence of complex or composite faults. A verifier for fault pattern detection based on the synchronous product of a system and a fault pattern is proposed. By removing all silent events, a silent closure is calculated based on the synchronous product, which offers computational advantages for systems that have a large number of silent events. An NSC/FSC pair verifier is then computed by taking the product of a normal silent closure and an accepted silent closure. By studying indeterminate cycles of the NSC/FSC pair verifier, necessary and sufficient verification conditions are established, asserting that a system is diagnosable with respect to a fault pattern if and only if there is no indeterminate cycle in the NSC/FSC pair verifier. It is shown that the proposed method requires polynomial time at most. Finally, a case study to illustrate the results is provided.

A protocol for Decentralized Synchronous Diagnosis with Coordination

Recently, a new architecture for decentralized diagnosis of Discrete Event Systems, called Decentralized Synchronous Diagnosis, has been proposed in the literature. In this scheme, local diagnosers are computed based on the fault-free behavior of the system components, which, in comparison with the classical diagnosis approach, reduces the size of local diagnosers for implementation. Although this method has been successfully implemented, its main drawback is the emergence of an exceeding fault-free language of the system for diagnosis, which reduces the diagnosis efficiency. To circumvent this problem, in this paper, a protocol for Decentralized Synchronous Diagnosis with Coordination (DSDC), which refines the diagnosis status using cluster automata of the local component models, is proposed. The protocol is based on the implementation of local state estimators of the fault-free component models of the system which communicate cluster automata to a coordinator. The indication of the fault occurrence is carried out by the coordinator according to the observed behavior of the system. To do so, communication rules between local state estimators and the coordinator are proposed. The protocol prevents the emergence of an exceeding fault-free language for diagnosis, which guarantees the same diagnosis performance as the traditional monolithic approach. In addition, the implementation of the DSDC in a didactic manufacturing system is presented to illustrate the results.

Polynomial-Time Verification of Decentralized Fault Pattern Diagnosability for Discrete-Event Systems

This paper considers the verification of decentralized fault pattern diagnosability for discrete event systems, where the pattern is modeled as a finite automaton whose accepted language is the objective to be diagnosed. We introduce a notion of codiagnosability to formalize the decentralized fault pattern diagnosability, which requires the pattern to be detected by one of the external local observers within a bounded delay. To this end, a structure, namely a verifier, is proposed to verify the codiagnosability of the system and the fault pattern. By studying an indeterminate cycle of the verifier, sufficient and necessary conditions are provided to test the codiagnosability. It is shown that the proposed method requires polynomial time at most. In addition, we present an approach to extend the proposed verifier structure so that it can be applied to centralized cases.

Model-Based Diagnosis of Real-Time Systems: Robustness Against Varying Latency, Clock Drift, and Out-of-Order Observations

Online fault diagnosis techniques are a key enabler of effective failure mitigation. For real-time systems, the problem of identifying faults is aggravated by timing imprecisions such as varying latency between events and their observation. This paper tackles the challenge of diagnosing faults based on partial observations which are subject to timing imprecisions and potentially made out-of-order due to latency. In this paper, we develop a theory of robust real-time diagnosis importing well-established notions from timed automata theory and the diagnosis of discrete event systems. The theory itself enables a foundational understanding and investigation of the problem and its intricacies. Based on this theory, we further devise an online diagnosis algorithm consuming observations incrementally as they are made and enabling diagnosis, whenever possible, within a bounded worst-case delay. We prove the correctness of the algorithm and its properties with respect to the theory. Aiming at practical feasibility, we also show how to obtain sound but not necessarily complete diagnosis results with space and time requirements bounded by the size of the system model and independent of the number of observations. Finally, using a prototypical implementation, we report on first empirical results obtained by simulation of a small excerpt of an industrial automation example.

Optimal sensor selection for diagnosability enforcement of discrete event systems using labeled petri net

AbstractThis paper addresses the problem of optimal sensor selection for ensuring diagnosability of discrete event systems. Given a nondiagnosable discrete event system modeled with labeled Petri nets, a new labeling function can be designed to enforce the system to be diagnosable. An ad‐hoc parallel composition of non‐deterministic finite automata (derived from the original labeled Petri net model) that are observed at multiple observation sites is employed. In order to optimize a given labeling function as well as a sensor selection, an integer linear programming problem is formulated to associate numerical sensor costs with transition relabelings. In a multi‐fault system, this method can determine the type of faults that occur and the fault sequences pertaining to this fault type if multiple faults occur. Examples are presented to demonstrate the proposed method.

The application of the statistical model and radial basis function neural network for the diagnosis of discrete event systems under temporal constraints

The application of the statistical model and radial basis function neural network for the diagnosis of discrete event systems under temporal constraints

Diagnosability of discrete event systems modeled by time-interval automata

We address, in this paper, the problem of diagnosability of time-interval discrete event systems (TIDES), a class of discrete event system that has a single clock structure and whose event occurrence takes place within a time interval after the previous event occurrence. The idea behind the diagnosability of TIDES is to leverage time information to distinguish faulty traces from non-faulty ones, increasing the accuracy of the fault diagnosis system. For this purpose, we use a recently proposed timed model called time-interval automaton (TIA), and, based on this model, we present necessary and sufficient conditions for diagnosability of discrete event systems modeled by TIA and an algorithm for its verification. Examples illustrate all of the results present in the paper.

Codiagnosability for Intersection-Based Decentralized Diagnosis of Discrete Event Systems

In this paper, we examine two intersection-based architectures, named the normal-state-estimator-intersection-based architecture (N-SEI architecture) and the failure-state-estimator-intersection-based architecture (F-SEI architecture), for decentralized diagnosis of discrete event systems. For each of these architectures, the corresponding notion of codiag-nosability is defined. The defined notions of codiagnosability are incomparable with inference diagnosability for the inference-based architecture. Besides, codiagnosability for the N-SEI architecture is weaker than the existing notion of intersection-based codiagnosability.

Fault Diagnosis in Discrete Event Systems Modeled by Signal Interpreted Petri Nets

This paper proposes a new approach for detecting and isolating faults in manufacturing systems using fault-free models of the plant and the controller. The plant and the controller are both modeled by signal interpreted Petri nets. It is shown how both models can be used to predict sensor and actuator signals. By comparing the predicted sensor signals to the observed sensor signals, the occurrence of a fault can be inferred. Further information about the fault can be obtained by determining the actuator signals required by the plant to generate the observed sensor signals and comparing these signals to the expected actuator signals. The proposed approach has a linear time complexity.

Diagnosability of Discrete Event Systems under Sensor Attacks*

This paper considers fault diagnosis in discrete event systems modeled by finite-state automata, according to the theory of diagnosability, but it assumes that an attacker has compromised the communication channel from the system's sensors to the diagnostic engine. The attacker operates according to a general attack model that has been studied previously in the context of supervisory control, but not in the context of fault diagnosis. Specifically, the attacker is able to replace each occurrence of a compromised observable event with a string in an attack sublanguage; in particular, this general model embeds event insertion and deletion as well as static and dynamic attacks. The new notion of CA-diagnosability is defined in order to formally capture the ability of the diagnostic engine to still diagnose the occurrences of faults in the presence of the attacker, as captured by its attack model. This extends prior results on supervisory control under attack, where the corresponding properties of CA-controllability and CA-observability were introduced, to the realm of fault diagnosis. A testing procedure for CA-diagnosability is developed and its correctness is proved. Then, diagnosability theory is used to study conditions under which the presence of the attacker can be detected based on the corrupted observations. The results in the paper are illustrated using an example of a protection relay and a circuit breaker in a power system, where the faults are the failures of the protection relay or of the circuit breaker.

Modular diagnosability of Discrete Event Systems Synchronized by Observable or Unobservable Events

Discrete Event Systems often have a modular architecture. Modularity is indeed an engineering method that allows to split a problem into sub-systems and to favor the reuse of modeling primitives. To take into account this modularity the notion of modular diagnosability arose. However, all of the existing works rely on a strong assumption, namely the observability of the events common to the modules. This work aims at relaxing this assumption and thus at proposing a broader framework for the application of modular diagnosability to systems characterized by regular languages.

Component level diagnosability of discrete event systems based on observations

Component level diagnosability of discrete event systems based on observations

Diagnoser design strategy for discrete event system: Case study of neutralization system

AbstractIn fault diagnosis, we study model‐based approaches to diagnose the discrete event system that efficiently represents the determined ambiguity, unpredictability, and the observation and judgment of several real‐life problems. Diagnosability is a crucial task for system reliability. This article presents a strategy for verifying the diagnosability of discrete event systems (DESs) using conjunctive normal form (CNF). This strategy introduces CNF‐based finite state machine (FSM). First, the scheme considers the model of the system for diagnosis, and CNF represents all transitions of the DES. CNF‐based FSM constructs a diagnoser known as CNF‐based diagnoser. Diagnoser tests whether faulty events can be detected or not in a given system model, that is, DES. The diagnoser verifies the diagnosability of the given DES‐based FSM using the resolve rule. The construction of diagnoser and diagnosability verification with respect to a real‐world industrial system is illustrated. The complexity of the diagnoser construction and diagnosability verification are shown to be efficient.

Multifault Diagnosis of Li-Ion Battery Pack Based on Hybrid System

Multifault diagnosis of a battery pack is a difficult issue, especially when different faults are with similar characteristics. This article presents a multifault diagnosis approach based on a hybrid system for diagnosing cell parametric faults, sensor faults, and relay faults in a Li-ion battery pack. The battery pack is defined as a hybrid system with hierarchical and decentralized diagnostic structure in order to reduce model complexity and improve diagnosability. The multifault detection and isolation is achieved by integrating an equivalent circuit model (ECM) into hybrid automata and tracing continuous/discrete dynamics from cell level to module level. By using a parameter identification-based method, cell parametric faults are distinguished and diagnosed from sensor faults in cell level. Sensor faults and relay faults are diagnosed through the hybrid system mode estimation and discrete event system diagnosis method in module level. The results of similar fault isolation test and random multifault injection experiments on a real Li-ion battery pack show that the different faults are detected and isolated in reasonable time with high accuracy.

Optimistic Fault Diagnosis in Discrete Event Systems by Labeled Petri Nets and Basis Markings

Optimistic Fault Diagnosis in Discrete Event Systems by Labeled Petri Nets and Basis Markings

Online Fault Diagnosis of Discrete Event Systems Modeled by Labeled Petri Nets Using Labeled Priority Petri Nets*

The problem of fault diagnosis of discrete event systems (DESs) concerns the capacity to detect the occurrence of fault events. In this work, we present a new fault diagnoser that uses labeled priority and λ-free diagnoser Petri nets to diagnose the fault occurrences of DESs modeled by Petri nets, which may be bounded or unbounded. By analyzing the diagnoser Petri net possible states reached after firing transitions labeled by the observable events, the diagnoser can infer about the fault occurrence. In order to allow the diagnoser to analyze the possible reachable states of the λ-free diagnoser Petri net and diagnose the fault occurrence, we propose a new approach for the state estimation of λ-free Petri nets. For a class of Petri nets, including some unbounded Petri nets, the diagnoser proposed here has a structure that does not grow indefinitely due to event observations, making the diagnoser able to perform the online diagnosis of a class of Petri nets that previous works are only able to diagnose with structures that are likely to grow indefinitely for specific sequences of event observations.

Fault Diagnosis of Discrete-Event Systems from Abstract Observations

Fault Diagnosis of Discrete-Event Systems from Abstract Observations

Forcing faults diagnosis in K steps in discrete event systems

Discrete Event Systems occur naturally in engineering practice and include industrial processes, production systems, robotics, among others. So, its diagnosis is important. The aim of this work is to present a new approach to enforce faults diagnosis in Discrete Event Systems modeled by Interpreted Petri nets without limited to much its language. This approach is based on a Regulation Circuit that reduces the relative distance between any two pair of transitions that prevent the detection of the firing of transitions whose occurrence indicates that a fault occurred. By modifying the initial marks of the net that make up the Regulation Circuit, language limitation is avoided, and the diagnosis is established in k steps. First, the terminology used in the Interpreted Petri nets and the diagnosis are established, then the relationship between theses nets and the Regulation Circuit is reviewed to determine the characterization of nets where they can be added, they are analysed and then the proposal is made. The proposal is an option to force faults diagnosis in binary Interpreted Petri nets that are not diagnosable without limiting so much its language that represents their behavior.

Correction to: A Communication Protocol for Decentralized Fault Diagnosis of Discrete Event Systems

Correction to: A Communication Protocol for Decentralized Fault Diagnosis of Discrete Event Systems