Diagnosis Of Discrete Event Systems Research Articles

How to use Model Checking for Diagnosing Fault Patterns in Petri nets

Abstract This paper deals with the problem of fault pattern diagnosis in discrete event systems modelled by Petri nets. A general framework for implementing a pattern diagnosis function by model-checking is proposed. Two different approaches with experimental results are presented and compared to a third approach from the literature.

Is LTL model-checking effective for Diagnosability Verification?

Abstract In this work we propose a model checking approach to deal with the problem of the diagnosability verification. In this approach we consider the normal and the faulty behavior of a transition system (TS). We describe the diagnosability property by using an unique linear temporal logic (LTL) formula. Using LTL model checking we can test if the system is not diagnosable if it does not satisfy our proposed LTL-formula. Our approach can be carried out in SPIN model checker which is a tool used for formal verification of models. One advantage of SPIN is that it can handle a large state space which can be useful for diagnosability verification of complex system. We illustrate the effectiveness of our approach by means of a scalable benchmark of a railway level crossing system for n tracks and compare the results found with the ones found using typical tools for verification of diagnosability of Discrete Event Systems (DES).

Weak Diagnosability of Discrete Event Systems

Abstract In this paper, we propose a weaker property for failure diagnosis called weak diagnosability. Weak diagnosability says that, after the occurrence of a fault event, for any continuation, there always exists at least one trajectory of sufficient length along which we can diagnose the occurrence of the fault event. We derive a necessary and sufficient condition for weak diagnosability and develop an algorithm to check the condition. We then extend weak diagnosability to networked discrete event systems which have packet losses and communication delays and call the extension weak network diagnosability. We adopt the transformed automaton technique to deal with packet losses and develop a method for checking weak network diagnosability with respect to packet losses. For communication delays, our results show that they do not change weak network diagnosability when delays are first-in-first-out, but may add time to diagnose the occurrences of fault events.

Fault Diagnosis of Discrete Event Systems Under Unknown Initial Conditions

This paper proposes a novel diagnosis technique for discrete event systems (DESs) plant models. The developed diagnosis tool, so called diagnoser, is able to detect and isolate the occurrence of system's faults without the knowledge of the system's past behavior. This allows the diagnoser to asynchronously begin its diagnosis of a system's behavior at any time instance of system operation (including postfault occurrences); consequently removing the generally required synchronous initialization between a diagnoser and the system under diagnosis. The necessary and sufficient conditions are derived for the diagnosability of a given DES plant under this asynchronous situation. Several examples are provided to illustrate the details of the proposed diagnosis framework.

Diagnosability Analysis of Intermittent Faults in Discrete Event Systems Using a Twin-plant Structure

Most research in fault diagnosis of discrete event systems has been focused on permanent failures. However, experience with monitoring of dynamic systems shows that intermittent faults are predominant, and that their diagnosis constitutes one of the most challenging tasks for surveillance activities. Among the main existing approaches to deal with permanent faults, two were widely investigated while considering different settings: the Diagnoser based approach, and the Twin-plant based approach. The latter was developed to cope with some complexity limitations of the former. In the present paper, we propose a twin-plant based approach to deal with diagnosability of intermittent faults. Firstly, we discuss various notions of diagnosability, while considering the occurrence of faults, their recovery, and the identification of the system status. Then, we establish the necessary and sufficient conditions for each notion, and develop on-the-fly algorithms to check these properties. The discussed approach is implemented in a prototype tool that is used to conduct experiments on a railway control benchmark.

Diagnosability of a class of discrete event systems based on observations

The diagnosability of discrete event systems has been a topic of interest to many researchers. The diagnosability conditions for various systems have evolved based on a regularity condition that is imposed on faulty traces with respect to their observable continuations. Improving upon this weak but necessary condition, a new model of diagnosability that is based on sensor outputs, which are called observations, upon a command input is proposed in this paper. Necessary and sufficient conditions are derived for the proposed diagnosability model. The search performance of the proposed diagnosability condition is of linear complexity in terms of the power set of the system events and observations, compared to the exponential complexity of the search with the existing diagnosability regularity condition. Moreover, a system that is not diagnosable according to the existing diagnosability condition may be diagnosable in the proposed diagnosability model, which includes observations.

Distributed Fault Diagnosis in Discrete Event Systems via Set Intersection Refinements

We extend and verify diagnosability for a class of set intersection refinement strategies, which can be used for distributed state estimation and fault diagnosis in nondeterministic finite automata that are observed at multiple observation sites. These strategies allow observation sites to (periodically) communicate their diagnostic information (i.e., possible states along with corresponding status, such as normal operation and/or fault type) to other observation sites, which subsequently fuse all available diagnostic information (via set intersection operations), and continue operation based on the refined diagnostic information. We verify diagnosability using the proposed distributed protocol, with polynomial complexity, via compositions of (extended versions of) local verifiers, which are capable of capturing the refinement of information under the set intersection operations, as well as the influence of the refinement process on immediate or future diagnosis decisions.

Online Fault Diagnosis in Discrete Event Systems with Partially Observed Petri Nets

This paper investigates the fault detection problem for Discrete Event Systems (DES) which can be modeled by Partially Observed Petri Nets (POPN). To overcome the problem of low diagnosability in the POPN online fault diagnoser in current use, we propose an improved online fault diagnosis algorithm that integrates Generalized Mutual Exclusion Constraints (GMEC) and Integer Linear Programming (ILP).We assume that the POPN structure and its initial markings are known, and the faults are modeled as unobservable transitions. First, the event sequence is observed and recorded. We use GMEC for elementary diagnosis of the system behavior,then the ILP problem of POPN is solved for further diagnosis. Finally, we modeled and analyzed an example of a real DES to test the new fault diagnoser. The proposed algorithm increased the diagnosability of the DES remarkably, and the effectiveness of the new algorithm integrating GMEC and ILP was verified.

On the history of diagnosability and opacity in discrete event systems

This paper presents historical remarks on key projects and papers that led to the development of a theory of event diagnosis for discrete event systems modeled by finite-state automata or Petri nets in the 1990s. The goal in event diagnosis is to develop algorithmic procedures for deducing the occurrence of unobservable events, based on a formal model of the system and on-line observations of its behavior. It also presents historical remarks on the early works on the property of opacity, which occurred about ten years later. Opacity can be seen as a strong version of lack of diagnosability and it has been used to capture security and privacy requirements. Finally, diagnosability is connected with the property of observability that arises in supervisory control. This paper is part of set of papers that review the emergence of discrete event systems as an area of research in control engineering.

Formal verification of intermittent fault diagnosability of discrete-event systems using model-checking

Fault diagnosis of complex and dynamic systems is a crucial and challenging task, essentially with respect to guaranteeing the reliable, safe and efficient operation of such systems. Most research in this field has been focused on permanent failure diagnosis, i.e., once a fault occurs, the system remains indefinitely faulty. However, experience with fault diagnosis in real-life systems shows that intermittent faults, i.e., faults that can be automatically recovered, are predominant and are among the most challenging kinds of faults to detect and isolate. In this paper, we address the formal verification of intermittent fault diagnosability in discrete-event systems. The system is modelled by a finite state automaton and intermittent faults are defined as faults that can automatically recover once they have occurred. Two definitions of diagnosability, regarding the detection of fault occurrence within a finite delay and the detection of fault occurrence before its recovery, are discussed. The diagnosability is analysed on the basis of the twin-plant structure, which is encoded as a Kripke structure, while diagnosability conditions are formulated using LTL temporal logic.

Fault Diagnosis of Discrete-Event systems Based on the Symbolic Observation Graph

Fault diagnosis of discrete-event systems (DESs) has received a lot of attention in industry and academia during the last two decades. In DES based diagnosis, the two main discussed topics are offline diagnosability analysis and online diagnosis. A pioneering approach that led to the development of various techniques is based on the so-called diagnose. However, this approach suffers from the combinatorial explosion problem due to the exponential complexity of construction. To partially overcome this problem, an efficient approach to construct a symbolic diagnoser is proposed in this paper. The proposed approach consists in constructing a diagnoser based on the symbolic observation graph (SOG), which combines symbolic and enumarative representations. The construction of the diagnoser as well as the verification of diagnosability are performed simultaneously on the fly, which can considerably reduce the state space of the diagnoser and thus the overall running time. To evaluate the efficiency and the scalability of the approach, some experimental results are presented and discussed based on a DES benchmark.

Relative diagnosability of discrete-event systems and its opacity-based test algorithm

The notion of relative diagnosability is proposed for logical automata and the concept of diagnosable rate is introduced to characterize the diagnosability property of a discrete-event system, which takes values in the interval [0, 1]. The relationship between relative diagnosability and diagnosability introduced by Sampath et al. that relative diagnosability is weaker than diagnosability for discrete-event systems is analysed. Furthermore, a necessary and sufficient condition for relative diagnosability is presented. In particular, an opacity-based algorithm is developed to test the relative diagnosability, which is polynomial in the number of states of the system. In addition, the proposed algorithm may be used to deal with the problem of diagnosability.

Formula omitted]-diagnosability for active on-line diagnosis in discrete event systems

In this paper, we investigate active on-line diagnosis in discrete event systems. Active diagnosis can be used for fault detection, fault localization, fault-tolerant control, among others. Discrete event systems are general models for complex manmade systems. For the active on-line diagnosis, we do not construct the entire diagnostic automaton off-line. Instead, we look N steps ahead to determine active diagnosability and calculate diagnostic strategies. Thus, we define active N-diagnosability and investigate the relation between active diagnosability and active N-diagnosability. We also develop an algorithm to check active N-diagnosability. If a system is actively N-diagnosable, the algorithm will also give the control that diagnoses the system. We show that there are significant computational advantages for using the on-line approach.

Diagnosability and online diagnosis of discrete-event systems modeled by acyclic labeled Petri nets

We address in this paper the problems of online diagnosis and verification of language diagnosability of discrete event systems (DES) modeled by acyclic labeled Petri nets, in which, different transitions can be labeled by the same event (observable, unobservable and failure). The proposed diagnoser makes its decision regarding the failure occurrence by storing the sequence of observed events and, after each occurrence of observable event, it verifiers if two sets of inequalities are satisfied; the first set accounts for the normal whereas the second one accounts for the faulty behavior of the system. We also consider the problem of diagnosability verification by creating new sets of inequalities that, when satisfied, allow us to decide whether the language generated by the Petri net is diagnosable. Our method for online diagnosis has the advantage over previously ones for relying only on the verification of set of inequalities. Regarding language diagnosability, our verification algorithm does not require any knowledge of automaton theory, being self-contained within the Petri net formalism.

From Diagnosability to Opacity: A Brief History of Diagnosability or Lack Thereof

In the context of the special historical session on the emergence of discrete event systems as an area of research within the control engineering community, this paper presents historical remarks on key projects and papers that led to the development of a theory of event diagnosis for discrete event systems modeled by finite-state automata in the 1990s. About 10 years later, this work was followed by related investigations on the property of opacity, a strong version of lack of diagnosability that captures security and privacy requirements in cyber and cyber-physical systems.

Polynomial Time Verification of Modular Diagnosability of Discrete Event Systems

This paper addresses the modular diagnosability verification of discrete event system (DES). We assume that the language generated by each automaton module can have deadlocks and that the common events among modules are observable. A new algorithm is developed based on the proposition of a necessary and sufficient condition for modular diagnosability, whose complexity is polynomial and is lower than that of other approaches in literature.

Synchronous Codiagnosability of Modular Discrete-Event Systems

Automated systems are subject to failures that can alter their expected behavior, leading to a decrease in their performance, and possibly causing equipment damages. In this work, we propose a decentralized synchronized Petri net diagnosis scheme for discrete-event systems modeled as automata. In order to do so, we define the notion of synchronous codiagnosability and propose an algorithm to verify this property. We also discuss the relation between synchronous codiagnosability, synchronous diagnosability and diagnosability of discrete-event systems.

On‐the‐Fly and Incremental Technique for Fault Diagnosis of Discrete Event Systems Modeled by Labeled Petri Nets

AbstractIn this paper, we develop an on‐the‐fly and incremental technique for fault diagnosis of discrete event systems modeled by labeled Petri nets, in order to tackle the combinatorial explosion problem. K‐diagnosability, diagnosability, Kmin (the minimum K ensuring diagnosability) and on‐line diagnosis are solved on the basis of the on‐the‐fly and incremental building of two structures, called respectively fault marking graph and fault marking set graph, in parallel. We build on existing results, namely those establishing necessary and sufficient conditions for diagnosability, but we bring mechanisms to make the checking of such conditions potentially more efficient. We show that, in general, analyzing or even building the whole reachability graph is unnecessary to analyze diagnosability and build an on‐line diagnoser. Our technique was implemented in a prototype tool called OF‐PENDA, and a railway level crossing benchmark is used to make a comparative discussion pertaining to efficiency in terms of time and memory relative to some existing approaches.

Fault Diagnosis for Discrete Event Systems Modeled By Bounded Petri Nets

AbstractFault diagnosis is an important problem in the manufacturing industry. It has been extensively studied in the past few decades both in time‐driven systems and discrete event systems. This paper presents a Petri net diagnoser for online fault diagnosis of discrete event systems modeled by bounded labeled Petri nets. First, we present the concept and some properties of an extended basis reachability graph. Next, based on such a graph, we construct a Petri net diagnoser that is used to determine if a fault has occurred. Finally, an example is given to illustrate the application of the proposed diagnoser.

Computation of the delay bounds and synthesis of diagnosers for decentralized diagnosis with conditional decisions

We consider decentralized diagnosis of discrete event systems in the conditional disjunctive and conjunctive architectures, where the local failure decision and local nonfailure decision are conditional, respectively. For each of these architectures, a notion of conditional codiagnosability, which guarantees the detection of any failure by conditional decentralized diagnosis within a bounded number of steps, has been defined in the literature. In this paper, we compute the minimum number of steps, called the delay bound, within which the occurrence of any failure can be detected in a conditionally codiagnosable system. The delay bound is important to evaluate the ability of diagnosis. In addition, we use the computed delay bound to synthesize local diagnosers with conditional decisions.