Zineb Simeu-Abazi

A methodology of alarm filtering using dynamic fault tree

This article presents a new approach for filtering the faults, thanks to the defined dynamic fault tree (DFT). The proposed methodology includes the dependencies between fault events in the models. Two problems must thus be solved: they relate to the filtering of false alarms, and the reduction of the size of the ambiguity of fault isolation related to the occurrence of a failure. In response to the expressed need for diagnosis, as well as for the need for filtering and localization of the failures, it is necessary to introduce new dynamic gates, making it possible to translate new dependencies, relationships. Based on previous techniques, the approach presented in this paper is based on four peculiar powerful features. First, the concept of the precedence between events is taken into account in order to resort to an adapted configuration for the fault isolation. Second, another relevant data to establish a diagnosis is to take into account the concepts of redundancies between various sets. The appearance of the same phenomenon on various sets can make it possible to refine the fault isolation. The knowledge of the character of the failures is a third important concept; indeed according to the character of the identified breakdowns, one will be able for example to refine the localization or to filter certain events considered non-representative of the character of the breakdown. Fourth, the time duration of the alarm is a more interesting resource to be exploited. The proposed DFT model can be modularized and each module translated into a High Level Petri Net (HLPN). Translation of DFT modules into HLPN has proved to be very flexible and various kinds of new dependencies can be easily accommodated. In order to exploit this flexibility a new representation, called the event diagram, is introduced.

Fault diagnosis for discrete event systems: Modelling and verification

Abstract This paper proposes an effective way for diagnosis of discrete-event systems using a timed-automaton. It is based on the model-checking technique, thanks to time analysis of the timed model. The paper proposes a method to construct all the timed models and details the different steps used to obtain the diagnosis path. A dynamic model with temporal transitions is proposed in order to model the system. By “dynamical model”, we mean an extension of timed automata for which the faulty states are identified. The model of the studied system contains the faultless functioning states and all the faulty states. Our method is based on the backward exploitation of the dynamic model, where all possible reverse paths are searched. The reverse path is the connection of the faulty state to the initial state. The diagnosis method is based on the coherence between the faulty occurrence time and the reverse path length. A real-world batch process is used to demonstrate the modelling steps and the proposed backward time analysis method to reach the diagnosis results.

Maintenance integration in manufacturing systems by using stochastic Petri nets

Research into both increased performance and improvements in dependability of production systems involves maintenance integration, which allows one to protect their availability and their durability. In the case of large manufacturing systems, maintenance integration is essential from conception, and so it involves a particular development concerned with both model complexity and computing time. A modular modelling approach, based on a cellular decomposition of the system, using stochastic Petri nets and Markov chains has been adopted to implement various maintenance strategies in complex production workshops, with the aim of studying their influence on the system dependability and performance.

Modular Modeling for the Diagnostic of Complex Discrete-Event Systems

For the complex systems, the development of a methodology of fault diagnosis is important. Indeed, for such systems, an efficient diagnosis contributes to the improvement of the availability, the growth of production, and, of course, the reduction of maintenance costs. It is a key action in the improvement of performance of industrial feature. This paper proposes a new approach to diagnose complex systems modeled by communicating timed automata. Each component has been modeled separately by a timed automaton integrating various operating modes while the communication between the various components is carried out by the control module. Starting from each module of the complex system, a single deterministic automaton, called a diagnoser, is constructed that uses observable events to detect the occurrence of a failure. This modeling formalism provides means for formal verification of the complex system model and its diagnoser. The model-checking methods are used to check correctness properties. The steps of the method are described by an algorithm and illustrated through a batch neutralization process. The implementation of the algorithm is also discussed.

Normal signature characterization for system health assessment

Helicopter is a system on which operates a large number of specialities from electrical domain to mechanical domain. Today, diagnosis methods are segmented by field of expertise. Each expert treats the sub-system whose he is responsible for regardless of the results of other specialities. Therefore, there is no relation between specialities. Thus, diagnosis at system level is efficient but, due to the lack of correlation between subsystems, it is incomplete. Within the framework of a helicopter, the operator of maintenance uses all the data recorded during the flight, the results of expert treatments, but also his knowledge, his experience and his capacities of observation and analysis to provide an effective global diagnosis. In order to build relation between fields of expertise and so, to obtain a diagnosis at aircraft level which could be relevant, we try to set up a concept which gets closer, at most, to human judgment but which is not well adapted to industrial environment: the normality. During our study, we established that to reach our objective and to stick at best with the concept of normality, the most relevant solution consists in building a global normal signature. This signature could be illustrated as the image of the aircraft health, qualified as normal. This paper defines the normal signature and explains, in part, the process of its building.

Maintenance in the Manufacturing Systems

Abstract Maintenance can be defined as the combination of all technical and associated admini~trative acti~ns, i~clu~ing supervision actions, intended to keep an item or system in, or restore It to, a sta!e In whIch It can perform its required function. For production equipment, ensuring t?e system functIon should be a prime maintenance target. Maintenance should try to provide the rIght CRAMP pararne~ers (Cost, Rel~ability, Availability, Maintainability, and Productivity) for any automated manufactUrIng system. ThIS paper gives an overview of different methods of maintenance applied to the manufacturing systems and the decision support systems used in the industry.

Identification of abnormal events by data monitoring

Many maintenance actions, such as mechanical, electrical, and hydraulic skills, are mandatory to maintain complex systems in operational conditions. Considerable research has been conducted in these fields to optimize maintenance actions. Most research proposes approaches based on physics: physical model of a specific failure, law of aging, etc. In spite of their performance, these approaches are quite difficult to implement on a complex integrated system. Each field of expertise assesses the good health of a system part using its own experts, its own methods, and, in some cases, its own data. Nevertheless, these fields all make up the same machine, and no interaction between systems is considered. Our study is not based on physical approaches but uses operational data and mathematical tools to diagnose, off-line, the current state of the system. The proposed paper concerns a new concept consisting in characterizing normal system functioning by using data recorded during monitoring. The life profile of this complex system is described by employing all the available data to determine, on the one hand, all normal events and, on the other, to identify abnormal events according to their position compared to the normal envelope defined. The recorded data are then specifically analyzed to characterize the level of criticism of an event considered to be abnormal. This abnormal event could then be assimilated to a global behavioral drift of the studied behavior, which is different to usual behavior. This approach is applied to helicopters by use of all flight recorded data.

Multiple fault diagnosis using mathematical models

Multiple fault diagnosis is a challenging problem because the number of candidates grows exponentially in the number of faults. The multiple fault problems is important, since the single fault assumption can lead to incorrect or failed diagnoses when multiple faults occur. In this work, we present an approach for diagnosing multi faults based on model using techniques of detection and localization. We use an observer to generate residuals for a decision in a stage of monitoring and diagnostic system when disruptions or defects occur. Our contribution is the proposal of a diagnostic method when multiple faults type actuators or sensors affect the system.

Diagnostic and prognostic of hybrid dynamic systems: Modeling and RUL evaluation for two maintenance policies

In the industrial sector, maintenance plays a very important role in carrying out production by increasing system reliability and availability. Thee maintenance decision is based primarily on diagnostic modules, prognostics and decision support. Diagnostic consists of detection and isolation of faults, while prognostic consists of prediction of the remaining useful life of systems. Moreover, recent industrial systems are naturally hybrid: their dynamic behavior is both continuous and discrete. This paper presents an integrating architecture of diagnostic and prognostic in a hybrid dynamic system. Indeed, the diagnostic system is based on controlling task execution times during system operation. This method is based on a general modeling approach using hybrid automata. The model proposed is detailed by studying a two-tank system. To validate the model, a Stateflow controller is used. These failures are anticipated by a prognostics process based on a prediction of the remaining life for each component by taking maintenance policy into account. Two new methods are compared: ABAO (As Bad As Old) and AGAN (As Good As New), based on the type of repair strategy.

Performance Evaluation of Centralized Maintenance Workshop by Using Queuing Networks

Maintenance workshop integrated in the job-shop system is required in order to maintain the production machines, to ensure the continuity of service and thus to contribute to the improvement of the availability. The objective is to propose a maintenance workshop where all corrective maintenance activities are centralized. When a failed equipment from the production workshop occurs, it is sent to the maintenance workshop in order to be repaired with a given arrival rate. The aim is to maximize the operational availability of the production workshop by reducing the sojourn time in the maintenance workshop. This paper proposes a methodology to design the Central Maintenance Workshop, which enables us to evaluate the performance in terms of cost and sojourn time, for a given budget. For that, we propose a modeling framework based on the queuing network models. Simulation results are given in order to illustrate the influence of different parameters, like arrival rate of the failed equipments and the waiting time of the equipment.