Alexandre Philippot

Decentralized diagnosis based on Boolean discrete event models: application on manufacturing systems

This paper proposes a decentralized structure based on a set of local diagnosers to monitor discrete event manufacturing systems with discrete sensors and actuators. The occurrence of any fault entailing the violation of the desired behaviour must be detected and isolated by at least one local diagnoser using its own local observation about the system execution. These local diagnosers infer the faults occurrence using event sequences, time delays between correlated events and state conditions, characterized by sensors readings and commands issued by the controller. A very limited communication among diagnosers is permitted through a simple coordinator based on a set of rules. The goal is to solve the problem of diagnosiss decision ambiguity due to the partial observation of diagnosers. An adapted codiagnosability notion is formally defined in order to ensure that the set of local diagnosers is able to diagnose all faults entailing the violation of the desired behaviour within a bounded delay. We show that the construction of these diagnosers is polynomial in the size of the system. An example is used to illustrate the proposed approach.

UNCONDITIONAL DECENTRALIZED STRUCTURE FOR THE FAULT DIAGNOSIS OF DISCRETE EVENT SYSTEMS

Abstract This paper proposes an unconditional decentralized structure to realize the fault diagnosis of Discrete Event Systems (DES), specially manufacturing systems with discrete sensors and actuators. This structure is composed on the use of a set of local diagnosers, each one of them is responsible of a specific part of the plant. These local diagnosers are based on a modular modelling of the plant in order to reduce the state explosion. Each local diagnoser uses event-based, state based and timed models to take a decision about faults occurrences. These models are obtained using the information provided by the plant, the controller and the actuators reactivity. All local diagnosis decisions are then merged by a Boolean operator in order to obtain one global diagnosis decision. Finally, the diagnosers are polynomial-time in the cardinality of the state space of the system. This approach is illustrated using an example of manufacturing system.

Evolution of plant modeling approaches for discrete manufacturing systems

The paper presents the evolution of plant modeling approaches for discrete manufacturing systems. In order to compensate the difficulties of a general modeling, we present modular approaches which allow us to express simple causalities between the plant elements (PE) We make a comparison between intuitive, theoretical, and practical approaches around a manufacturing example.

Centralised controller for manufacturing systems through liveness extraction approach

Following our work on the control synthesis based on the Ramadge and Wonham theory, we propose a refined centralised control approach. This approach requires modelling the plant, and some safety and liveness constraints. We propose to model the plant by a method based on occurrence and precedence rules to model the plant, and to use logical equations to model the constraints. The control synthesis approach makes it possible to use liveness extraction of controller to get functional behaviour of manufacturing discrete systems. To illustrate these new concepts, we propose to test them on an example of transfer system.

A synthesis approach to distributed supervisory control design for manufacturing systems with Grafcet implementation

In supervisory control, computational complexity and implementation flexibility represent major challenges when a large number of local components compose a target system. To overcome these difficulties, we propose a formal approach to distributed control synthesis and implementation for automated manufacturing systems (AMS). We assume that the system is modelled with automata in a local modular fashion. Local control specifications are defined for each local subsystem by means of logical equations to construct local controllers (LCs). Then, global control specifications, stated as logical implications, are applied to the LCs, which allows synchronisation and cooperative interaction among the subsystems. This paper makes two contributions. First, it outlines a formal method for constructing minimally restrictive and deadlock-free distributed controllers (DCs). Second, it proposes a method for the interpretation of these DCs into Grafcet, which is a graphical modelling formalism widely used to design the controller’s dynamic behaviour for AMS. An experimental manufacturing system illustrates the approach.

Causal Temporal Signature from diagnoser model for online diagnosis of Discrete Event Systems

In Discrete Events System (DES), there are two basic approaches to diagnosis: the first approach is the diagnosers and the second approach is Causal Temporal Signature (CTS) and chronicles. The first approach has limitations including the issue of combinatorial explosion. On the other side, it offers tools to study the diagnosability of the models constructed. CTS are easier to write but pose the problem of the guarantee of the completeness of a given base. This means that there is at least one CTS in the set of CTS for all the faults in the monitored system. This study aims to propose a method to garantee the completeness of a set of CTS. The method is based on a translation of formalism and model of a diagnoser into CTS. From these CTS, a recognition algorithm based on the concept of “world” is used. A “world” is defined as a set of coherent hypotheses of assignment of the event received by the diagnostic task.

Survey on diagnosis of a pick and place benchmark: Special session on diagnosis of Discrete Event Systems: Application on a benchmark

This paper is a survey to present a benchmark for a special session on diagnosis of Discrete Event Systems. This session consists in comparing several diagnosis approaches around a common example. A Pick and Place station from a simulating tool is used to simulate approaches without any risk of injury to man or damage to machine. The paper presents the benchmark components and defines control specification. It also deals with Key Performance Indicators to evaluate the proposition.

Modelling of a discrete manufacturing system by Parts of plant

Abstract The paper presents an original approach to model a discrete manufacturing system by Parts of Plant (PoP). This approach takes into account technical and technological specifications of each plant element. The aim of this works is to realize a reliable simulation of discrete manufacturing systems during the design stage before producing them. Models are established from the process functional chain. They take into account the information distribution of manufacturing system through each PoP with its sensors, pre-actuators and actuators. A PoP library is proposed with their corresponding model. An application example is used to illustrate the approach.

The Standardized Generation and the Robust Filtering of the Command As Tools of Optimization of the Mental Workload of the Systems Engineer

Abstract This paper presents an original approach to standardize the work of electric traction of railway transportation. This approach is developed within an industrial thesis, financed by the SNCF (French acronym for National Society of French Railways) in association with the CReSTIC (Research Centre in Information and Communication Science and Technologies). This approach is composed of two axes. The first axis is the standardized generation of deliverables made by the systems engineers in order to help them keeping their concentration on cognitive task and to avoid repetitive tasks which can lead to mental underload. The second axis is the integration of a robust filter is based on the use of safety constraints within the methodology of standardization. This controller is then constrained by the functional programs, already established and used by the SNCF. The systems engineers can be serene and this filter avoids the stress of decision-making which can lead to a mental overload.

Separated approach for “active” monitoring of discrete event systems

Abstract In this paper, we propose an approach to realize “active” monitoring for Discrete Event Systems. The “active” monitoring means that human operators have a good perception of process state and detect failures by their own observation of process running. This approach uses one model of the operative part to design two complementary monitoring tools. The first tool is used to follow the normal operating function. Indeed only sensors are not sufficient to guarantee an accurate perception of the process state. The idea is to supply high-level information to human operator or automated system in order to improve the situation awareness. However, it is important to supply reliable information. For that, the second complementary tool detects failures of sensors and consequently generates an alarm. This approach is a separated one since it models the behavior of the operative part independently of the command part. Additionally this approach is simple to be implemented and it better controls the combinatory explosion. An experimental prototype based on this approach is developed. Finally two examples of implementation are studied to illustrate the functionality of this approach.