MULTI-AGENTS BASED REFERENCE MODEL FOR FAULT
MANAGEMENT SYSTEM IN INDUSTRIAL PROCESSES
Mariela Cerrada, Juan Cardillo
Universidad de Los Andes, Facultad de Ingeniería, Dpto. de Sistemas de Control, Av.Tulio Febres, Mérida,Venezuela
Jose Aguilar, Raúl Faneite
Universidad de Los Andes, Facultad de Ingeniería, Dpto. de Computación, Av.Tulio Febres, Mérida,Venezuela
Keywords: Multi-Agent Systems, Fault Management, Distributed Control Systems, Automation
Abstract: Nowadays, industrial necessities claims global
management procedures integrating information systems in
order to manage and to use the controlled-processes information and thus, to assure a good process
behaviour. These aspects aim to the development of fault detection and diagnosis systems and making-
decision systems. In this work, a reference model for fault management in industrial processes is proposed.
This model is based on a generic framework using multi-agent systems for distributed control systems; in
this sense, the fault management problem is viewed like a feedback control process and the actions are
related to the making-decision in the preventive maintenance task scheduling and the running of preventive
and corrective specific maintenance tasks.
1 INTRODUCTION
Automation is an important aspect that permits to
improve the industrial processes performance
(Williams et al, 1994) and Fault Management
Systems (FMS) are vital part of the automations
process. In (Bravo et al, 2003), a multi-agent based
automation model is proposed, taking into account
the general objects of production processes:
production planning, production factors
management, processes control, fault management
and abnormal situations management.
In this work, an agent-based reference model for
FMS is p
roposed, as part of the automation model
proposed in (Bravo et al, 2003), and it is based on
the generic framework proposed in (Aguilar et al,
2001). The FMS objectives are achieved by the
coordinated interaction of the agents. This way, the
agent-based FMS provides the assistance in the
detection-diagnosis-decision process, as well as in
the planning and running of maintenance tasks.
2 REFERENCE MODEL FOR
FAULT MANAGEMENT
SYSTEMS
The FMS proposed in this work is composed by two
blocks Monitoring and Fault Analysis Tasks
(MFAT) and the Maintenance Management Tasks
(MMT).The MFAT block is concerned to the
detection, isolation, diagnosis, prediction and
planning. MMT block is concerned to the set up and
running of the maintenance tasks according to the
maintenance plan. A reference model permitting the
adequate interaction between the previous blocks is
showed in figure 1. (Cerrada et al, 2003)
3 MULTI-AGENT SYSTEMS AND
DESIGN METHODOLOGY
Multi- Agents Systems (MAS) theory can be viewed
as an evolution of artificial intelligence, in order to
attain autonomous computational systems. Although
the agent’s definition has been argued into the
Distributed Artificial Intelligence (DAI) researchers
community, it is accorded the autonomy is the main
342
Cerrada M., Cardillo J., Aguilar J. and Faneite R. (2004).
MULTI-AGENTS BASED REFERENCE MODEL FOR FAULT MANAGEMENT SYSTEM IN INDUSTRIAL PROCESSES.
In Proceedings of the First International Conference on Informatics in Control, Automation and Robotics, pages 342-345
DOI: 10.5220/0001139703420345
Copyright
c
 SciTePress

characteristic describing an agent, the autonomy
being the ability to accomplish a task and to reach its
objectives without human or any other assistance
(Weiss, 1999;Waterbury, 2002).
Figure 1: Reference Model for Fault Management System.
MAS-CommonKADS is a general methodology for
the agents modelling (Iglesias, 1998). In (Aguilar et
al, 2004), the MAS-CommonKADS methodology
has been enhanced by incorporating new aspects for
agent modelling. The enhanced methodology
MASINA (Multi-Agent Systems-based INtegrated
Automation) is a new approach for MAS modelling
in industrial automation processes. This new
modelling approach allows the developing of the
Agent Model, Tasks Model, Intelligence Model,
Communication Model and Coordination Model.
4 MAS-BASED REFERENCE
MODEL FOR FMS
The Conception and Analysis phase is only
developed here, and the MASINA methodology is
used in order to propose the models mentioned
above. Actors and cases of use are listed in Table 1.
Table 1: Actor and cases of use
Actor Case of Use
Detector System Monitoring
State Identification
Finder Finding-Failure
Diagnostician Failure Analysis
Predictor Failure Occurrence
Estimation
Scheduler Tasks Scheduling
Redefining Plans
Executor Running Tasks
Reporting Tasks
4.1 Agent Model
The roles of the mentioned actors can be embedded
into the agents defined in the generic conceptual
framework for the Agents-based Intelligent
Distributed Control Systems (AIDCS) (Aguilar et al,
2002). Thus, the following agents are defined,
(figure 2): Specialized Agent Detector, Specialized
Agent Finder, Specialized Agent Diagnostician,
Specialized Agent Predictor, Agent Coordinator,
Agent Controller, Agent Actuator, Agent Observer.
Characteristics and requirements of the Agent
Coordinator are presented in tables 2 and 3.
Figure 2: MAS-based FMS.
Table 2: Agent Coordinator
Agent: Coordinator
Name: Coordinator
Type: Software agent.
Roles: Making decision for the maintenance tasks
planning.
Description: It gathers the information, from the
specialized agents, about the process’ items and it
schedules the maintenance tasks. The timeline should
be defined according to the item’s reliability, the
failure effects and urgent tasks.
Table 3: Agent Coordinator Objective
Objective – Coordinator Agent
Name: To schedule the maintenance tasks according
to the item’s reliability, the failure effects and urgent
tasks.
Type: On-condition objective
Input Parameters: Data from the Specialized
Agents, Observer Agent, and Controller Agent.
Output Parameters: Maintenance plan and/or
corrective maintenance order (urgent tasks).
Activation Condition: Information is received from
Specialized Agents, Observer agent, and Controller
agent.
End Condition: The maintenance plan or corrective
actions are stated.
MULTI-AGENTS BASED REFERENCE MODEL FOR FAULT MANAGEMENT SYSTEM IN INDUSTRIAL
PROCESSES
343

Success Condition: A maintenance plan or
corrective actions are proposed, permitting to achieve
adequate reliability levels on the item’s process.
Failure Condition: ¬Success Condition.
Representation Language: Natural Language.
Description: Coordinator Agent assesses the
different setting in order to create the best general
maintenance plan on a long time horizon (Long Term
Plan (LTP)) and it produces the best set of corrective
actions (urgent tasks) in case of emergency.
Coo ices:
aintenance Plan Proposition, Maintenance Plan
d tasks for the MAS based
FMS.
Table 4: MAS-based FMS Tasks
Agent Tasks
rdinator Agent provides the following serv
M
Redefinition and Specialized Agent Calling
4.2 Tasks Model
Table 4 shows the define
Observer failures
ation.
ate estimation
Abrupt o functional
identific
Operational index estimation.
Maintenance st
Detector t failures
n methods incorporation.
Statistical estimation abou
occurrence.
Selection of detection techniques.
New detectio
Finder Finding Failure
Diagnostician es
is models readjustment.
Failure modes and failure caus
identification.
Statistical estimation about failures
modes.
Statistical estimation about failures
causes.
Failure consequences analysis.
Diagnos
New diagnosis models incorporation
New failure modes identification
New failure causes identification
Predictor of the Reliability curves estimation
process’ items.
Reliability index estimation of the
process.
New prediction models incorporation
Controller enance plan proposition ST maint
ST maintenance plan processing
Coordinator
rective action
LT maintenance plan proposition
Resources Evaluation.
Application order of DIDP tasks
Application order of cor
Maintenance plans redefinition
Actuator On-time maintenance tasks run
Urgent tasks run
4.3 Intellig
ents in the FMS
lligent. A general
structure of the intelligence model is presented
ence Model
Excepting the Actuator Agent, all ag
may be sensitive to be inte
bellow:
Experience
Represe
ntation: Rules.
ype: Based on cases.
is related to the data completeness.
edge parameters tuning
ation.
T
Reliability: It
Processing scheme: Knowl
and new models incorpor
Learning Mechanism
Type: Adaptive
Representation: Rules, neural
techniques, genetic
s: Success or failures during DIDP
hanism: Experiences feedback.
techniques.
Learning Source
tasks.
Update Mec
Reasoning Mechanism
Inform
ation Source: Previous results from de
duling tasks, DIDP tasks.
onship: It decides if the used
n
ed.
ns.
cations scheme of
s and associated
languages. The following conversations have been
the
FMS’s agents.
Activation Source: Sche
Type of Inference: Based on rules.
Task-Objective Relati
algorithm is adequate for DIDP tasks or if a
adequate maintenance plan is propos
Reasoning Strategy: It can be deductive or
inductive: evaluate the causes of success or failure in
DIDP tasks or confront unknown situatio
4.4 Coordination Model
This model describes the communi
the MAS: conversations, protocol
defined: On-Condition Maintenance, Maintenance
Plan, Urgent Maintenance Tasks, Maintenance Plan
Redefinition, Maintenance State, Functional Failure
Identification. The conversation Maintenance Plan
Redefinition (MPR) is presented in table 5.
Table 5: Conversation “MPR”
Objective: To redefine the execution timeline of
maintenance tasks that have not been put up and
running on the process.
Agents: Coordinator, Data-base (MAS-based
Middleware), Human.
Beginner: Coordinator Agent
Speaking interactions: Expecting Maintenance
Tasks Search, Alarm, Maintenance Plan Shipment
Precondition: A particular maintenance task has not
been put up and running on the process.
ICINCO 2004 - INTELLIGENT CONTROL SYSTEMS AND OPTIMIZATION
344

End Condition: A new timeline has been proposed
and a new plan is sent to Data-Base, else, an Alarm is
sent to the Human Agent (User).
Description: This conversation permits the
information search on the Data-Base about the
maintenance tasks that have not put up and running
(Expecting Tasks)
4.5 tion Model
been defined
and they are suitably arranged into the conversations
Spe king Interaction: Expecting Maintenance Tasks
Communica
A set of 21 speaking interaction has
in the Coordination Model. In the case of the
previous conversation Maintenance Plan
Redefinition, the following speaking interactions are
performed: Expecting Maintenance Tasks Search
(Table 6), Alarm and Maintenance Plan Sending
Table 6: Speaking Interaction
a
Search.
Type: Query.
Objective: To search in the Data-Base the expecting
maintenance tasks that have not executed on the process.
Agents: Coordinator, Data-Base (MAS-based
Middleware)
Beginner oordinator Agent. : C
Precondition: An active flag about expecting tasks.
End Condition: The Coordinator Agent receives, from the
Data-Base Agent, the whole information about the
expecting tasks.
Conversations: Maintenance Plan Redefinition, Urgent
Maintenance Tasks.
Description: The Coordinator Agent requests the whole
information about the expecting tasks reported by the
Observer Agent.
4.6 Conclusions
tion and analysis of a Multi-
Agents System-based reference model for Fault
In this work, the concep
Management System has been proposed. This model
has been developed into a generic framework
proposed for Intelligent Distributed Control
Systems. In this sense, the system performs a set of
tasks (actions) permitting the maintenance tasks
planning and the application of specific maintenance
tasks as fault detection, isolation, diagnosis and
prediction. The enhanced methodology MASINA
has provided a set of models permitting to describe
the main characteristics of the MAS. The resulting
models have a generic structure that permits to
incorporate it into the automation process of a
distributed control systems.
REFERENCES
T.J Williams, P. Bernous, J Brosvic, D. Chen, G,
Doumeingts, 1994. Architectures for integrating
manufacturing activities and enterprises, Computer in
Industry 24, pp.11-39.
C. Iglesias, 1998. Definición de una metodología para el
desarrollo de sistemas multiagentes, Ph.D Thesis,
Dpto. de Ingeniería de Sistemas Telemáticos,
Universidad Politécnica de Madrid-España.
G. Weiss, 1999. "Multiagent Systems". The MIT Press.
Massachusets, USA.
J. Aguilar, M. Cerrada, F. Hidrobo, G. Mousalli, F. Rivas,
2001. Application of the Agent Reference Model for
Intelligent Distributed Control System, in: N.
Mastorakis and L. Pecorelli-Peres (Eds.), “Advances
in System Sciences: Measurements, Circuits and
Control”, p.p. 204-210, WSES Press.
B. Waterbury, 2002. Artificial Intelligence Expands
Frontiers in Asset Management, Control Magazine.
J. Aguilar, G. Mousalli, V. Bravo y H. Díaz, 2002.
Agentes de control y de gestión de servicio para el
modelo de referencia SCDIA, Memorias del II
Simposio Internacional de Automatización y Nuevas
Tecnologías, TECNO2002, pp. 45-50, Mérida-
Venezuela.
M. Cerrada, J. Aguilar, J Cardillo, R. Faneite, 2003.
Agent-based reference model for fault management,
technical report, Grant Project I-621-98-02-
A,CDCHT-ULA, Mérida-Venezuela.
C. Bravo, J. Aguilar, F. Rivas, M. Cerrada.2003. Diseño
de una Arquitectura de Automatización Industrial
Basada en Sistemas Multiagentes. Memorias del IV
Congresos de Automatización y Control CAC’2003.
Universidad de los Andes. Mérida-Venezuela.
J. Aguilar, M. Cerrada, F. Hidrobo, F. Rivas, 2004.
Especificación detallada de los agentes del SCDIA,
Technical Report, Grant Project Agenda Petróleo No.
97003817, CEMISID-Universidad de Los Andes.
Mérida-Venezuela.
MULTI-AGENTS BASED REFERENCE MODEL FOR FAULT MANAGEMENT SYSTEM IN INDUSTRIAL
PROCESSES
345