A Petri‐net based machine tool maintenance management system
| Published date | 01 June 1997 |
| Pages | 143-149 |
| DOI | https://doi.org/10.1108/02635579710173211 |
| Date | 01 June 1997 |
| Author | Paul Prickett |
| Subject Matter | Economics,Information & knowledge management,Management science & operations |
[ 143 ]
Industrial Management &
Data Systems
97/4 [1997] 143–149
© MCB University Press
[ISSN 0263-5577]
A Petri-net based machine tool maintenance
management system
Paul Prickett
Lecturer, Cardiff School of Engineering, University of Wales, Cardiff
Considers an integrated
approach to maintenance
management, based on the
development of an advanced
machine tool failure diagnosis
and recovery system. Outlines
how such systems can be
utilized to effectively monitor
the ever more complicated
machining facilities used in
modern manufacturing.
Illustrates how a Petri-net
based system has been devel-
oped to meet the needs of a
modern computer-based
maintenance facility and the
potential that such systems
provide in a move towards
total productive maintenance.
Deals with the steps taken to
ensure that such a system
can be successfully deployed,
and considers how the result-
ing integrated maintenance
management system can
benefit existing users of
advanced manufacturing
technology.
Introduction
This paper outlines the development of an
integrated maintenance management sys-
tem. The need for such systems and their
importance within modern manufacturing
environments was identified by an initial
study[1]. The system itself was then devel-
oped by a collaborative research project,
funded under the European EUREKA Maine
project initiative as EU744: Integrated Condi-
tion and Machining Process Monitoring Sys-
tem for Flexible Manufacturing Systems and
for Stand-alone CNC Machine Tools[2].
The system is based on the use of Petri-net
models. Using a computer-aided modelling
tool developed during the Eureka Project it is
possible to represent the machine tool or
other system being monitored as fully func-
tional models that can be linked directly to
the machine elements or to the controller.
The developed system can then model every
action taken by the actual machine, system or
sub-system. In this way the initial goal set for
the developed system can be met, namely that
of fault condition determination. Since it is
possible to match machine status to the
expected status at any time, any deviation
may be identified, thus ensuring that fault
conditions may be indicated almost instantly.
On detecting such a failure the most appro-
priate recovery action may be initiated using
the system which has associated decision-
making facilities. All such actions may then
be input into a maintenance management
database.
In addition to the above functions, associ-
ated with failure diagnostics, the developed
system can also play a major role in the
deployment of total productive maintenance
(TPM). Since maintenance management
functions are now able to access information
pertaining to the operation of a machine
system it is possible that key attributes, such
as cycle times, cycle counts and other selected
operational variables can be monitored. This
then allows the assessment and maintenance
of machine and or process condition, so as to
facilitate the continuous provision of accu-
rate manufacture.
Petri-net based monitoring
systems
Petri-net models are widely utilized as a
process control mechanism[3]. They are a
powerful tool, enabling users to graphically
design and monitor complicated manufactur-
ing or process-based activities in a simple, yet
comprehensive, manner. As such they are
ideally suited to the application described in
this paper, namely in the monitoring of
machine tool related operations and proce-
dures.
Using the Petri-net approach a graphical
representation of any logical process can be
produced[4]. The developed models are easily
understood. As pre-set conditions are satis-
fied events proceed. In this way the current
status of every element within the system
being modelled is represented on the Petri-
net. It is thus possible to monitor any process
via such nets and, as described in this paper,
to then link the Petri-net monitoring system
to maintenance management tools. The
resulting system can then be deployed as part
of an integrated maintenance management
system to monitor aspects of the process, and
provide diagnostic routines on machine or
process failure.
The advantage of the developed system is
that the above method can be used to provide
a history of events that lead up to any particu-
lar failure. In this way any maintenance
action is not based purely on the state of the
system when the action is attempted, as with
other existing methods, but on the state of the
system just prior to failure. An added tool
that is made available at this stage is the
ability of the developed system to “play back”
the events occurring prior to, and during, the
system failure. Such information is of great
benefit to those attempting to regain the func-
tionality of the machine or system in ques-
tion.
A simple example of a Petri-net model pro-
duced using the Petri-net analyser developed
in the Cardiff School of Engineering, is
shown in Figure 1. In this Figure conditions
or machine states are modelled as places in
the net and represented as circles such as
PX1. Events, i.e. actions that contribute to the
completion of a machine or system cycle are
To continue reading
Request your trial