04-06-2012, 02:09 PM
Modeling, Analysis and Simulation of Failures in a
Materials Handling System With Extended Petri Nets
Modeling, Analysis and Simulation of Failures.pdf (Size: 1.56 MB / Downloads: 46)
Abstract
This paper models, analyzes, accommodates and
simulates potential failures which may occur in a materials
handling system with Extended Petri nets (EPNs). EPNs are
shown to be able to cope with various important failure situations
that occur in a manufacturing system through the definition
of six different types of places. Detailed EPN based subnets
are provided to study and accommodate occurred failures.
INTRODUCTION
FL EXIBILITY, autonomy and reliability are important system characteristics [4]. The mean-time between failures
(MTBF) or the failure rate within a system represent
suitable single reliability measurements [ 151. A (potential)
system failure may be classified in general as extemal (due
to factors extemal to the system) or internal failure (directly
associated with the system itself). Internal failures may be
further classified as hard or soft failures, which in turn may
be hardware failures or sofnvare failures.
THE MATERIALSH ANDLINGS YSTEM
The system hardware configuration and components are
shown in Fig. 2. The specific typeskind of robots, cameras,
laser spreader, etc., are not important at this stage. Avoiding
lengthy details, which are beyond the scope of this paper, the
proposed system characteristics are summarized in the sequel.
CONCLUSION
The objective of this paper has been twofold:
1) To demonstrate the modeling power of EPNs and their
ability to cope with several failure situations within a
system (through the action, subnet and switch places),
and,
2) To apply EPNs to study potential failures within a
complicated materials handling system.
Potential system failures have been modeled and accommodated
with EPN subnets. Failure accommodation at the
lowest possible level has been shown to lead to interacting
subnets. This implies that subnets defined at one level could
communicate and thus interact with each other during the
process of error recovery. It is envisioned that continued
research in this direction could lead to the development of
well defined hybrid synthesis techniques for the modeling and
analysis of intelligent systems.