A survey of Petri net applications in batch processes
Tianlong Gu
a,b,*
, Parisa A. Bahri
a
a
School of Engineering, Murdoch University, Rockingham Campus, Perth, WA 6150, Australia
b
School of Computer Science, Guilin University of Electronic Technology, Guilin 541004, PR China
Received 23 August 2000;accepted 15 June 2001
Abstract
Batch processes are fundamentally different from continuous chemical processes in both their design and operations, which
pose a diversity of problems to system engineering and process control. As an emerging methodology for batch processes,
Petri net has found its applications in different aspects of modelling, qualitative and quantitative analysis, supervisory and co-
ordinate control, planning and scheduling, and hybrid system design. In this paper, recent research on Petri net applications in
batch processes is presented, and future directions are discussed. # 2002 Elsevier Science B.V. All rights reserved.
Keywords: Batch processes;Petri nets;Modelling;Supervisory control;Discrete event systems;Scheduling
1. Introduction
Petri net has been originated from Carl Adam
Petri's doctoral dissertation work on communication
with automata in 1962, which was initially used as a
general purpose mathematical tool to describe the
casual relationships between conditions and events in
a computer system. Starting in the late 1970s, Petri
net became a very active research ®eld. During that
period, most of the scientists and scholars were from
information processing and computer science com-
munity, and the earlier applications and theory of
Petri nets were aimed primarily at computer and
information processing systems [46,47,49]. Prelimin-
ary research showed that Petri net is a powerful
mathematical and graphical tool in describing
event-driven systems, i.e. discrete event (dynamic)
systems. These systems may be asynchronous, con-
tain sequential and concurrent operations, and
involve con¯icts, mutual exclusion and non-deter-
minism. The probe into applications of Petri net in
industrial engineering systems, particularly manufac-
turing systems, started in the early 1980s. Since then,
Petri net methodology and its applications in auto-
mated manufacturing systems have been thoroughly
explored, and several results could be observed
[22,27,41,51,64±66]. In an industrial manufacturing
system, the product is typically composed of discrete
parts, and the behaviour to transform raw materials
into ®nished goods is dominated by discrete event
activities. The range of problems encountered in its
design and operations is mainly event-related or
event-driven.
In contrast, product in industrial chemical processes
is characterised by ¯uid, and two categories of such
processes are usually distinguished: the continuous
and the batch processes [11,50,51]. In a continuous
process, raw materials are fed and products are deliv-
ered continuously at known ¯ow-rates, and the plant
Computers in Industry 47 (2002) 99±111
*
Corresponding author. Tel.: 61-8-93607-120;
fax: 61-8-93607-104.
E-mail addresses: gu@eng.murdoch.edu.au (T. Gu),
parisa@eng.murdoch.edu.au (P.A. Bahri).
0166-3615/02/$ ± see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0166-3615(01)00142-7