Growing need for flexible and responsive development has created a special importance of information-driven analysis and design of manufacturing systems. They are often described as a hybrid system of
continuous and discrete event systems. With this point of view, this study concerns with autonomous
distributed manufacturing system (ADMS), Petri net applications for sequential control design and
recipe engineering of batch processes. Another interest in this area is project evaluation and review
using integrated functional structure modeling methods like IDEF0 besides application of various
simulation tools.
ADMS is considered as a practical resolution to achieve better flexibility and productivity in a new
paradigm known as intelligent manufacturing system. By noticing a variety of objects such as robots,
AGVs, machine tools and jobs can compose autonomous agents in ADMS, this study aims at developing
effective cooperation and niche strategies among groups and each of agents, adaptive learning mechanism
against the changes of production environment, and much more. Since production instructions are usually given to each production stage consisting of more than one unit,
each stage is to accumulate enough information and knowledge about its operation necessary for creating
a schedule of its own. Therefore, a scheduling system with better adaptability to various changes can be
derived more quickly if schedules could be created on the stage-by-stage basis. Then as shown in the
figure, we develop a distributed scheduling system that requires no supervisory control system managing
the creation of a whole schedule. Instead of it, each production stage will revise its own schedule
autonomously toward the goal through communication to other stages.
We can model many production systems including chemical batch processes as discrete event systems.
For the design and control of such systems, we focus on the Petri net that can provide a graphical tool
and a mathematical basis for the development at the same time. By using Petri nets, we concern with
the studies in terms of modeling, sequence logic validation, and simulation/evaluation of sequence
control systems for responsive production.
Standard known as "ISA S-88" has already equipped with hierarchical data structure of production
recipe for chemical batch processes. However, any methodology has not given regarding the recipe
management, e.g., how to derive actual sequence control programs and some instruction to a plant
operator (a control recipe) from a master recipe. Associated with such problems, this study develops a
reasonable data-handling method in batch control systems, which needs recipe information for their
own production activities, such as production planning, scheduling, batch management and recipe
management.
"A Supporting System for Evaluation and Review of
Business Process through Activity-Based Approach", Proc. of 7th Int.Symp. on Process Systems Engineering (Denver, July, 2000).
(pdf file 386KB)
Multi-Agent Oriented Manufacturing System
Autonomous Distributed Scheduling System
Design and Control of Discrete Event System Using Petri Nets
Recipe Engineering of Batch Processing
System Create Lab. Research Areas
(sc-wm—)