Modeling automated manufacturing systems from specification to simulation
Hsin Rau · 1993
This research aims to develop a CIM (computer integrated manufacturing) control reference model for a manufacturing enterprise, and through it to develop a methodology for modeling the dynamic behavior of an automated manufacturing system (AMS), from specification to simulation. In implementing a CIM system, one needs a rational control reference model to manage the complex information flow with other activities efficiently. The model developed in this study is more realistic than previous ones because it has the broadest control domain, deals with production planning and production control under separate control levels, proposes a special organizational structure for the level which handles production planning, and proposes an alternative organizational structure when the manufacturing entity becomes intelligent enough to make its own decision. Traditional methods (control-theoretic methods, queueing theory, and Petrinets) only allow for modeling certain tasks performed in an AMS. To capture all the activities occurring in an AMS (such as information flow and physical movements), a modeling technique, namely specification language or description technique, is chosen. Due to their well-defined and well-constructed nature, specification languages force people to describe their systems in a manner that is unambiguous, consistent, and precise. These features result in a correct, effective, and economical implementation. After reviewing a number of existing modeling tools, this study has concluded that Estelle, LOTOS and Statecharts/Statemate are useful for modeling the activities in an AMS, but are not adequate for our needs. Therefore, a message-action specification language (MASL) has been developed in this study. MASL has constructs, such as module, message-passing, external-action, and internal-action, so that it can capture different component activities and interactions occurring in AMSs. In general, MASL provides the means to model the control or operation of a dynamic system, such as an AMS. This study also proposes to turn powerful specifications into simulations, that is, to make specifications executable by developing a message-action specification tool (MAST), which serves two main purposes: verifying specifications and running simulations. In addition, a generic AMS that is general enough to represent several actual factories has been developed as a test case for applying both MASL and MAST. The effects of various parameters on the AMS simulations have been examined, and the results provide insightful information to help design and implement AMSs. Using the simulation results, an optimization technique has been applied to determine cases of maximum-profit and/or maximum-productivity.