Hierarchical and distributed discrete event control of manufacturing processes

Abstract

The supervisory controllers used in manufacturing processes to establish nominal sequences of production steps can theoretically be synthesized based on the supervisory control theory (SCT). The fact, that in practice manual design procedures are preferred over the SCT, is (among other reasons) due to the fact that the models used in the SCT are of an abstract type rather than reflecting the signal flow within a discretely controlled process, and that sets of permissible behaviors are obtained rather than selected (optimal) behaviors within this set. This paper proposes a modeling and synthesis scheme for discrete control which aims at retaining the closed loop structure of the controlled process. An algebraic representation of the discrete event dynamics is chosen to enable the transfer of concepts from continuous control to the supervisory control task. In addition, the scheme considers that the plant to be controlled often comprises continuous quantities (from which events need to be extracted), and that larger manufacturing processes typically have a modular and hierarchical structure.