A graph-theoretic method for integration of process and control system
(IPCS) syntheses with different controllability notions has been prop
osed in the present paper. The foundation of this integration is a wel
l-established, graph-theoretic approach to process synthesis in conjun
ction with the analysis of structural controllability based on digraph
-type process models. Unambiguous structural representation of the res
ultant integrated process and control systems, IPXS structures in brie
f, has been introduced for unambiguous representation of a process str
ucture, it is rendered possible as an extension of the directed bipart
ite graph, the P-graph. Different set of axioms are proposed for descr
ibing the case of disturbance-rejective regulable and the combinatoria
lly feasible and controllable structures in the special cases consider
ed: the case of structural controllability and the case of fault-toler
ant controllability. These axioms make the synthesis computationally m
ore effective by considering very simple engineering knowledge. The ma
ximal controllable structure of an IPXS synthesis problem has been def
ined as the union of combinatorially feasible and controllable IPXS st
ructures. Thus, the mathematical programming model, e.g. MINLP model,
of an IPXS synthesis problem can be and should be derived from the max
imal controllable structure. Different versions of a fundamental polyn
omial time, combinatorial algorithm are presented for identifying the
maximal controllable structure. The resultant IPXS structures are comp
ared with the structures synthesized without considering their control
systems. (C) 1998 Elsevier Science Ltd. All rights reserved.