Distributed strategy for integration of process design and control: A knowledge engineering approach to the incorporation of controllability into exchanger network synthesis

Abstract

The control system of a process cannot be rationally and completely designed without prior knowledge of the structure of the process. This is particularly true for a process composed of highly interconnected process units, e.g. heat and mass exchanger networks. The design of any process and that of its control system are usually carried out in series as two distinct activities. While integration of these two activities is of the utmost importance, purely sequential integration would give rise to a time-consuming and tedious procedure requiring extensive iteration. At every iterative step of such a procedure, the structure of the process optimally designed in the first step tends to be altered and ceases to be optimal in satisfying the control-related criteria, such as controllability and robustness, in the second step; this procedure is indeed passive. Moreover, the lack of controllability and robustness is the major cause of a catastrophic event, energy inefficiency and generation of “off-spec” products to be treated as waste or to be recycled. Obviously, it is highly desirable that a new strategy be developed for actively integrating the two design activities. In one such strategy, namely, the distributed strategy presented in this paper, the criterion for structural controllability is taken into account simultaneously with the criteria for process design throughout the process synthesis phase, which is the first and most important phase in process design. This distributed strategy aims at the synthesis of a highly controllable process structure, which meets the design criteria, from the outset of design activities. Nevertheless, two major difficulties are encountered in developing the distributed strategy: (1) the complete process structure is unknown until completion of the process synthesis; and (2) the available information relevant to designing the process is almost always incomplete, imprecise and uncertain. Thus, traditional algorithmic techniques are unsuitable for developing the distributed strategy, whereas the knowledge engineering (KE) approach appears to be a viable alternative. A KE-based comprehensive strategy has been successfully applied in the present work to the design of two types of process networks, heat exchanger networks (HENs) and mass exchanger networks (MENs). Each of the resultant networks satisfies the traditionally accepted design criterion of lowest possible total cost; meanwhile, each network also possesses the highest possible degree of structural controllability. This prevents the process structure from being repeatedly modified and also simplifies the design of the process' control system.