Robust distributed model predictive control for systems of parallel structure within process networks

Abstract

The parallel structure is one of the basic system architectures found in process networks. In order to achieve robust control of complex process networks, it is necessary to formulate control strategies that specifically accommodate the characteristics of such parallel systems. In this paper, the competitive coupling and competitive constraints in parallel systems are initially defined. A novel robust distributed model predictive control algorithm is then developed for such parallel systems which deals explicitly with competitive couplings, competitive constraints and uncertainties. The Lyapunov Method is used for the theoretical analysis which produces tractable linear matrix inequalities (LMI). Two simulation studies and an experimental trial are provided to validate the effectiveness of the proposed approach. These consider control of 40 user and 100 user gas boiler heating systems as well as control of two continuous stirred tank reactors (CSTRs) which are connected in parallel.

Decentralized Coordination of Temperature Control in Multiarea Premises

“…[9][10][11] in this direction. Some authors consider general issues of coordination in systems with a certain structural organization of coordinated processes: parallel [12,13], sequential [14], and serial-parallel [15].…”

Distributed static output feedback robust model predictive control for process networks

“…V111 is the header tank which contains acetic ether and V112 is the header tank containing sodium hydroxide. These raw materials are processed in the CSTR simultaneously [23]. The reaction function is A + B = C + D and the output of this experiment is the temperature of the reactor.…”

“…This paradigm has been extensively used for process networks [19], [18], [20], [21], [22]. For example, distributed RMPC has been used in [23] to control process networks with a parallel structure.…”

Output feedback robust distributed model predictive control for parallel systems in process networks with competitive characteristics

“…It is necessary to clearly describe the couplings and constraints within a parallel system before embarking on a control design [6], [7]. The authors have previously studied the parallel system and proposed a state-feedback based distributed model predictive control (DMPC) in which the competitive couplings and constraints have been initially defined to describe characteristics of the parallel system [8]. Assume that two subsystems are connected in parallel in the same bus.…”

“…The authors have solved the competitive coupling and competitive constraint problems by designing a state feedback DMPC for a parallel system in [8]. In this paper all of the states are assumed to be measurable.…”