Decomposing complex plants for distributed control: Perspectives from network theory

Abstract

This paper reviews recent research on the application of methods from the theory of networks for developing distributed control architectures for complex plants. The problem is defined as one of decomposing process networks into constituent subnetworks with strong intra-subnetwork and weak inter-subnetwork interactions. These interactions are quantified based on connectivity and response sensitivity information. This perspective is inspired by the community detection problem in networks. Several approaches are discussed based on hierarchical clustering and modularity optimization. The concepts and potential of these methods for developing control architectures for complex plants are illustrated through a case study. Future research directions are also discussed.