Input/output hierarchical clustering in process networks based on relative degrees

In this paper, a divisive hierarchical clustering method is proposed to address the problem of pairing of manipulated inputs and controlled outputs. Specifically, an integer nonlinear optimization problem is formulated to identify groups of inputs and outputs with strong structural coupling quantified by relative degrees. This optimization problem can be solved in a hierarchical manner to generate a hierarchy of block decentralized control configurations. The application of the method is illustrated through a case study on an example chemical process network.