A systematic method is proposed for control-relevant decomposition of complex process networks. Specifically, hierarchical clustering methods are adopted to identify constituent subnetworks such that the components of each subnetwork are strongly interacting while different subnetworks are loosely coupled. Optimal clustering is determined through the solution of integer optimization problems. The concept of relative degree is used to measure distance between subnetworks and compactness of subnetworks. The application of the proposed method is illustrated using an example process network.
Bibliographical noteFunding Information:
Financial support for this work by the National Science Foundation is gratefully acknowledged.
© 2016 American Institute of Chemical Engineers
- community detection
- hierarchical clustering
- process control