Generalized Hessenberg transformations for reduced-order modelling of large-scale systems†

The problem of constructing a reduced-order model induced by the information structure is considered. A new system representation, denoted as the Generalized Hessenberg Representation (GHR), is introduced and chained aggregation which transforms the system representation into the GHR is described. A ninth-order model of a two-thermal plant power system used in load-frequency studies is presented to demonstrate the potential of the GHR in model reduction. The role of GHR in identifying model structure and suggesting appropriate subsets of adjustable parameters is established.