Promises and pitfalls of decoupling procedures

In this paper, the decoupling problem, i.e. the identification of the dynamic behaviour of a structural subsystem that is part of a larger system, is considered. The dynamic behaviour of the whole system, in terms of Frequency Response Functions (FRFs), is assumed to be known at the coupling DoFs, and eventually at some internal DoFs of one subsystem, together with the physical (finite element) model of the second subsystem. Reliable solutions of the decoupling problem could lead to promising developments, both in the field of diagnostics and in the field of vibration control. Decoupling procedures may either start with impedance-like matrices (impedance based approach) or with mobility-like matrices (mobility based approach). The paper highlights pitfalls due essentially to lack of measured DoFs, that produce ill-conditioning in the neighbourhood of certain frequencies.