Taking advantage of multiple solutions in configuration-dependent substructure decoupling

Substructure decoupling allows identifying the unknown dynamic behavior of a subsystem starting from the dynamic behavior of both the whole system and the residual part of the system. Substructuring techniques were originally developed to deal with invariant systems. The feasibility of performing substructure decoupling when the coupling conditions among invariant mechanical subsystems are configuration-dependent was recently investigated. Typical examples of such systems could be a lifting crane or a cartesian robot. In principle, when the unknown subsystem is identified by performing substructure decoupling on different configurations of the assembled system, one would expect to get the same solution. In practice, this is not true because the decoupling problem may be affected by ill-conditioning that depends on the configuration of the assembled system. In this work, the redundancy of information provided by multiple solutions is exploited through different strategies to reduce the effects of ill-conditioning.