A Flexibility-Based Inverse Algorithm for Identification of Substructural and Joint Properties

This paper presents an inverse problem methodology for the identification of structural joint characteristics. The underlying theory employs a substructural flexibility method that allows statistical uncertainties in joints and interfaces. The method partitions structures into continuum and localized joint/interface substructural regions. The former are modeled by continuum finite elements and built up with the standard direct-stiffness method. Joint and interface regions are constructed from continuum and/or special elements. The novel aspect of the method is that the coupling of substructures is effected in terms of node-collocated Lagrange multipliers, which leads naturally to a boundary-flexibility formulation. Coupling through interaction forces addresses the concern that delicate localized effects are not masked out in the overall structural model. This heightened sensitivity is significant for inverse problems in which localized properties and uncertainties are deduced indirectly, through a hierarchical “peeling off” process. The method is applied to identify joint flexibilities in a model engine mount, which demonstrates key features of the method and its high-fidelity capability.