Finding local non-linearities using error localization from model updating theory

Paper in proceeding, 2012

Within the aerospace industry, linear finite element models are traditionally used to describe the global structural dynamics of an aircraft. Ground vibration test data serve to facilitate the validation of models which are then used to characterize the aeroelastic behavior of the aircraft and to predict the responses due to dynamic loads. Thus, it is vital that the models contain the essential dynamics of the aircraft. Observed nonlinearities are judged to be local in nature whereas the main part of the structure behaves linearly under normal loading. In this work we focus on the identification of nonlinear effects and do that based on model updating theory. That includes methods for error localization with proper selection of candidate error parameters. The nonlinearities are treated as local modeling errors not considered in the linear system model. The error localization behavior is studied using synthetic test data from a simple system, known as the ECL Benchmark, with known nonlinear properties.