Nonlinear dynamics of jointed structures: a multiscale approach to predict fretting wear and its effects on the dynamic response

Abstract

Accurate prediction of the vibration response of friction joints is of great importance when estimating both the performance and the life of built-up structures. The relative motion at the frictional interfaces can lead to a highly nonlinear dynamic response and cause fretting wear at the contact. The latter, by changing the contact surface geometry, affects the contact conditions of the interface and consequently impacts the nonlinear dynamic response of the entire assembly, which today is ignored in the analysis. To address the above issue, a novel multiscale approach that incorporates wear into the nonlinear dynamic analysis is presented. A contact solver, based on boundary integral equations, is implemented to compute local contact stresses and stiffness which, in combination with an energy wear approach, allow to compute fretting wear at the contact interface. The nonlinear dynamic response of the whole system is computed using a multi-harmonic balance approach and a continued iteration between the contact and nonlinear dynamic solvers allows the prediction of the nonlinear dynamic response over time. After describing its implementation in detail, the contact solver results are fully verified against a range of test cases for which an analytical solution is available. A comparison against finite element simulations demonstrates the accuracy and computational benefits of the implemented solver. The limitations of the solver due to its underlying half-space assumption are also discussed. The proposed multiscale approach is applied to an underplatform damper-blade system. A significant impact of fretting wear on the nonlinear dynamic behaviour of the blade-damper system is observed, highlighting the sensitivity of the nonlinear dynamic response to changes at the contact interface due to wear. A strong effect of rough interfaces on the wear rate and the resulting interface parameters was also discovered, making them a crucial component for nonlinear dynamic response predictions over time.