Uncertainty model for contact instability prediction.

Contact between sliding bodies can cause vibrations leading to instability. The problem of squeal due to high frequency noise from brake systems is due to unstable vibrations generated at the contact interface between the pad and disk. Squeal noise is characterized by extreme unpredictability due to large uncertainties on the values of parameters of the system. Parametrical complex eigenvalue analysis is a common tool used to predict squeal instability. In this paper a substructured linear finite element model of a simplified brake system is studied. A parametrical analysis is focused on a test case and compared to experimental results. The analysis is developed as a function of the parameters assumed to be the most influential but also the most uncertain: friction coefficient and the parameters driving the dynamics of the system. The uncertainties are accounted for by considering parameters such as random variables. A Monte Carlo simulation and a probabilistic technique are performed simultaneously to study the probability of squeal occurrence. Finally, a reduced model based on the transfer function calculated at the contact is developed to perform the analysis with reduced computational effort.