Uncertain finite element model for the brake squeal prediction

Brake squeal has been under investigation for decades due to the consistent costs suffered by automotive industries. Although the actual mechanism that leads to squeal noise is still unknown, different approaches have been developed for squeal prediction. A commonly used numerical approach is the complex eigenvalues analysis. Squeal events are predicted as unstable eigenvalues with positive real part. Drawback of this method is that the prediction is sensitive to slight variations in the system model, so that it can only predict squeal occurrence at the exact modelled condition. Since the brake apparatus is a highly complex system subjected to variability due to the mass production, an accurate model is nearly impossible. Therefore it is convenient to use an uncertain model in order to cover the lack of knowledge inherent to the model. In this paper, squeal occurring in an appropriately designed experimental rig is numerically studied by using an uncertain finite element model. Uncertainties derive from the main driving parameters such as the friction coefficient, or brake material properties. The unstable parameter regions are then provided along with their likelihood level.