Uncertainty Analysis for the Prediction of Disc Brake Squeal Propensity

Since brake squeal was first investigated in the 1930s, it has been a noise, vibration and harshness (NVH) problem plaguing the automotive industry due to warranty-related claims and customer dissatisfaction. Accelerating research efforts in the last decade, represented by almost 75% of the papers published in the open literature, have improved the understanding of the generation mechanisms of brake squeal, resulting in better analysis of the problem and better development of countermeasures by combining numerical simulations with noise dynamometer tests. However, it is still a challenge to predict brake squeal propensity with any confidence. This is because of modelling difficulties that include the often transient and nonlinear nature of brake squeal, and uncertainties in material properties, operating conditions (brake pad pressure and temperature, speed), contact conditions between pad and disc, and friction. Although the conventional Complex Eigenvalue Analysis (CEA) method, widely used in industry, is a good linear analysis tool for identifying unstable vibration modes to complement noise dynamometer tests, it is not a predictive tool as it may either over-predict or under-predict the number of unstable vibration modes. In addition, there is no correlation between the magnitude of the positive real part of a complex eigenvalue and the likelihood that the unstable vibration mode will squeal. Transient nonlinear simulations are still computationally too expensive to be implemented in industries for even exploratory predictions. In this paper, a stochastic approach, incorporating uncertainties in the surface roughness of the lining, material properties and the friction coefficient, is applied to predict the squeal propensity of a full disc brake system by using CEA on a finite element model updated by experimental modal testing results. Results compared with noise dynamometer squeal tests illustrate the potential of the stochastic CEA approach over the traditional deterministic CEA approach.