Mixture design optimization of low-noise pavements

This paper proposes a tool to optimize the mix design of low-noise pavements. An experimental model was developed to predict the rolling noise of a reference car tire as a function of the composition and volumetric characteristics of mixes obtained from in-service pavements. The model enables an analyst to identify which composition parameters need to be altered to improve the acoustic performance of a low-noise pavement. To define the experimental model, several types of asphalt surface layers composed of hot-mix asphalt mixtures with different void contents, different aggregate grading, and different bitumen percentages were analyzed in situ and by laboratory tests. The acoustical properties of pavement surfaces were evaluated by the close-proximity method. The model was defined by using a multivariate nonlinear regression technique to relate composition and volumetric characteristics of asphalt mixtures with rolling noise levels recorded at different speeds. This model, which is a function of several significant parameters of asphalt mixture composition and tire speed, has proved to be highly reliable in predicting car tire rolling noise. Because the model enables the identification of mixture characteristics that require modification in relation to the specific value of the mean traffic speed, it is particularly useful for the optimization of low-noise pavement mix design.