Vehicles generate noise through their power-train, aerodynamics, exhausts systems and
tire pavement interaction. Of these sources, tire pavement interaction is by far the most
dominant source at regular cruising highway speeds. A standard means of mitigating the
environmental hazards of freeway traffic noise is through the use of noise abatement
walls. Such infrastructure, however, can be very expensive and difficult to maintain.
Hence the objective of this thesis is to investigate the possibility of reducing traffic noise
associated with tire pavement interactions through pavement surface modification.
This work was focused on an investigation of pavement surfaces to determine what
texture variables affect pavement noise and examine ways of modifying these to improve
pavement quietness. The first step was the identification and physical conceptualization
of possibly significant noise inducing variables with an emphasis on pavement surface
texture. This resulted in a hypothesized model-form for predicting of noise related to tire
pavement interactions. It was followed by a large scale field campaign that performed
numerous on-board sound intensity (OBSI) measurements on various texture types under
various atmospheric conditions. These measurements along with the proposed modelform
were then used in an unforced stepwise regression process. It was ascertained that
asperity interval (a measure of texture wavelength), texture direction relative to the traffic
direction, and texture spikiness (a measure of the probability density function of the
texture amplitude) were the major surface finish contributors to tire pavement noise.
Contrary to previously held belief, however, the profile depth was not identified as a
significant surface finish texture variable. This analysis also identified air temperature
and pavement ride quality (measured through the international roughness index IRI) as
the significant non-textured contributors to tire pavement interaction noise. The complete regression analysis resulted in a model for predicting OBSI from
measurable pavement surface variables, air temperature and ride quality. This model was
able to reproduce over 90% of the field measurements to within 1.5 dBA which is the
band of the typical human noise detection. It was consequently used to determine the
optimum surface texture for a quiet pavement. In addition, the model was used to predict
the OBSI for the design of two large scale pavement rehabilitation projects. Moreover,
the design pavement texture resulted in a post-construction noise level drop of
approximately 5 dBA. The predicted OBSI pre-construction and post-construction were
within 1 dBA of the field measurements.